{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "aa001122",
   "metadata": {},
   "source": [
    "# V2 Spec & Registry: Declarative Configuration\n",
    "\n",
    "BaseAttentive v2.2.0 uses a **spec / registry / resolver** workflow\n",
    "that decouples model *description* from model *construction*. This\n",
    "notebook walks through:\n",
    "\n",
    "1. `BaseAttentiveSpec` — backend-neutral model configuration\n",
    "2. JSON serialisation and reload\n",
    "3. `BaseAttentiveComponentSpec` — selecting component keys\n",
    "4. `ComponentRegistry` — registering a custom builder\n",
    "5. `assemble_model()` — resolving a `BaseAttentiveV2Assembly`\n",
    "6. `BaseAttentiveV2` — building a trainable model from the spec\n",
    "\n",
    "## Why declarative configuration?\n",
    "\n",
    "- **Version-control** model configs as plain JSON\n",
    "- **Reproduce** experiments exactly from a saved spec\n",
    "- **Swap backends** (`tensorflow` / `torch` / `jax`) without changing model code\n",
    "- **Override components** through registry keys instead of constructor rewrites"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb223344",
   "metadata": {},
   "source": [
    "## Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8667d7e7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:10.798936Z",
     "iopub.status.busy": "2026-04-21T02:46:10.798936Z",
     "iopub.status.idle": "2026-04-21T02:46:14.307776Z",
     "shell.execute_reply": "2026-04-21T02:46:14.306504Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Backend: tensorflow\n"
     ]
    }
   ],
   "source": [
    "# ── v2.2.0 Backend Setup ─────────────────────────────────────────────────────\n",
    "# BASE_ATTENTIVE_BACKEND must be set *before* importing base_attentive.\n",
    "# Choose your installed backend: \"tensorflow\" | \"torch\" | \"jax\" | \"auto\"\n",
    "import os\n",
    "os.environ.setdefault(\"BASE_ATTENTIVE_BACKEND\", \"tensorflow\")\n",
    "os.environ.setdefault(\"KERAS_BACKEND\", os.environ[\"BASE_ATTENTIVE_BACKEND\"])\n",
    "import keras  # initialise Keras 3 backend before base_attentive\n",
    "BACKEND = os.environ[\"BASE_ATTENTIVE_BACKEND\"]\n",
    "print(f\"Backend: {BACKEND}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cc334455",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.310804Z",
     "iopub.status.busy": "2026-04-21T02:46:14.310804Z",
     "iopub.status.idle": "2026-04-21T02:46:14.385789Z",
     "shell.execute_reply": "2026-04-21T02:46:14.384699Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BaseAttentive 2.2.0 spec imports OK\n",
      "Resolver backend: tensorflow\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "import pathlib\n",
    "from dataclasses import replace\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "from base_attentive import BaseAttentive, __version__\n",
    "from base_attentive.config import (\n",
    "    BaseAttentiveArchitectureSpec,\n",
    "    BaseAttentiveComponentSpec,\n",
    "    BaseAttentiveRuntimeSpec,\n",
    "    BaseAttentiveSpec,\n",
    "    normalize_base_attentive_spec,\n",
    "    serialize_base_attentive_spec,\n",
    ")\n",
    "from base_attentive.experimental import BaseAttentiveV2\n",
    "from base_attentive.registry import (\n",
    "    ComponentRegistry,\n",
    "    ModelRegistry,\n",
    ")\n",
    "from base_attentive.resolver import (\n",
    "    BackendContext,\n",
    "    BaseAttentiveV2Assembly,\n",
    "    assemble_model,\n",
    "    ensure_backend_registrations,\n",
    ")\n",
    "\n",
    "component_registry = ComponentRegistry()\n",
    "model_registry = ModelRegistry()\n",
    "backend_context = BackendContext.current(BACKEND)\n",
    "component_registry, model_registry = (\n",
    "    ensure_backend_registrations(\n",
    "        backend_context=backend_context,\n",
    "        component_registry=component_registry,\n",
    "        model_registry=model_registry,\n",
    "    )\n",
    ")\n",
    "\n",
    "print(f\"BaseAttentive {__version__} spec imports OK\")\n",
    "print(\"Resolver backend:\", backend_context.name)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd445566",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 1 — Keyword approach (recap)\n",
    "\n",
    "The classic keyword approach is fine for quick experiments.  Everything\n",
    "shown here also applies to v1-style construction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ee556677",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.388464Z",
     "iopub.status.busy": "2026-04-21T02:46:14.388464Z",
     "iopub.status.idle": "2026-04-21T02:46:14.652598Z",
     "shell.execute_reply": "2026-04-21T02:46:14.651380Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\projects\\base-attentive\\src\\base_attentive\\core\\base_attentive.py:148: DeprecatedParameterWarning: BaseAttentive: 'static_input_dim' is deprecated since 2.1.0 and will be removed in 3.0.0. Use 'static_dim' instead.\n",
      "  resolved = resolve_deprecated_kwargs(\n",
      "D:\\projects\\base-attentive\\src\\base_attentive\\core\\base_attentive.py:148: DeprecatedParameterWarning: BaseAttentive: 'dynamic_input_dim' is deprecated since 2.1.0 and will be removed in 3.0.0. Use 'dynamic_dim' instead.\n",
      "  resolved = resolve_deprecated_kwargs(\n",
      "D:\\projects\\base-attentive\\src\\base_attentive\\core\\base_attentive.py:148: DeprecatedParameterWarning: BaseAttentive: 'future_input_dim' is deprecated since 2.1.0 and will be removed in 3.0.0. Use 'future_dim' instead.\n",
      "  resolved = resolve_deprecated_kwargs(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Keyword model: BaseAttentive\n"
     ]
    }
   ],
   "source": [
    "model_kw = BaseAttentive(\n",
    "    static_input_dim=4,\n",
    "    dynamic_input_dim=8,\n",
    "    future_input_dim=6,\n",
    "    output_dim=2,\n",
    "    forecast_horizon=24,\n",
    "    embed_dim=64,\n",
    "    num_heads=8,\n",
    "    dropout_rate=0.15,\n",
    ")\n",
    "print(\"Keyword model:\", model_kw.name)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ff667788",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 2 — `BaseAttentiveSpec`\n",
    "\n",
    "`BaseAttentiveSpec` is a frozen dataclass that stores the full model\n",
    "configuration. The current schema groups fields as follows:\n",
    "\n",
    "| Group | Fields |\n",
    "|-------|--------|\n",
    "| **Required I/O** | `static_input_dim`, `dynamic_input_dim`, `future_input_dim`, `output_dim`, `forecast_horizon` |\n",
    "| **Capacity** | `embed_dim`, `hidden_units`, `attention_heads`, `dropout_rate`, `activation` |\n",
    "| **Architecture** | `architecture.encoder_type`, `architecture.decoder_attention_stack`, `architecture.feature_processing` |\n",
    "| **Runtime** | `runtime.num_encoder_layers`, `runtime.scales`, `runtime.multi_scale_agg`, `runtime.memory_size`, `runtime.use_residuals` |\n",
    "| **Output** | `head_type`, `quantiles` |\n",
    "| **Backend** | `backend_name` |\n",
    "| **Component keys** | `components.*` |\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "gg778899",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.654914Z",
     "iopub.status.busy": "2026-04-21T02:46:14.654914Z",
     "iopub.status.idle": "2026-04-21T02:46:14.668038Z",
     "shell.execute_reply": "2026-04-21T02:46:14.666924Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "spec.embed_dim      : 64\n",
      "spec.quantiles      : (0.1, 0.5, 0.9)\n",
      "spec.backend_name   : tensorflow\n",
      "spec.objective      : hybrid\n",
      "spec.attention_lvls : ('cross', 'hierarchical')\n"
     ]
    }
   ],
   "source": [
    "spec = BaseAttentiveSpec(\n",
    "    static_input_dim=4,\n",
    "    dynamic_input_dim=8,\n",
    "    future_input_dim=6,\n",
    "    output_dim=2,\n",
    "    forecast_horizon=24,\n",
    "    embed_dim=64,\n",
    "    hidden_units=96,\n",
    "    attention_heads=8,\n",
    "    dropout_rate=0.15,\n",
    "    head_type=\"quantile\",\n",
    "    quantiles=(0.1, 0.5, 0.9),\n",
    "    backend_name=\"tensorflow\",\n",
    "    architecture=BaseAttentiveArchitectureSpec(\n",
    "        encoder_type=\"hybrid\",\n",
    "        decoder_attention_stack=(\"cross\", \"hierarchical\"),\n",
    "        feature_processing=\"dense\",\n",
    "    ),\n",
    "    runtime=BaseAttentiveRuntimeSpec(\n",
    "        num_encoder_layers=2,\n",
    "        memory_size=64,\n",
    "        multi_scale_agg=\"last\",\n",
    "    ),\n",
    ")\n",
    "\n",
    "print(\"spec.embed_dim      :\", spec.embed_dim)\n",
    "print(\"spec.quantiles      :\", spec.quantiles)\n",
    "print(\"spec.backend_name   :\", spec.backend_name)\n",
    "print(\"spec.objective      :\", spec.objective)\n",
    "print(\"spec.attention_lvls :\", spec.attention_levels)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "hh889900",
   "metadata": {},
   "source": [
    "### Specs are immutable\n",
    "\n",
    "Specs are frozen dataclasses — attempting to mutate one raises a\n",
    "`FrozenInstanceError`.  To \"modify\" a spec, use `dataclasses.replace`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ii990011",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.670095Z",
     "iopub.status.busy": "2026-04-21T02:46:14.670095Z",
     "iopub.status.idle": "2026-04-21T02:46:14.684715Z",
     "shell.execute_reply": "2026-04-21T02:46:14.682206Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original backend: tensorflow dropout: 0.15\n",
      "Modified backend: torch dropout: 0.2\n"
     ]
    }
   ],
   "source": [
    "spec_torch = replace(\n",
    "    spec, backend_name=\"torch\", dropout_rate=0.2\n",
    ")\n",
    "\n",
    "print(\n",
    "    \"Original backend:\",\n",
    "    spec.backend_name,\n",
    "    \"dropout:\",\n",
    "    spec.dropout_rate,\n",
    ")\n",
    "print(\n",
    "    \"Modified backend:\",\n",
    "    spec_torch.backend_name,\n",
    "    \"dropout:\",\n",
    "    spec_torch.dropout_rate,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "jj001122",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 3 — JSON Serialisation\n",
    "\n",
    "Specs serialise to/from plain JSON — no pickles, no framework artefacts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "kk112233",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.686954Z",
     "iopub.status.busy": "2026-04-21T02:46:14.686954Z",
     "iopub.status.idle": "2026-04-21T02:46:14.699433Z",
     "shell.execute_reply": "2026-04-21T02:46:14.697984Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{\n",
      "  \"static_input_dim\": 4,\n",
      "  \"dynamic_input_dim\": 8,\n",
      "  \"future_input_dim\": 6,\n",
      "  \"output_dim\": 2,\n",
      "  \"forecast_horizon\": 24,\n",
      "  \"embed_dim\": 64,\n",
      "  \"hidden_units\": 96,\n",
      "  \"attention_heads\": 8,\n",
      "  \"layer_norm_epsilon\": 1e-06,\n",
      "  \"dropout_rate\": 0.15,\n",
      "  \"activation\": \"relu\",\n",
      "  \"backend_name\": \"tensorflow\",\n",
      "  \"head_type\": \"quantile\",\n",
      "  \"quantiles\": [\n",
      "    0.1,\n",
      "    0.5,\n",
      "    0.9\n",
      "  ],\n",
      "  \"lstm_units\": 64,\n",
      "  \"attention_units\": 32,\n",
      "  \"vsn_units\": null,\n",
      "  \"architecture\": {\n",
      "    \"encoder_type\": \"hybrid\",\n",
      "    \"decoder_attention_stack\": [\n",
      "      \"cross\",\n",
      "      \"hierarchical\"\n",
      "    ],\n",
      "    \"feature_processing\": \"dense\"\n",
      " ...\n"
     ]
    }
   ],
   "source": [
    "# Serialise to dict -> JSON string\n",
    "spec_dict = serialize_base_attentive_spec(spec)\n",
    "spec_json = json.dumps(spec_dict, indent=2)\n",
    "\n",
    "print(spec_json[:600], \"...\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ll223344",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.701461Z",
     "iopub.status.busy": "2026-04-21T02:46:14.701461Z",
     "iopub.status.idle": "2026-04-21T02:46:14.714040Z",
     "shell.execute_reply": "2026-04-21T02:46:14.713194Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Spec round-trip: OK\n",
      "Reloaded backend: tensorflow\n"
     ]
    }
   ],
   "source": [
    "# Reload from JSON\n",
    "reloaded_dict = json.loads(spec_json)\n",
    "spec_reload = normalize_base_attentive_spec(reloaded_dict)\n",
    "\n",
    "assert spec_reload.embed_dim == spec.embed_dim\n",
    "assert spec_reload.quantiles == spec.quantiles\n",
    "print(\"Spec round-trip: OK\")\n",
    "print(\"Reloaded backend:\", spec_reload.backend_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "mm334455",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.716545Z",
     "iopub.status.busy": "2026-04-21T02:46:14.716545Z",
     "iopub.status.idle": "2026-04-21T02:46:14.729813Z",
     "shell.execute_reply": "2026-04-21T02:46:14.728453Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Spec saved to: D:\\projects\\base-attentive\\examples\\model_spec.json\n",
      "Loaded from file — embed_dim: 64\n"
     ]
    }
   ],
   "source": [
    "# Save to file\n",
    "spec_path = pathlib.Path(\"model_spec.json\")\n",
    "spec_path.write_text(spec_json)\n",
    "print(f\"Spec saved to: {spec_path.resolve()}\")\n",
    "\n",
    "# Load from file\n",
    "spec_from_file = normalize_base_attentive_spec(\n",
    "    json.loads(spec_path.read_text())\n",
    ")\n",
    "print(\n",
    "    \"Loaded from file — embed_dim:\", spec_from_file.embed_dim\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "nn445566",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 4 — `BaseAttentiveComponentSpec`\n",
    "\n",
    "`BaseAttentiveComponentSpec` stores the registry keys used for each\n",
    "logical part of the model. Instead of passing free-form per-component\n",
    "parameter dictionaries, the current API selects named builders such as\n",
    "`\"encoder.temporal_self_attention\"` or `\"pool.mean\"`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "oo556677",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.732189Z",
     "iopub.status.busy": "2026-04-21T02:46:14.732189Z",
     "iopub.status.idle": "2026-04-21T02:46:14.745250Z",
     "shell.execute_reply": "2026-04-21T02:46:14.743755Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dynamic encoder key: encoder.temporal_self_attention\n",
      "Sequence pool key : pool.mean\n",
      "Quantile head key : head.quantile_forecast\n"
     ]
    }
   ],
   "source": [
    "component_keys = BaseAttentiveComponentSpec(\n",
    "    dynamic_encoder=\"encoder.temporal_self_attention\",\n",
    "    future_encoder=\"encoder.temporal_self_attention\",\n",
    "    sequence_pooling=\"pool.mean\",\n",
    "    fusion=\"fusion.concat\",\n",
    "    quantile_head=\"head.quantile_forecast\",\n",
    ")\n",
    "\n",
    "spec_with_components = replace(\n",
    "    spec, components=component_keys\n",
    ")\n",
    "\n",
    "print(\n",
    "    \"Dynamic encoder key:\",\n",
    "    spec_with_components.components.dynamic_encoder,\n",
    ")\n",
    "print(\n",
    "    \"Sequence pool key :\",\n",
    "    spec_with_components.components.sequence_pooling,\n",
    ")\n",
    "print(\n",
    "    \"Quantile head key :\",\n",
    "    spec_with_components.components.quantile_head,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "pp667788",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 5 — `ComponentRegistry`: Registering a Custom Builder\n",
    "\n",
    "The registry stores builder callables keyed by a backend-neutral string.\n",
    "Below, we register a tiny demo pooling builder into a local registry.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "qq778899",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.747635Z",
     "iopub.status.busy": "2026-04-21T02:46:14.747635Z",
     "iopub.status.idle": "2026-04-21T02:46:14.760892Z",
     "shell.execute_reply": "2026-04-21T02:46:14.759799Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Custom builder defined\n",
      "Registered pool.demo_last\n"
     ]
    }
   ],
   "source": [
    "def build_demo_last_pool(\n",
    "    *,\n",
    "    context=None,\n",
    "    spec=None,\n",
    "    axis=1,\n",
    "    keepdims=False,\n",
    "    **kwargs,\n",
    "):\n",
    "    def _pool(x, training=False):\n",
    "        return x[:, -1:, :] if keepdims else x[:, -1, :]\n",
    "\n",
    "    return _pool\n",
    "\n",
    "\n",
    "print(\"Custom builder defined\")\n",
    "\n",
    "component_registry.register(\"pool.demo_last\", build_demo_last_pool)\n",
    "print(\"Registered pool.demo_last\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "ss990011",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.764027Z",
     "iopub.status.busy": "2026-04-21T02:46:14.764027Z",
     "iopub.status.idle": "2026-04-21T02:46:14.838805Z",
     "shell.execute_reply": "2026-04-21T02:46:14.837593Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Assembly type       : BaseAttentiveV2Assembly\n",
      "Custom pool in spec : pool.demo_last\n",
      "Resolved pool type  : function\n"
     ]
    }
   ],
   "source": [
    "## Use the custom component in a spec\n",
    "spec_custom = replace(\n",
    "    spec_with_components,\n",
    "    components=replace(\n",
    "        spec_with_components.components,\n",
    "        sequence_pooling=\"pool.demo_last\",\n",
    "    ),\n",
    ")\n",
    "\n",
    "assembly = assemble_model(\n",
    "    \"base_attentive.v2\",\n",
    "    spec=spec_custom,\n",
    "    backend_context=backend_context,\n",
    "    component_registry=component_registry,\n",
    "    model_registry=model_registry,\n",
    ")\n",
    "\n",
    "print(\"Assembly type       :\", type(assembly).__name__)\n",
    "print(\n",
    "    \"Custom pool in spec :\",\n",
    "    spec_custom.components.sequence_pooling,\n",
    ")\n",
    "print(\n",
    "    \"Resolved pool type  :\",\n",
    "    type(assembly.sequence_pool).__name__,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "uu112233",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.841952Z",
     "iopub.status.busy": "2026-04-21T02:46:14.841952Z",
     "iopub.status.idle": "2026-04-21T02:46:14.854711Z",
     "shell.execute_reply": "2026-04-21T02:46:14.853440Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Registered encoder keys:\n",
      "  encoder.dynamic_window\n",
      "  encoder.hybrid_multiscale\n",
      "  encoder.residual_bilstm\n",
      "  encoder.temporal_self_attention\n"
     ]
    }
   ],
   "source": [
    "REGISTRY_KEY = \"encoder.residual_bilstm\"\n",
    "\n",
    "\n",
    "def build_residual_bilstm(\n",
    "    *,\n",
    "    context=None,\n",
    "    spec=None,\n",
    "    units=64,\n",
    "    num_layers=1,\n",
    "    name=None,\n",
    "    **kwargs,\n",
    "):\n",
    "    layers_ns = (\n",
    "        context.layers\n",
    "        if context is not None\n",
    "        and getattr(context, \"layers\", None) is not None\n",
    "        else BackendContext.current(\"tensorflow\").layers\n",
    "    )\n",
    "    dense_cls = layers_ns.Dense\n",
    "    layer_cls = getattr(layers_ns, \"Layer\", object)\n",
    "\n",
    "    class ResidualBiLSTM(layer_cls):\n",
    "        def __init__(self, units, num_layers, name=None):\n",
    "            try:\n",
    "                super().__init__(name=name)\n",
    "            except TypeError:\n",
    "                super().__init__()\n",
    "                if name is not None:\n",
    "                    self.name = name\n",
    "            self.units = units\n",
    "            self.num_layers = num_layers\n",
    "            self.proj = dense_cls(\n",
    "                units, name=f\"{name}_proj\" if name else None\n",
    "            )\n",
    "\n",
    "        def call(self, inputs, training=False):\n",
    "            projected = self.proj(inputs)\n",
    "            input_shape = getattr(inputs, \"shape\", None)\n",
    "            output_shape = getattr(projected, \"shape\", None)\n",
    "            if input_shape == output_shape:\n",
    "                return projected + inputs\n",
    "            return projected\n",
    "\n",
    "        def get_config(self):\n",
    "            base_config = {}\n",
    "            parent = getattr(super(), \"get_config\", None)\n",
    "            if callable(parent):\n",
    "                base_config = dict(parent())\n",
    "            base_config.update(\n",
    "                {\n",
    "                    \"units\": self.units,\n",
    "                    \"num_layers\": self.num_layers,\n",
    "                }\n",
    "            )\n",
    "            return base_config\n",
    "\n",
    "    return ResidualBiLSTM(\n",
    "        units=units,\n",
    "        num_layers=num_layers,\n",
    "        name=name,\n",
    "    )\n",
    "\n",
    "\n",
    "component_registry.register(\n",
    "    key=REGISTRY_KEY,\n",
    "    builder=build_residual_bilstm,\n",
    "    backend=\"generic\",\n",
    "    description=\"Bidirectional LSTM-style demo encoder with residual projection\",\n",
    "    replace=True,\n",
    ")\n",
    "\n",
    "print(\"Registered encoder keys:\")\n",
    "for key in component_registry.list_keys():\n",
    "    if key.startswith(\"encoder.\"):\n",
    "        print(\" \", key)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "vv223344",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.857373Z",
     "iopub.status.busy": "2026-04-21T02:46:14.856365Z",
     "iopub.status.idle": "2026-04-21T02:46:14.870288Z",
     "shell.execute_reply": "2026-04-21T02:46:14.869135Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Resolved layer type: ResidualBiLSTM\n",
      "Layer units: 128\n",
      "Layer depth: 3\n"
     ]
    }
   ],
   "source": [
    "# Test the builder directly\n",
    "registration = component_registry.resolve(\n",
    "    REGISTRY_KEY,\n",
    "    backend=backend_context.name,\n",
    "    allow_generic=True,\n",
    ")\n",
    "layer = registration.builder(\n",
    "    context=backend_context,\n",
    "    spec=spec_with_components,\n",
    "    units=128,\n",
    "    num_layers=3,\n",
    "    name=\"test_enc\",\n",
    ")\n",
    "print(\"Resolved layer type:\", type(layer).__name__)\n",
    "print(\"Layer units:\", layer.units)\n",
    "print(\"Layer depth:\", layer.num_layers)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ww334455",
   "metadata": {},
   "source": [
    "### Use the custom encoder in a spec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "xx445566",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.872649Z",
     "iopub.status.busy": "2026-04-21T02:46:14.872649Z",
     "iopub.status.idle": "2026-04-21T02:46:14.886128Z",
     "shell.execute_reply": "2026-04-21T02:46:14.884942Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Custom encoder component type: encoder.residual_bilstm\n"
     ]
    }
   ],
   "source": [
    "spec_custom = replace(\n",
    "    spec,\n",
    "    components=replace(\n",
    "        spec.components,\n",
    "        dynamic_encoder=REGISTRY_KEY,\n",
    "        future_encoder=REGISTRY_KEY,\n",
    "    ),\n",
    ")\n",
    "\n",
    "print(\n",
    "    \"Custom encoder component type:\",\n",
    "    spec_custom.components.dynamic_encoder,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "yy556677",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 6 — Building `BaseAttentiveV2` from a Spec\n",
    "\n",
    "The resolved `BaseAttentiveV2Assembly` is a low-level view of the\n",
    "components. For normal use, instantiate `BaseAttentiveV2` with the same\n",
    "spec and let it manage the assembly internally.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "zz667788",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.889280Z",
     "iopub.status.busy": "2026-04-21T02:46:14.888271Z",
     "iopub.status.idle": "2026-04-21T02:46:14.962717Z",
     "shell.execute_reply": "2026-04-21T02:46:14.961566Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model type      : BaseAttentiveV2\n",
      "Forecast horizon: 24\n",
      "Head type       : point\n"
     ]
    }
   ],
   "source": [
    "model_from_spec = BaseAttentiveV2(\n",
    "    static_input_dim=spec_custom.static_input_dim,\n",
    "    dynamic_input_dim=spec_custom.dynamic_input_dim,\n",
    "    future_input_dim=spec_custom.future_input_dim,\n",
    "    output_dim=spec_custom.output_dim,\n",
    "    forecast_horizon=spec_custom.forecast_horizon,\n",
    "    spec=spec_custom,\n",
    "    backend_name=spec_custom.backend_name,\n",
    "    name=\"spec_driven_v2\",\n",
    ")\n",
    "\n",
    "print(\"Model type      :\", type(model_from_spec).__name__)\n",
    "print(\"Forecast horizon:\", model_from_spec.spec.forecast_horizon)\n",
    "print(\"Head type       :\", model_from_spec.spec.head_type)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "aaa00111",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:14.964648Z",
     "iopub.status.busy": "2026-04-21T02:46:14.964648Z",
     "iopub.status.idle": "2026-04-21T02:46:15.666425Z",
     "shell.execute_reply": "2026-04-21T02:46:15.663419Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output shape: (8, 24, 2)\n"
     ]
    }
   ],
   "source": [
    "# Quick smoke-test: forward pass\n",
    "B, T, H = 8, 24, 24\n",
    "x_s = np.random.randn(B, 4).astype(\"float32\")\n",
    "x_d = np.random.randn(B, T, 8).astype(\"float32\")\n",
    "x_f = np.random.randn(B, H, 6).astype(\"float32\")\n",
    "\n",
    "out = model_from_spec([x_s, x_d, x_f])\n",
    "print(\"Output shape:\", np.array(out).shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bbb11222",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 7 — Inspecting the Assembly\n",
    "\n",
    "The assembly object exposes the concrete components chosen by the\n",
    "resolver. This is useful for debugging, backend inspection, and custom\n",
    "pipelines.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "ccc22333",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:15.668517Z",
     "iopub.status.busy": "2026-04-21T02:46:15.668517Z",
     "iopub.status.idle": "2026-04-21T02:46:15.680214Z",
     "shell.execute_reply": "2026-04-21T02:46:15.679298Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Assembled model name  : spec_driven_v2\n",
      "Embed dim from spec   : 64\n",
      "Encoder key in spec   : encoder.residual_bilstm\n",
      "Assembly dataclass    : BaseAttentiveV2Assembly\n",
      "Sequence pool object  : function\n"
     ]
    }
   ],
   "source": [
    "print(\"Assembled model name  :\", model_from_spec.name)\n",
    "print(\"Embed dim from spec   :\", spec_custom.embed_dim)\n",
    "print(\n",
    "    \"Encoder key in spec   :\",\n",
    "    spec_custom.components.dynamic_encoder,\n",
    ")\n",
    "print(\n",
    "    \"Assembly dataclass    :\",\n",
    "    BaseAttentiveV2Assembly.__name__,\n",
    ")\n",
    "print(\n",
    "    \"Sequence pool object  :\",\n",
    "    type(assembly.sequence_pool).__name__,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "ddd33444",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:15.683266Z",
     "iopub.status.busy": "2026-04-21T02:46:15.682166Z",
     "iopub.status.idle": "2026-04-21T02:46:15.695747Z",
     "shell.execute_reply": "2026-04-21T02:46:15.694696Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hidden projection: Dense\n",
      "Output head      : MultiDecoder\n",
      "Backend context  : tensorflow\n"
     ]
    }
   ],
   "source": [
    "print(\n",
    "    \"Hidden projection:\",\n",
    "    type(assembly.hidden_projection).__name__,\n",
    ")\n",
    "print(\n",
    "    \"Output head      :\", type(assembly.output_head).__name__\n",
    ")\n",
    "print(\"Backend context  :\", assembly.backend_context.name)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eee44555",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 8 — Sharing Specs Across Experiments\n",
    "\n",
    "A common pattern is to define a **base spec** and derive per-experiment\n",
    "variants with `dataclasses.replace`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "fff55666",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:15.698582Z",
     "iopub.status.busy": "2026-04-21T02:46:15.698582Z",
     "iopub.status.idle": "2026-04-21T02:46:15.713267Z",
     "shell.execute_reply": "2026-04-21T02:46:15.710139Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "spec_exp_A.json saved — backend=tensorflow, head_type=quantile\n",
      "spec_exp_B.json saved — backend=torch, head_type=point\n",
      "spec_exp_C.json saved — backend=jax, head_type=quantile\n"
     ]
    }
   ],
   "source": [
    "BASE_SPEC = BaseAttentiveSpec(\n",
    "    static_input_dim=4,\n",
    "    dynamic_input_dim=8,\n",
    "    future_input_dim=6,\n",
    "    output_dim=2,\n",
    "    forecast_horizon=24,\n",
    "    embed_dim=64,\n",
    "    attention_heads=8,\n",
    "    dropout_rate=0.1,\n",
    ")\n",
    "\n",
    "# Experiment A: TensorFlow, quantile output\n",
    "spec_A = replace(\n",
    "    BASE_SPEC,\n",
    "    backend_name=\"tensorflow\",\n",
    "    head_type=\"quantile\",\n",
    "    quantiles=(0.1, 0.5, 0.9),\n",
    ")\n",
    "\n",
    "# Experiment B: PyTorch, point output, bigger capacity\n",
    "spec_B = replace(\n",
    "    BASE_SPEC,\n",
    "    backend_name=\"torch\",\n",
    "    embed_dim=128,\n",
    "    hidden_units=128,\n",
    "    dropout_rate=0.2,\n",
    ")\n",
    "\n",
    "# Experiment C: JAX, quantile output\n",
    "spec_C = replace(\n",
    "    BASE_SPEC,\n",
    "    backend_name=\"jax\",\n",
    "    head_type=\"quantile\",\n",
    "    quantiles=(0.2, 0.5, 0.8),\n",
    ")\n",
    "\n",
    "experiment_specs = {\"A\": spec_A, \"B\": spec_B, \"C\": spec_C}\n",
    "for name, s in experiment_specs.items():\n",
    "    pathlib.Path(f\"spec_exp_{name}.json\").write_text(\n",
    "        json.dumps(serialize_base_attentive_spec(s), indent=2)\n",
    "    )\n",
    "    print(\n",
    "        f\"spec_exp_{name}.json saved — backend={s.backend_name}, \"\n",
    "        f\"head_type={s.head_type}\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6f9e6a9",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 9 — Training a Spec-Built Model\n",
    "\n",
    "A `BaseAttentiveV2` model compiles and trains exactly like any Keras model.\n",
    "Here we train the spec-built model on a small synthetic dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "83600c9e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:15.715339Z",
     "iopub.status.busy": "2026-04-21T02:46:15.715339Z",
     "iopub.status.idle": "2026-04-21T02:46:54.138110Z",
     "shell.execute_reply": "2026-04-21T02:46:54.136820Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Final train MSE : 0.9570\n",
      "Final val   MSE : 1.0494\n",
      "Prediction shape: (64, 24, 2)\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "\n",
    "# Synthetic training data\n",
    "rng = np.random.default_rng(0)\n",
    "B, T, H = 64, 24, 24\n",
    "x_s_train = rng.standard_normal((B, 4)).astype('float32')\n",
    "x_d_train = rng.standard_normal((B, T, 8)).astype('float32')\n",
    "x_f_train = rng.standard_normal((B, H, 6)).astype('float32')\n",
    "y_train   = rng.standard_normal((B, H, 2)).astype('float32')\n",
    "\n",
    "# Build a fresh spec-based model for training\n",
    "train_spec = replace(\n",
    "    BASE_SPEC,\n",
    "    embed_dim=32, attention_heads=4,\n",
    ")\n",
    "model_train = BaseAttentive(\n",
    "    static_input_dim=train_spec.static_input_dim,\n",
    "    dynamic_input_dim=train_spec.dynamic_input_dim,\n",
    "    future_input_dim=train_spec.future_input_dim,\n",
    "    output_dim=train_spec.output_dim,\n",
    "    forecast_horizon=train_spec.forecast_horizon,\n",
    "    embed_dim=train_spec.embed_dim,\n",
    "    num_heads=train_spec.attention_heads,\n",
    "    dropout_rate=train_spec.dropout_rate,\n",
    "    name='spec_trained',\n",
    ")\n",
    "_ = model_train([x_s_train, x_d_train, x_f_train])  # build weights\n",
    "model_train.compile(\n",
    "    optimizer=keras.optimizers.Adam(1e-3),\n",
    "    loss='mse', metrics=['mae'],\n",
    ")\n",
    "history = model_train.fit(\n",
    "    [x_s_train, x_d_train, x_f_train], y_train,\n",
    "    epochs=10, batch_size=16, validation_split=0.2, verbose=0,\n",
    ")\n",
    "print(f'Final train MSE : {history.history[\"loss\"][-1]:.4f}')\n",
    "print(f'Final val   MSE : {history.history[\"val_loss\"][-1]:.4f}')\n",
    "\n",
    "y_pred = model_train.predict(\n",
    "    [x_s_train, x_d_train, x_f_train], verbose=0)\n",
    "print(f'Prediction shape: {y_pred.shape}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "5a0673a1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:54.141307Z",
     "iopub.status.busy": "2026-04-21T02:46:54.141307Z",
     "iopub.status.idle": "2026-04-21T02:46:54.433540Z",
     "shell.execute_reply": "2026-04-21T02:46:54.432205Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABQkAAAGMCAYAAAB5zCSNAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzs3Qd4U/X+BvC3u6UFCqXsvfcUERyAigiIinvvcZ336v+619XrvG6vXvW6x3XvLQ4EBVH23pRRZhmFDujM/3l/6UlO0pW2STP6fp4n2oQ0OclJml++5zuiHA6HAyIiIiIiIiIiItJgRQd7A0RERERERERERCS4FCQUERERERERERFp4BQkFBERERERERERaeAUJBQREREREREREWngFCQUERERERERERFp4BQkFBERERERERERaeAUJBQREREREREREWngFCQUERERERERERFp4BQkFBERERERERERaeAUJBQREWkgOnfujDFjxgR7Mxqsf/zjH4iKisKGDRtqfRv8/Ysuusiv2yWhia8T7m++biLxNfDLL7+Yx/L666/79fkRERGR2lOQUERExI/Wr1+PK664Ar1790ajRo3QrFkz9OnTBxdeeCGmTZsW8OeaX5g/++wzhLKff/4Zxx57LJo2bWqeo0MOOQRvvvlmvW4DAxMMMPhyUmA1MGbMmIETTzzRBK8TEhLQsmVL81q4/vrrzfsoHPC14f16adGiBUaMGIEXXngBJSUl9bIdNX3fWwE2nk444YQKr1NUVIT09HRzHe4jERERiXyxwd4AERGRSDF37lyMHj0acXFxuOCCC9CvXz8cOHAAa9aswdSpU9G4cWOMHTs2oNtw7733moDkySefXO7fVq1aZb7wB9O7776Lc889F126dMFtt92G5ORkfPLJJ2abMzMzcfvtt9fLdhx11FF46623PC574IEHsHLlynKXt2rVyi/3eeedd+LWW281AbHa4uspJiYG4e7555/H1Vdfja5du5p936FDB2RlZWHFihXmNcL9w38LB9yfL7/8svnZ4XBgx44deO+993DVVVeZx/P000/X6nY7depk9ndsbGyd3vdVSUxMxHfffYdt27ahTZs2Hv/2xRdfYNeuXeY6IiIi0jAoSCgiIuIn/KKen5+PhQsXYtCgQeX+ffv27UF9rusSnPIHZiYxS4wZY/PmzUNqaqq5/Nprr8XEiRNNNtRZZ51VL8Eh3of3/TDQwyDheeedV+XvMhCUl5eHlJSUGt0ngz2+BHyqEgkBm+LiYhMM7tixIxYsWIAmTZp4/HthYSFyc3MRLrhPvV8zfE3z9cWM1doGCRnQD/T+ZhYhMxAZGL/55ps9/u3VV1/FwIEDTTZkOO0PERERqT2VG4uIiPgJMwbT0tIqDBBS69aty132448/4rjjjjMBMwYE+KWcZYoVYUDl9NNPN5ltDPgx++rss8/GunXrXOWD9MYbb3iUP1bXk5BBgsMPP9xk9THwxZ8///zzctezfp+BtEmTJpnMSJYMn3baaT4FQJcuXWoyk5jtZAUIidvIzEsGEf/3v/8hVPumPffcc+jbt6/ZT4899pj59z///NP0h+vZs6cpneZzwufv008/9aknoXUZszwZOGvfvr3Zt3wNffPNN+Vuo6J+dNZlv//+u8lk5X7k6/Cyyy6rMLgzffp0jBw5EklJSeY1+de//hXLli3zub/b1q1b8X//938YPHiwKafn88Hn5ZFHHvGpvJavgezsbAwfPrxcgJDi4+PRvHnzCvfBv//9b/Nc8z75f56v7L14/vnnm+w43h5fuzfddJMJ7nrja5fBawb1rLLncePG4YcffkBtcfv4GHjfdnz/VFS6W1F/PV967vnyvq8K/5YwQP/aa695XM7Mwu+//x4XX3xxleXifJ74N4CvpaFDh+KVV16p8Lr8ezJkyBDzvPDv1l133WXe7xUpKCjAgw8+aDKxeX3+rZg8ebL5+yciIiKBpUxCERERP+nWrZsJ9rB89pRTTqn2+v/973/xl7/8BYcddhjuuOMOE9xhYIJligz8Pfroo67rfvXVVzj11FPNdRj86d69uwlu8Is8g2/s8cdsIAZGjjzySNMX0Rf/+c9/cM0115geinfffbe5jMEYBvJefPHFcrezZcsWE+iYMmWK2b5FixaZ6+3fv9+UVFeFX/6JwTRv1mWzZ89GKHrqqaewe/duXH755SawxkAHMRjIoOkZZ5xhykN5HQZruP8Z8DznnHN8un2WirJM/e9//7vJpOP9cR+sXr3ap35wzF5lVhiDOrxPBtYYsImOjjavM8tvv/1mgtIM7rH0mQGYDz74ADNnzvT5uVi8eLF5jfM1wNc8gz0sWeXtsZcgXw/VBaYYjGaQie+XXr16+XS/DAjyNX/llVeaYCzLkhnc27NnD+655x7X9ZilevTRR5vHxuu2a9fOvE6feeYZ8zgZJOVzbQXZGNRliTAD1eyJyEAiX4cM4DMI5gsGPq0sU5ZN8zXAwCvf14HEnoG1ed/bXXLJJea1xiAzg8fE7WdZOzMkrVJquy+//NLsf74XGDDm/mCJNf828TXA0n0L3yP828XXMf/GMPOSQcmvv/663O3ytXT88cdj1qxZ5jExI3Pfvn146aWXzH7ia4b7SERERALEISIiIn4xa9YsR1xcnIMfrz169HBcfPHFjv/85z+O5cuXl7vu1q1bHQkJCY6zzz673L9df/31jujoaMe6devM+by8PEeLFi0c6enpjszMzHLXLykpcf3M+77wwgsr3L5OnTo5Ro8e7Tq/Z88eR3JysqNbt26Offv2uS7nz127dnWkpKQ49u7d6/H7vP3333/f43avvvpqc/nKlSurfH54fzExMY4hQ4Y4SktLPf7tr3/9q7mNAQMGOIKFz4330mjatGnmsmbNmjl27NhR7ndyc3PLXcb91bNnT0efPn08Lr/nnnvMbWVkZJS7bNKkSR7PyZ9//mkuv/XWWz1uo6L9y8uioqIcs2fP9rh84sSJjtjYWEdOTo7rsuHDh5vXnfXaosLCQseoUaPM7XB7qpOfn19u/9F5551nXrd8bVfnscceM/fH1wO3ia/5t99+27Ft27Zy17X2AV+Pmzdvdl1eUFBgfpeP0X75wIEDHb169XLs37/f43Y++eQTczuvvfaa67IJEyaYy7777rsq31fVvWa8T3xcFT2XvD7fR974mvB+/iu6rKrXQGXv+4pYt33NNdc4ioqKHK1atXJcfvnlrn/n6/fUU081P/fr189jm4uLix0dO3Z0NG3a1LFlyxaP/cHXEV8Dq1evdl23Q4cOjrS0NEdWVpbrutnZ2eY2vPfHE088UeH+4N8k3o7971dlz4+IiIjUnsqNRURE/IRZOMxiYlYYs1+YLcPhDCzF5CAG+8TWjz76yGTWXXrppSYLyX5iaV1paanJZCJmC/JyZuwwK8obs8Vqg1mLzJpiNpa97JM/8zKWqlrbYGnbtq3JmrNj1pZV4lkVZq8xa4llgyyPZXbX2rVr8a9//ctkChF7OoYiZpmxDNUbMzst3HZmEvL/fE44tIIZlr5gya+9RJSluMy2q+45tb/2OFHXjtvA/n9WeTOz5ebMmYOTTjrJox8js+p4/75iaam1rcx6ZCYfX5/jx483r1sO8KkOX8scjMGsxuXLl5ssP2atsdya74mKXgcceMN/t7CU94YbbjCPkZlttGTJEpPpyGxKvr/s76sjjjjC7C8r45XbzQxIZq5x22v7vmJJLN9L1untt982zzF7lN53330IdczsY9be+++/bwalMNuSGax8r1aEf+M2bdpk/p1/D+z7g30N+Rqw2hXwups3bzYZrpz6bGGJMrOovfG5Y1bzsGHDPPYdX2fM6mQmLLdRREREAkPlxiIiIn40YMAAU65LGzduNKWNLNf79ddfTeCAX5r5ZZoBJGKZcGUY1CErUMSeXv6UkZFh/s/eX96sy+yBTapoqAj73xEDZNVhMIgBJg5FePPNN10lk3yOGNipqEednRWUsmMwjSdezn+vrg9kbbD/XUV27txpphYzKMKfvbH3XnWPqarn1ZfntKrfJ+s2rP1dUXmvryW/xKDcww8/bPYfg7zORDa3vXv3+nQ7DIbzxD6GDBT+9NNPZsgHXxsMXHmXLffp06fcbTAAb3+dWu8rlh/bS5Arel9Z217d+4rBcu/ejvZ+gyzL9X4fM6A5YcIE00+QPTut7QxVDOKxz+bHH3+MadOmmeBfRYHTmv7dsP7PwJ+3ip4T7j8GAfk3oTIMGlrl/iIiIuJfChKKiIgECHvUMQPN6hfGDB0OumBGkxVYYaCFwxUqUh9TfmuKAZHKeAeLKsu6YvCHQSYGhhho4ZAOBmwqCybYsVfZ2LFjPS5jMIjBGPYBZFC2ptvki4r6KPK2mQnHwAYz8dgrjRlSfI6YRfrOO++YrKq6PK++bn9d90tN3HjjjaY/4Jlnnml67jHDktmI8+fPxy233OLzY7ZvO4PrPDGbkP022ROP/TKrelxVPVZmKjJDsLKM1ppg8IxZgXYMpFU0BMiOQTZmKrI/pBUQq2ygCAOvwcTtYyYqh/Owxyl7Adb0ufcH7j++Dp544olKr1NVAFFERETqRkFCERGRAGNggF/AGSTk4A/q0aOH+T9L8KrKJrRnsXE4BYNS/mIFITlg4ZhjjvH4Nwbw7NfxNwZqOIjAYk3y5aTVqjCg6D111trGxx9/3OcsNn9gWStLpjmMwTuIVNGwh2CzBqBwWIi3ii6rDAdlsHyegyrsrEBvXfD9wGEoDDgyY4xDTixWlmBVr1PrfVVRdp83BiP53uT7qioM9DOwb1fZBHM7a3pvTk6ORwYis4m9eWfsBgPLhznoxfrZl78b1e0P6/8c7lPZde24/zj4haXytW2jICIiIrWnT18RERE/YfCqoowgls9ZfdCsjCL29UtISDBZcBX12GJPQ2saMAODDJ4wCLZt27YqM8WssltfsMcXe7QxK8weyODPvIy35et017pg+eIjjzxigqGnn356tcFFBn/sJysQwT5m3v8WSFamlXemHjOxONE11LD0mtmOLI22B6UYzGKZb00et/djZm/LJ5980qffZ79B74xPC0vrGTzi6907Y4zTojMzM13nWVrO++T2cLIzsXS4f//+eOGFFyoMvPH9ab0/GLBjSfC3335brvcmWY+Rry/v11V12Yj8XasvH1+XFr7G+f5iRrGFmZe+PneVqcn7vjJnnXWW+XvE14IVbK3I0KFD0bFjR5Mty2nT9tcRJ54z8MrWCtZjZx9JXteaAE3s1cl9VFFAlrdZWSahVSouIiIigaFMQhERET/hEAX2fzvxxBNNyRxLVNm0n2WnHATAL8C8nPjF+fnnn8dll11meq2xJJnlycyi4fCFzz77zARLmP3F23nllVdMbzMGQPg7zILidTnUhOWf1pfyww47zAQ8GHTjF3l+YeeX/4qkpqaaoSHXXHONyXTkMBFiT0VmhbEsmOWz/sTb/Oqrr0z5NQNBzDDi0BL2oPvwww9N4DRccL+xBxufQwa+2NeP+5mPkfu5ooyxYGPpLAO/o0aNMkN1uH8/+OADVy/Hysph7fg65GNkuTEDZgzcsI+g1QOxOnyuWKrL1zJLghmQYlCNrwWW3x88eNCUvXpnkjHAxtcpB140btzYvK84iOWuu+5y9ajj9jPTkZloAwcONBlx3Ee8T76mP/nkEzz00EOu1/qzzz5rngsGCzlwiEEtBu3/+OMP897j+6g6DDxy4IaFvSl5P8wcZoDfnqV7xRVXmGD/lClTTIk6y+05xKiu5cY1ed9Xhr0zWbZfHQZl+bzxMXDADh8T9wcHn8yePRu33367R0YnA6A8KHLooYfi8ssvN+916/XCASh2fE54sOWmm27Czz//bPYjt4vXY89KtitgqbeIiIgESB0mI4uIiIjN999/77j66qsdAwcOdKSlpTliYmIczZs3d4wZM8bxyiuvOEpKSso9X7/99pvj5JNPdqSnpzvi4uIcbdq0Mdd/7LHHHAcOHPC47h9//OE46aSTzG3Hx8c7OnTo4DjnnHMc69atc11n9erVjnHjxjkaN27MNChzsnTq1MkxevToctvwySefOEaOHOlo1KiROfHnTz/9tNz1Kvv9adOmmft57bXXqn09zJgxwzy+Fi1amMfQsWNHx1VXXeXYsmVL0F9LfGzeS6PqHtuGDRscp512mnk8SUlJjuHDh5vn85577jG/l5GR4bqur5dV9XzzuhdeeGG1lxG3mf/Gx2D3008/OUaMGOFISEhwtGrVynH99dc7Zs+eba77yCOPVPs85eXlOf7+97+bfcfb6N69u+Ohhx5y/Pjjjz69DoqKihyvvvqq46yzznL07NnTvFb52m/btq1jypQpjp9//rnSffD000+b++Nrh/9/6qmnKt0vV155pXkOedt8Hw4dOtRx6623OjZt2uRx3czMTHNdvp943ZYtW5r3EB+Pr68Z+ykxMdHRv39/xwMPPOA4ePBgud/5+uuvHYMGDTKPge/3m2++2bFy5Urzu3w9WPia8L6ssv1d1fu+ItZtX3PNNdU+xn79+pnn0dsvv/ziOPbYY8198nUwePBgx8svv1zhbXz88ceux9y+fXvHnXfe6Zg6dWqFrxe+PrifDznkENffJO5r/q3j39jqnh8RERGpvSj+J1ABSBEREREJfZxqywzBd999t8YZaIHGwR8cVsOSVSsDUERERET8Tz0JRURERBoIHhtmOa8de8mxBxzLQKub2CsiIiIikUs9CUVEREQaCA7DYe/Lc8891/RQZA9N9pLjpOZbbrnFDDcRERERkYZJQUIRERGRBiIuLg6TJk0yk3c5KZuZhQwWclAIB5mIiIiISMOlnoQiIiIiIiIiIiINnHoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSioiIiIiIiIiINHAKEoqIiIiIiIiIiDRwChKKiIiIiIiIiIg0cAoSiojfde7cGRdddJHr/C+//IKoqCjz/3Dzj3/8w2x7bfA54HMRquqyXzZs2GB+9/XXXw/ItomIiEj94uc61z0SPP/617/Qu3dvlJaWajfUAtelfB1znRoqDjvsMNx8883B3gwRnylIKOInS5YswWmnnYZOnTohMTER7dq1w7hx4/Dvf/87ZJ/jMWPGmA9S6xQfH48uXbrgiiuuwObNmwN63++88w6eeuopn6/PYBu38dhjj63w31966SXX45g7dy7CGYOL9v1S2ckeiG2IduzYgb///e9mMd2oUSMkJydj2LBhuP/++5GdnR3szRMRkQYoHNeDgWKt3So6HTx4EJFs69atJuC6cOFCn39n//79eOSRR3DLLbcgOtrza3peXh7++c9/YuDAgWbN07RpUxx55JF488034XA4ar2d33zzTb0FhvPz8819+Xpw2jqYbZ0SEhLQqlUr8/3lwQcfRFZWFoKJgVwGdfndie917pt333233PW4P5977jls3749KNspUlOxNf4NESln1qxZGDt2LDp27IjLL78crVu3NkG22bNn4+mnn8Z1110Xss9a+/bt8dBDD5mfCwsLsXz5crzwwgv4/vvvsWLFCrMQqalVq1aVW9xUFCRcunQp/va3v/l8u/wAnjZtmvmQ5XNs97///c/8eyQsOq+88kqPYGhGRgbuvvtuE7zlgtDSrVu3Ot3PUUcdhQMHDpjgcE3xyw9/Ny4uDsEwZ84cTJw4Ebm5uTjvvPNMcJAYIH744YcxY8YMTJ06NSjbJiIiDVM4rwcDZfDgwfi///u/cpfXZu0RbkHCe++91wRK+Rz44tVXX0VxcTHOPvvscgdFjznmGLMuP+uss3Dttdea9e7HH3+MCy+80AT6uA6OiYmp8XbydxnAqo9AIYOEfE6IgT5fXX/99Rg+fDhKSkpMYJDvs3vuuQdPPPEEPvjgAxx99NGu655//vnmOWJAMdDuuOMOs+bke53b9/nnn+Occ84xAU1ug+Wkk05CkyZN8J///Af33XdfwLdLpK4UJBTxgwceeMAc0WPgIjU11ePfdu7cGdLPMbebQRY7HhHjAmTmzJnm6HdNBeqD+fDDDzfP8fvvv4+//vWvrsszMzPx66+/YsqUKWbBFO5GjhxpThYGvhgk5GXe+8r7KDOz6XzFQC4Dq7XBBVBtf7eumCXIfc3F8IIFC0wmoff7kZml/lDT51RERBqucF4PBgozKatau9Qli4sHt4O1FgmE1157DSeeeGK5x8RAIAOEn376qfl3e/DspptuwmOPPYYhQ4aYjLVIxAPkzM61W7RoEY477jiceuqpJsGhTZs25nKuDWsTLK2pLVu24PHHH8c111yDZ5991lx22WWXYfTo0WafnH766a7t4Hqb28+sTwZJa9vGSKS+qNxYxA/WrVuHfv36lVsQUsuWLT3O84OBATge8evVq5dZCDALiplPFX0AXXLJJSa1noE33gePMnrj0UQeAezZs6e5PX5QnnLKKWa7asPK0ouNja22v15FPfu8exJ649HDr7/+Ghs3bnSVEPjSu4+PjY+LWYh2TO1v1qwZxo8fX+Hv/fzzz2aBwWAP9xGP6HGx5e23334zRwJ5P8zSe/HFFyvdlrffftvst6SkJDRv3twcMfSlRHvbtm1YuXIlioqK4I+eK9OnT8fVV19tXmfMCiU+r7yMry9uX1pamlmsePdnqagnIfdN//79zYKL2RDMJOUCn+UU1fUk5D5PSUkxr9uTTz7Z/Jyenm5Kgnn012737t3maC+PrHKfcAHMBZ8vfQ65X3gfPILsHSAkvl/uvPPOansseb9OK3tOP/roI9flFW0L/41ZsRbuXy4G+brga+mQQw7BF1984fF73P9cKPbo0cNch/voiCOOwA8//FDlYxcRkchYDzIgxAwoXs41Xt++ffH8889X+Fl1wgknmM9qfp7wc33AgAGuz+5PPvnEnLfWkzx4Zmd9Nq9fv96sk7gWatu2rclo8qVM1de1aF0OxjHTsEOHDub2uXZh0Mt72+zrZ24Dr/vdd9/5fb3M+x41apT5XOZzzeeU6wBv/Lzm5zb3NZ9fbvftt99u/o37hutJuvjii11r3arWN6waWbx4cbm2OsxCZXUP96M9QGhhNRDXEixTZoVHVT2nvdduvE1mEVrPr3WyX5fPx5NPPmkqSPh8MAhmX/NYa8eKMgPt3x14e1wTkhUoq0sPzEGDBpm2RTxwbAXpKutJWNf3UEWYNci1HNeLFt7vVVddZZIXfv/9d4/rM+mC6/OalJ+LBIsyCUX8gB+c/DDghyYDLNVhsIHZcDwCyMUM08+PP/54/Pnnn67fZ2kBG91aiyJ+sH777be49NJLTc8Sq0yXwRd+8P30008mUMUMu5ycHLN44fZUV5LK39+1a5f5mR92DJ4xhb979+4mcy9Q6fn79u0zH6JceBAXWL5gGj+PHHJBZz02Bg0ZlKmo9PXHH3/EhAkT0LVrV7MQ4QKKfYH42ObPn+9avLCHEG+XzzOvx3IPPg9ccFaUKXDXXXfhjDPOMEcNWfrA22T5LhcWFX05sNx222144403zGLQH0NNuDjhNjPTkAttYgYDSzH4emCQiwslfvHgAo7Bv+pKyPfu3Wtej1w48zFyccyj01xA8bms7vXELyEjRowwC0s+/zzSyn3FhZN19H/y5Mnm9c7LGOjjYouBQl8w4MYFnvdRZX/xfk4nTZpkXp8saeHi2I7vY34Zsd63y5YtM68tBlZvvfVW82WMv8egKbNcmQFJfI1xYc/Xz6GHHmre08wY5WuyNtm7IiISXutBfi7z84OBHx6U/fLLL83nDz8jmZ1kt3btWrP+YTsSZuXx85Wfo2wPw8CUFajg5wo/t73bvvCzmZ/rXFfyoB+Da1zjcK1TVfmjr2vRqnBtaa0zLVyH8MRAIB8/W8nwNlmWy4AYM7EY+LPWiPaDvvxM5ba0aNHCrKP8vV5mWTi36dxzzzWZiu+995450PrVV1+Z9YD1Wc/bYg86Pn9cy3MfsQKH+vTpYy73bhXD4GNluG6joUOHelzO1wVdcMEFFf4eXzt8bTDwxvuvrHd3Rfh6Ylk0n4O33nqrwusw+43PE1+TDLLy+WFwm+vmitbIleF+4Wue6z6uhbjGJD6HtcV1IPcz28twbV6Vur6HvHG9zzUe97Ud13TWvzOIbLHa4nAfMetTJKQ5RKTOpk6d6oiJiTGnkSNHOm6++WbH999/7ygsLCx3Xb7teJo7d67rso0bNzoSExMdU6ZMcV126aWXOtq0aePYtWuXx++fddZZjqZNmzry8/PN+VdffdXc3hNPPFHuvkpLS6vc7tGjR7u2x37q06ePY/369R7XvfDCCx2dOnUqdxv33HOP+R07Xo/Xt0ybNs1ch/+3TJo0qcLbqwyvy98pLi52tG7d2vHPf/7TXL58+XJz29OnT3e89tpr5uc5c+a4fm/w4MGOli1bOnbv3u26bNGiRY7o6GjHBRdc4Lrs5JNPNvuA+8LC2+Y+tT++DRs2mMseeOABj+1bsmSJIzY21uPyip4zXsbby8jI8Pmx8/Hwd/j4LNZjPeKII8xzYme9Nux+//13c/0333yzyv1ivSbs1ysoKDDP+amnnuq6jNvvvU3WY7vvvvs87nvIkCGOYcOGuc5//PHH5npPPfWU67KSkhLH0UcfXe42K9KsWTPHoEGDHL7ibfJ16s37dVrVc3r22Web15H98m3btpnXkf3xHnPMMY4BAwY4Dh486PE+HDVqlKNHjx6uy7j9fD2LiEjDXA9W9Fk9fvx4R9euXct9VvGzadasWa7LeJu8LCkpyWPd8uKLL5b7XLc+m6+77jqPzyV+BsXHxzuysrIq/bz0dS1aGWvbvU/WfXz22Wfm/P333+/xe6eddpojKirKsXbtWo9t42fusmXLPK7r7/Wy92Pivuvfv79Zo1iefPJJc1v2586XtVtV7rzzTnP9nJwcj8u5PuXle/furfR3P/nkE3OdZ555ptL1XWVrt2uuuabcOt5+Xb7GMjMzXZf/8ccf5vIbbrjBY+3IkzfvdTCfr8rWZBWxHseHH35Y6XW4nuK60HstZ19n1/U9VBG+f7zfq5SXl2d+/9Zbby33b3y/XXXVVdU8apHgU7mxiB8w84dHjnnkkSWTPErLbCpmE3mXGRJ7y1lHlIgNrlkCy6OnPNLJtRCzjniEiz/zCKx14u0yC48ZR8Tr8WhqRc2wfel5waOwPILIE4+8MnWft8+MsWBPDasI+3vwCJ81PYxlJyxRsQ/0sJf2Mq2f5Q4s/bTwqCX3GZs1E59zPvfM9uK+sPDooHcJM0sSeJSf22DfLyzRZrkHj4ZXhWUQ3Kf+yCIkNkv27r3CLDv7EXyW9jIzlBmO1uumKsyas/cPYnNxHhllqZIv/vKXv3ic576x/y4zGJj1yW238Gitd+ZEZZgZ0LhxYwRKRc/pmWeeafpJ2Ut3mGHJ1wL/jfbs2WOyHPja4FF367XB55+vozVr1pjMCOK+YCYCLxMRkYa3HrR/VnPdxc8LZqvz85Ln7ViKbO9VzGx9YkaXfd1iXV7R5zWz7CxW1h0z5ZjxX5GarEWrwm2y1pnWycqK4zqMn7esrLFj+THvk+tSOz4/fC5qs42+rpft+4WVFbwNrmPsj9WqGGEVBNcB/sC1ArMCvStruJ6gqtY91r9xfeRvXBvz9WvhepD71FpDBxufL+s5qoo/3kN2rEyqqAe71U/SKv22Y2sk76xakVCkcmMRP2HvEQaQuODiwpDNhVkmwVR4BqrsixoGk7yxPwqnfjEwx4AJe2z897//NaeKWA2wWXbLPij2/oHeeJv2nnD8QLUWIUyVt5cmsByF6fHs2cGJXSwVDTUsF3jmmWfM88xSY5aNVBQQZe8P4vPjjQFABgZZTsrFBT/MK9ov/F37QohBHS5EK7ou1fe0Xw6Z8cbHwnIJ9jtiUMre18f7i0dFWKLs/XxyYcNeOdXh4sjqOWP/XS607fuFfYC8y54ZyPQF+xj6siD053PK9wWb0bO8mBMGiT+zNIrvXauUhc81S9F5qux9y8U2y5B4YIC/y5I03j57NNal7EZERMJnPciyQ5b8MqjI9Z8dP6v5mWOxBzHI+jceJK3ocvtnLnFdybYrdtZnl3e/Yvva0de1aFUYmKusBJbrAfZH9A6AWSWc1jquss/nmmyjL+tlYlnx/fffb/ZVQUGB63L7uogHB19++WXTMoStRbguYPks93FVJaq1YT03XPdU1s7Gl0BibVX2nYVl36EgNzfXp8dd1/eQNwaT7a8PC0uyrX/3xjWihpZIOFCQUMTPmHXFBSJP/BBlw+IPP/zQLAR9ZR2VZDZXZX3aahJM4LbYF1rclqoaBTPLkR+S9mEqlX2oeQ+kqA88ysfeMewzw95+DBrWF+4bPhc8ul3R9DRfeyv6S0WLEB4lZ4CQzw+PmnJfcpsZTPXliHdlU+F8aXBeHxPl2MOQi3d+AeP7rbYqe+1W9JzyaDGPpvPLHnuIsgcSv+A9+OCDrutYzy0HtVQ2RMcKhLJ/Jb+wMAuBvXT4ZYNfItkbh186REQkcteD/PvPwBI/zziEi4EKXp8HJflZ4P1ZXdlna10+r+t7LeoP3p/P/t7GX3/91WSB8jOan/U8oMmDv1xT2YfmcTu4Rmb1CAfxsUKCBw6ZlcbP9NqshTgohT0iGfCzB70YMP3ss8/MgVpuV0Wsg7hWALq+1+y8v4pec4H+jsBqmdWrV/vUD97f7yG+Nrj/vQN/rGIiBr+9MaDNoLlIqFOQUCSAmI1n/8CwVFRiyA85ZlZZWVhcIPDDtboGxAyW/fHHH+aDsrIsNpbk2tPevY8mV4T3zaNz9mwwfrh58z7K66u6Hkk7++yzzZFeLp6YzVVZA3Fi82FvnEDLD2pmUjL7jQu+ivaL9+/y+eaCgEezraPwoYZlsFws27NAeWSzov0XDNwvXFgxc8KeTchMPF+wrIiZFywd4uugOhW9dhlg9H5fVoeZAxw6w6bnHPDD14FVamx/X/F96EvjcJbA80sjT3yvcfHP4L2ChCIikb0e5DAKZiGxBNme4VRdy5LaYjCN5ZP2dQvXnVRZ+xOuR31di9ZlPcByZ+/AGNdo1r9XpSbb6Mt6mesKrglZaWIvJWWQ0BszBhno5YmBXh405GA+7kNuS03XuQwYEw9+2wObHJDC6hAOEKkoSMjHzgAm1zrWwEH+TN5rn4rW7NVtZ2XfWeyvG95fReW53vfn7yw6rnf5/aayA7OBxO8ePMDL9aC9WoyvMevf7VjZw7Wn96ATkVCknoQifmAdSfJmlal6l7sywGHvbbJ582aTUcTpujyixdOpp55qFiucuObN3iuQ12N/i2effbbc9axt4qKBCxbrVF2QkI+HQYtBgwZ5LK5Y/mIvOeVil5lVtcHgnC+lr5VhIIVH46sqh+ZRPn5IM7BjXyjxOeWR3okTJ5rzfL65wOCR2k2bNrmuxw9+LhTtWE7C63OKnPc+53n2lKkKnzMufrlIDRRun/e2cfpyMLI+K8Lnmo//pZde8vgC89xzz/nc85D7lj2LrC853qVFDCDbX7v2rFhiWVJNnw++dxjYY7YAT+zLYy99atmypZkg/eKLL1YYgLS/b71fJ8xAZZZhRaUrIiISWetBK3vJux1IRcEof7GvE3m/PM9gmdVCw1tN1qK1xXUYP4u917DMpmRAif2xq+Lv9TJvj/drXx+wHJvrQzv2IPZmBYWsz3Guc8nXA7RWv7y5c+d6XM6JyFx/8LXBUmhvDExyLXTzzTe7Mi0ZXOVj8V77MDvSW3Xbycdu9VOmP//80wTC7PuG6yyube3PN0vtrWnPFuvAsD8OWvP2WTHDAKWvPa39iS1j+P6xP6d8HbEihG1lvCdZz5s3r9oJ1yKhQpmEIn7A8k5mRU2ZMsUcCeSRolmzZplAAo+0MVPIjmnxDJSwUTOPVFofMAw8WdgPkItNltZykAKPUnFRwuAij7paCxQ2f+bRxRtvvNF8cLO5Mvvs8TpXX321+RCrChelb7/9tvmZZQ7MnHv++efNQoN9ViwsVb3lllvMY+R28/Hyejwq7Uvj6opKmvn8cLtZisMgCTPEfMUFUFUl05ZHH33ULGS4+Lr00kvNEUcGzFiCa/99PvcsF+Hzx+eNzwWv169fP4/AKBdCDEDddtttZuHIElQexeaRXwZMr7jiClNuWhn+HoOWvL6/hpd441Hnt956yzxGvm4YlObrgaUsoYDPGQNsDPIxe5DvGWZTWK/p6o40c0HI55pfLrgoZ5mRNQiIr0UOtbE3p2ZAmYFFfkFgU3kuLBn8rWnJBxeDDBK/99575j322GOPlbsOA53s6TlgwADzvmVAnqXJ3AeZmZnmvon7hQFFbjcDj/xSwCPi9sbyIiISmetBHhRmeTHXPVdeeaU5MMsDZzzYVNMsd18wO45rHFYZcF3Jliksk7399tvL9RG283UtWlt8/GPHjjWBLq6peHCaB3F54JwBIK65quPP9fKkSZNMViD7BLOVDQ868nOdB/Hsa0H2FWYAjtfnepTX41qePZ25BiBuO3sIMmjEdSKDcdzGivoeE9cL/H7A7bnkkks8/o3bzWAut5HbxW1nMJK9LzlQjVUNN910k+v6XP+dfvrpZh3LNRW3hQHGinpIWusnru353YTBRa75LXzsfExXXXWVuU8OOOR6kkFJC7eXzxt/n2tt3g8fN9fQ9mEq/G7B/cP3A78/cP3Dx1xduTDLwFkRw+AtD7Iy+Mh1Ix8n14McHljfuK/5GuX3DB745ncZBlS5razg8i5j5sAeZg0PGTKk3rdVpMaCPV5ZJBJ8++23jksuucTRu3dvR0pKihlx3717d8d1113n2LFjh8d1+ba75pprHG+//bajR48ejoSEBMeQIUMc06ZNK3e7/F1et0OHDo64uDhH69atHcccc4zjv//9r8f18vPzHXfccYejS5curuuddtppjnXr1lW53aNHjzbbY52ioqIczZs3d5x44omOefPmlbv+1KlTHf379zePr1evXuYx3HPPPeZ37Tp16uS48MILXef52Hgd+2PMzc11nHPOOY7U1FTzb/ydqvDfJ02aVOV1XnvtNXNbc+bM8bj8xx9/dBx++OGOpKQkR5MmTRyTJ092LF++vNzvT58+3TFs2DDz+Lp27ep44YUXKnx89PHHHzuOOOIIR3Jysjlx33NfrVq1ynUdPgfej4uX8fYyMjIcvuLj4e/w8VX3WGnv3r2Oiy++2NGiRQvzehw/frxj5cqVPu0Xvib69etX7ja9Hwu333ubeB0+F94qeg6zsrLM/m/cuLGjadOmjosuusgxc+ZMc7333nvPp+dl69atjhtuuMHRs2dPR2JioqNRo0Zm/z3wwAOOffv2ua5XUlLiuOWWW8zzwevw+Vi7dm2556Oq59Tyww8/uN4rmzdvrvA6fN9dcMEF5n3I92O7du0cJ5xwguOjjz5yXef+++93HHrooeb1z9clXz/c7sLCQp8eu4iIhPd68IsvvnAMHDjQfH517tzZ8cgjjzheffXVcmuEytY/1nrSzvpsfvTRR8t9NvOz6bjjjjOfg61atTKfzfx89L5NXl6btWht1245OTnms7xt27bm9rk25vaXlpZW+3gDsV5+5ZVXXOtz7keuDbzXMT/99JPjpJNOMtvMfcz/n3322Y7Vq1d73N/nn3/u6Nu3ryM2NrbcmqkiTzzxhHndcDsrep7+8Y9/mDUa1w1cP3Ft+/rrr5d7rqx11qmnnmr2d7NmzRxXXnmlY+nSpeW2o7i42Lw+09PTzdrGepz219Ljjz9unls+J0ceeaRj0aJF5e6P3wm4dubzMXjwYMf3339f4Tp41qxZrrV2Ra83O2udap24z7idRx11lFkz7dy5s9zvWGs5f76HKsP3z4MPPmhun4+H+4bPQ0XXa9OmjePOO++s9jZFQkEU/1Pz0KKI1BaP6DEtvqJyB5GGjEdgmX3x22+/ufrqiIiISO1ddNFFJlPd3mdaQhOre5hR+K9//ctk5AUTMzuZ9chMuaoqZMS39S0zQDmwiO1yREKdehKKiEi9sw/SIZaQsCymSZMmGDp0qPaIiIiINCgsn2UZLwNz3hOuJXw98sgjpp2MAoQSLtSTUEREgtK3iYFC9g60+uqwbxOnA1qNt0VEREQaEvb/5kkiB/tSi4QTBQlFRKTeHX300WYyNRtpsxk1G2Mzk1CDO0RERERERIJDPQlFREREREREREQaOPUkFBERERERERERaeAaXLkxm8Bu3boVjRs3NlNmRURERCKVw+FATk4O2rZti+johnlsWGs/ERERaSgcdVz7NbggIQOEHTp0CPZmiIiIiNSbzZs3o3379g3yGdfaT0RERBqazbVc+zW4ICEzCK0nrEmTJgE9ap2VlYX09PQGe+Q+kmh/Rhbtz8ii/RlZtD/9a//+/ebgqLX+aYjsa7+UlBStzyKA/k6EP+3DyKD9GP60DyNvP+bm5tZp7dfggoRWiTEDhIEOEnJiJ+9DQcLwp/0ZWbQ/I4v2Z2TR/gyMhtxixb72Y5BQ67Pwp78T4U/7MDJoP4Y/7cPI3Y9RtVz7KcVNRERERERERESkgVOQUEREREREREREpIELapBwxowZmDx5spm6wlTIzz77zOffnTlzJmJjYzF48OCAbqOIiIiIiIiIiEikC2pPwry8PAwaNAiXXHIJTjnlFJ9/Lzs7GxdccAGOOeYY7NixI6DbKCIiEqlKSkpQVFSEUOupwm1iXxX19PVNfHy8nisRERHxea1VWFiodVeYiouLQ0xMTGQGCSdMmGBONfWXv/wF55xzjnliqss+LCgoMCf7lD/rjcFToPC2HQ5HQO9D6o/2Z2TR/ows2p81x88nHmTjQbdQ3ac5OTnB3oywwWBq586dTbDQm9YhIiIiYmFwMCMjw7U+sGIWXHc15CFn4SY1NRWtW7cOyD4Lu+nGr732GtavX4+3334b999/f7XXf+ihh3DvvfeWu5zjoZmlECh8o+3bt8+86ZQJEf60PyOL9mdk0f6sOS4EeQCtZcuWSExMDKlFobVY5WdnKG1XqOLztW3bNrPg54LR+zlTsFVERETsawYmW3Xo0MGstXhZcXGxaeWmdVfo4/7Kz8/Hzp07zfk2bdo07CDhmjVrcOutt+LXX381L2Jf3Hbbbbjxxhs9Mgn5hkhPTzfjoQOFX3D4JuP9KEgY/rQ/I4v2Z2TR/qx5ifGePXvM0ce0tDSEIpYbs5RCfF8wbtmyBc2bNy/3vDEILCIiIsJgIANMnAnRqFEj1xpCQcLwkpSUZP7PQCEP+Pu79Dg2nL7UsMSYWYE9e/b0+fcSEhLMyRsDd4EO3jFI6ECUgoQRgvuzPl43Uj+0PyOL9mfNykz4fCUnJ4fkEWMuVq3tCsXtC0UsMzZrjgqqF/SZJSKhqtThCPYmiDQojKlQRe1JJLw0Kgvy8sB6gw0Sslxm7ty5WLBgAa699lqPPlTMKpw6dSqOPvpohILcg0X4Ys4GzFiaif6dd+HaCf2DvUkiIiIeFICLHNqXIhJO9h8oxD3vzcXqrdm4cExPnHF492BvkkiDonVD+IsK4IH0sAkSsjR4yZIlHpf95z//wc8//4yPPvoIXbp0Qajg/vrfr2tQXOLA/oLtuOb4fnojioiIiIhIg/faz6uwPHOveR7enL4GJx7aBYlxgZvUKSIiYRIkzM3Nxdq1a13n2XR74cKFpqdOx44dTT9B9th58803TblM//6eGXlWw3Xvy4MtOSEOgzu3wNx1WdidU4A12/ahZ9vUYG+WiIiIiIhI0Kzbvg/fzt/kOl9UUorte/PRuWVj7RURkRAQ1OZqLB8eMmSIOREHjPDnu+++25zn5J1Nm9wfIuFkZM+Wrp9nrtwe1G0RERGR8jp37oynnnpKT42ISD1gm6gXpi6HdyfC7dn5ev5FJGzLfj/77DNEkqAGCceMGWM+LLxPr7/+uvl3/v+XX36p9Pf/8Y9/mMzDUHRYz1awqsRnrdoR5K0REREJ7wVYVSeuB2pjzpw5uOKKK+q8luE2PPzww+X+bdKkSeW2j1UTHMTGyYKshmjfvj1OOukkrFy50nWdyh7ne++9h4booYcewvDhw9G4cWNTRXLyySdj1apVwd4sEamh31Zux+KNe8pdvm2vgoQiUr3ff//dDOng+qomdFC4ZjSmNUCapySgW6tk8/OmXbnI3J0bqLsSERGJaKwssE7M/GOfYvtlf//7313X5cHG4uJin243PT3dNR2uLjp06OA6wGlhu5SffvoJbdq0cV3GCXTjxo3Dvn378Mknn5hA1/vvv48BAwYgOzvb4/dfe+01j8fIE4NjDdH06dNxzTXXYPbs2fjhhx/M83jcccchLy8v2JsmIj4qLC7BSz+uqPDflEkoIr545ZVXcN1112HGjBnYunWrnrSGPrgkHA3rnIq1O/Jc2YRnjEoJ9iaJiIiEndatW7t+btq0qcmqsy5jxcHYsWPxzTff4M477zRDzqZOnWoCd2xjwsASg0l9+vQxGWnHHnusx5Hlv/3tb+ZEvN2XXnoJX3/9Nb7//nu0a9cOjz/+OE488cQqt++EE07ABx98gJkzZ+Lwww83l73xxhsmkGVvm7Js2TKsW7fOBA87depkLuP/rd+xS01N9XjcDdl3333ncZ4BWWYUzps3D0cddVS56xcUFJiTZf/+/eb/paWl5sRAMv8v4Uv7Mfx8/Pt67Mg+UOG/bd2Tp/dkmNJ7MTz3l3XKO1iEjJ05KC4pRmxM/YeGurRsjOTEOJ/nWfDAKqtAtm/fbg6m3n777a5///LLL/HPf/7TrANTUlJw5JFHmgOyXCNu3LgRN9xwgzlZzwOrPD7//HMsWLDAdRs8EP3000+bqg/ifd1xxx3mOjxAOXjwYDzxxBMYOnSox7ZZz2d9su7TWtvY34t1XeMoSBhAQzul4v0/tpifZ63ajjNGdQvk3YmIiNTatS//hr257sBKfWiWkoBnLzvCL7d166234rHHHkPXrl3RrFkzbN68GRMnTsQDDzyAhIQEMwRt8uTJJnuPw9Eqc++99+Jf//oXHn30Ufz73//GueeeaxaXHKpWmfj4eHM9LlitgB8DWbwde6kxMxc5iO2jjz4ygUmWzEjNMROTKtsnDAZzP3rLyspCfn6++X0upLkvJDzxC5D2Y/jYm1eId39zD6tsmhSL9s2TsGxLjjmfuSsHO3fuDOIWSm3pvRheGOjiPmPFBU9rt2Xjlv/NCdr2PHLucPTr0Myn67LlSq9evdCtWzecddZZporkpptuMgd4eaD41FNPNWtBZhsWFhaaA4x8jAwsHnLIIbj00kvNiXi5FVSzV59YwbXisstY5cH1HQODvC6DiCx1Xr58uWmBYikpKfG5isVfrMewe/duxMXFebwX61ppoSBhALVumohO6SnYmJWLlZnZ2J1zEGmNEwN5lyIiIrXCAOGunINh++zdd999ppTXwgDSoEGDXOd5dPnTTz/FF198gWuvvbbS27noootw9tlnm58ffPBBPPPMM/jzzz9x/PHHV3n/l1xyiTlqzSPQzHDjQo0ZhvYgITMTeXs333yzCWJx0coj3FyAMrhpx23wDiJyUVpVgLMh4CKYAVYGY/v371/hdW677TaTRWrPJGRmKYO0zC7gFworYCvh+zrQfgwfb325GAXF7syWi4/ujc27cl1Bwl25heY9yX0q4UXvxfBy8OBB5OTkIDY21pyCfbCS98/t8AUPvp533nnm+gzUXX755aaCg72hH3nkERM45FrPMmzYMPN/Vh7wfliJwj7QFq4B+DfHfv/WuiC27DL7upJYbcID0bxfrvFq8zj8hffH7U1LSzM9ru3vRWZd1um2/baVUqFRPVuZICGTT2ev3oFJw5zlRSIiIqGEWX3hfJ8MuNlxgcQAHUuH2c+PR1wPHDjgUf5bkYEDB7p+Tk5ONv0PfclwYUCyR48eJktw2rRpOP/88ytcMLK33gUXXGDKpFkK/eGHH5pgJIOX9sXok08+6VEaTRx20tDx+Vu6dCl+++23Sq/DzFGevHExbX0psH6W8KX9GB5Wbc3Gj4udlVXUvXUTjB/SEV/P3ei6rLC4FNn5RUqmCFN6L4YP6zPQfgomX7eBVSA8YMuDvbw+M+fOPPNMvPrqq+ZgK4fZMmhY1W1535f1c1WX7dixw7Sy4ZqNa0FmDLIigdUq3r9X38+ldZ/29Yy/3osKEgbYyF6t8O7Mda6+hAoSiohIKPJX2W+wMKBnxzIUDrlgCXL37t2RlJSE0047zZSgVIULTzsuuHzt7cJswueee85k/HExWxmWqLD0maf7778f48ePN/+3BwnZj5DbLW7MAP3qq69Mw3J7NoCIhCaWvb3w/XKPy/4yvh+i2Ve2WVK54SWquBKp/56Aj11wmAl+MRuuvgNdvH9fsISYB3vtB0v594UHBJ999lmzxqspBtK8+wiyHNvuwgsvNOW8rBJhD2ne38iRI6tdS4Y7BQkDjEfL0pskImv/QSzM2GWag/ranFNERERqh6UgLB2eMmWKK7Nww4YNAX06zznnHBOcZFZh3759ffodLsh79+6NWbNmBXTbwhkX8ZxmyAwCHs3v0qVLsDdJRHzwy7KtWJ6513V+dN82GNDR2Uu0TarnZPlte/PRr0PlvV9FxP8Yl+jfsbkJwLH6IdiZhRXhtrGvNAfJcSCc3cknn4x3333XVIFwKNzFF19cae9oBkLtWJbLAShcY1iPmxmJ3mvJ//znP6bHNTGDcNeuXYh0ChIGGF9wo3q1xudzNqC41IE5a7Mwpr/KhURERAKJpb+casdsPX4W33XXXQGfnsk+NSxt9s5GtHDxec8995hSZAYRuWidPn26KZe55ZZbPK7LZtlcvHpnIHpnTDaUEuN33nnHTCHkc2A9L+wvVJvsAREJvIOFxXj5p5Wu8/Gx0bjs2D6u8+lNk8Dv5VYiz/a9+dotIlIOKwj27t1rho7wc9+Ow0qYZchhc8ccc4xrqAkDixxmYq2tOnfubKoQ+G/MBmzRooXpZciBZhwyx0oTDjr59ttvTZsZ+1ryrbfeMi1t2N+Yg1IawrpDzVjqwaherVw/c8qxiIiIBBYn0TFoN2rUKBMoZEnv0KFDA/60p6amVhrIY4ksF6ocWjJixAizPSxh4fk77rjD47o8Gt6mTRuPE6ctN0TPP/+8GQTDBb39+eDEQhEJTR/+vh679ruHYZ02sitaNnV/uY6LiUZacrzr/LZsBQlFpDwGAdmj2TtAaAUJ586da4bVsccz+zsPHjwYRx99tEfbFw63YzUJg4jMIKQ+ffqYLEG2iWEFCK/PahDv+2aAkus1HuC9/vrrzSCUSBfl8C7EjnCMAPMFxsWmPUrsb8xWYHNLvoj4BJ/x+I/IPViERvGxeP//jkV8bHAnCUnt96canYc/7c/Iov1Z88l2GRkZpmST09BCDZcloVz2Em77tL7WPaHM/hxwurE+z8Of/u6Htp37DuCy//zimmjconEiXrl6NBLjYz324Y2v/oYV25wTjvt1aIYnLhoVtG2W2tF7MfzXC1p3Rca+LLXFK9hipy5rP2US1oOY6Ggc1tMZcc4vLMaiDbvr425FRERERETq1Ss/rXQFCOmSo3t5BAgt6U3iPQaXiIhI8ClIWE/Yl9DCKcciIiIiIiKRZNnmPWZgiaVPu1SMHdCuwuu2bJzg+nl3TgEKijwHC4iISP1TkLCeDOuWjoRY59P9+6odKCltUFXeIiIiIiISwUodDjz//XKPy/4yvh+iK2kdkd7EHSSkHcomFBEJOgUJ60liXAyGdnU2ydybV4CVW/bW112LiIiIiIgE1I+LM7Fm2z7X+WMGtEPvdqmVXj/dlklIGl4iIhJ8ChLWo1G97VOOVXIsIiIiIiLhL7+gGK/9vMp1PiEuBpcc3bvK32nplUm4fa/6EoqIBJuChPXosB6tEF2WbT9z5XYzSUhERERERCScvTdzLfbkFrjOn3V4N7Ro4jlt3VtKQgyS4mNc57dlHwjoNoqISPUUJKxHTRrFo3/H5ubnbXvzsTErtz7vXkRERERExK/4veaT2Rmu862aJuHUw7pW+3tRUVFok9rI43ZERCS4FCQM6pTj7fV99yIiIiIiIn7z0o8rUFRS6jp/2bF9TLmxL1o3cwcJVW4sIhJ8ChLWs1G91JdQRERERETC36INu00bJQurpo7s406KqE5reyZhdr7aMUlYUzsxiQQKEtazVqmN0L11E/Mzp3/t3KfeGyIiIvVhzJgx+Nvf/qYnW0TED0pKHXhh6nLXebZev+q4vqaM2FdtUpNcPxcUlSA7r1D7RsLSjOXbcO7TP+GmN39HXkFRsDdHauGiiy7CySefHPR14y+//GL+jmZnZyMYFCQMcsnx7yo5FhERqdLkyZNx/PHHV/hvv/76q1lILV68uM7P4uuvv25uq0+fPuX+7cMPPzT/1rlzZ9dlJSUlePjhh9G7d28kJSWhefPmGDFiBF5++WWPBSd/z/tU2eMREQkX3y/cjPU79rvOjx/cAd3bNK3RbdjLja1sQpFwDJj/+5sl2J1TgMUb9+DLORuDvUkRxb6Wio+PR/fu3XHfffehuLg4oPf7ySef4J///GdYBPb8SUHCIFDJsYiIiO8uvfRS/PDDD8jMzCz3b6+99hoOOeQQDBw40C9PaXJyMnbu3Inff//d4/JXXnkFHTt29Ljs3nvvxZNPPmkWkMuXL8e0adNwxRVXlFsgMiC4bds2j9O7777rl+0VEQmG3INFeH3aKtf5RvGxuGhsrxrfjr3cmNSXUMJRxo792H/AnT24InNvULcnEllrqTVr1uD//u//8I9//AOPPvpouesVFvovG7l58+Zo3LgxGhoFCYOgc8vGaFN21IxHGvYfUFq9iIhIZU444QSkp6ebTD+73Nxck+HHIOLu3btx9tlno127dmjUqBEGDBhQq0BcbGwszjnnHLz66quuyxic5BFiXm73xRdf4Oqrr8bpp5+OLl26YNCgQWZb/v73v3tcLyEhAa1bt/Y4NWvWTDtcRMLW/35dg3357u8w5xzZHc1SEmp8Oy2bJpoyZYsmHEs4WrZ5j8f5Ndv3BW1bIpW1lurUqROuuuoqHHvssWYdZpUIP/DAA2jbti169XIerNi8eTPOOOMMpKammmDfSSedhA0bNnhUg9x4443m39PS0nDzzTeX6yk5xqvcuKCgALfccgs6dOhgtocZjTyIzNsdO3asuQ7Xd8wo5HZRaWkpHnroIbNOZNUJ14offfSRx/1888036Nmzp/l33o59O4MhNqj33kDxRcNswo9nZ6DU4cAfq3di3KD2wd4sERFp6OY+Acx7ovrrtRwKTPnC87JPTwR2zq/+d4fdCBxyY40DdxdccIEJEt5xxx2uflcMEHKRx+AgA4bDhg0zi7cmTZrg66+/xvnnn49u3brh0EMPrdH9XXLJJWZh+PTTT5uAI++XR7BbtXIPHyMuVn/++WcTKGQQU0SkIcjcnYvP/3R/iWXyw0mHulsx1ER8bAxaNElE1v6D5rzKjSUcLdnkmTnIsuM9uQfRPCUR4SB6/lPAwqdDbv1XFQbUeICYfvrpJ7P2Y9UJFRUVYfz48Rg5cqRpS8N15P3332/WcmxPw5Llxx9/3KzveFCYbWZ4/tNPP8XRRx9d6X1yLcpKk2eeecYE+zIyMrBr1y4TNPz4449x6qmnYtWqVWZbuH3EAOHbb7+NF154AT169MCMGTNw3nnnmXXj6NGjTTDzlFNOwTXXXGOqUebOnWsyJYNJQcIg9iVkkNDqS6ggoYiIBF3hfiB3S/XXa9yh/GUHsnz7Xd5HLTBwx7KS6dOnmwCeVWrMBVnTpk3NyZ7Bd9111+H777/HBx98UOMg4ZAhQ9C1a1dzpJeBRi4in3jiCaxfv97jerzstNNOM8HCfv36YdSoUeZI9YQJEzyu99VXXyElJcXjsttvv92cRETCzX9/WGF6sFmuGNfHBPtqi0FGK0i4Qz0JJcww+8w7k5DWbd+P5t3DI0jItVlUiK7/Knq+GRTkGo9rvaysLNMqhv2gGfwjBuWYwcfLrAPLXDMya5CVIccddxyeeuop3HbbbSZARwzi8TYrs3r1arOmZCCSWYzEtaKF2YrUsmVLcz9W5uGDDz6IH3/80QQsrd/57bff8OKLL5og4fPPP28OaDNIScyEXLJkCR555BEEi4KEQdKnfTM0bRRv0vTnrsvCwaISJMbV/sNVRESkzuKbACntqr9eUnrFl/nyu7yPWuBwEAbheMSXQcK1a9eao8NsXE3MKORCjAu4LVu2mJ40XJwxE7C2QUkuKNmHMC8vDxMnTsSzzz7rcZ2+ffti6dKlmDdvHmbOnGmODnPICktM7MNLWDrCRaCdtZgUEQkn/N7yx5qdrvODu6RhZE/PLOva9CVkCyZSubGEG75m9+QWlLt8zbZ9GN69JcJCfBM4Utp5lP6HyvrP+4ArswQZAGQLGPYlZAYeW8xYAUJatGiRWSd69xM8ePAg1q1bh3379pn+hhw2Z2G2IXtcO7xKji0LFy5ETEyMCez5ituQn5+PcePGeVzONSoPSNOKFSs8toOsgGKwKEgYJDHRURjZqxW+W7AZBcWlmL8+y2PqsYiISL07pA6lIN7lJwHAfn88avzcc8+ZAB6PvFqLNWYZsjyYR4a5WORRZfaRqW0D63PPPdf0p+EClNmEXDxWJDo6GsOHDzcn3h+PXvP6LItm/xnitrBvjYhIOCsuKcWLU5e7zkdHAVcd18+VqVNb9uElu/YfRGFxSZ0yE0Xq09IKsghp7bbw6UtYOvRviD707+yLFpLrP/sBVwYD2XvQvi7jOsvOakHzv//9r9zt1LY9TFJZ+XBNcDuILXDYM9uOPQ1DlQaXBJGmHIuIiPiODagZlHvnnXfw5ptvmmw/68spM/lY6ss+L+wTw3IOlobUFjP9TjzxRFPezPvxFbMLidmHIiKR5Ot5G7Fpl/NLL00a1skMZKwra6AjMYdn574Ddb5NkfqydFMlQcLt/imvFXgccGWFR2UHbi1Dhw41U5BZ+svfsZ+sFjVt2rTBH3/84fqd4uJiUxlSGR6AZgYj14UVsTIZWdliXxMyGLhp06Zy28E+hsR+iH/++afHbc2ePRvBpCBhEA3p0gJJ8c6jZH+s3oGS0tJgbo6IiEhIY5nJmWeeaXrIsEzEmhxHbAbNPjGzZs0ypRtXXnklduzYUaf7Yy9CNqRmqXNF2I/wySefNIvMjRs3mj43LHvhhDr777Dsefv27R4n3q6ISLjYn1+IN6evcZ1PSYzF+aN7+uW2W9uChKSSYwkny7yGllh27Dtg3jdS/1gN0qJFC3PwmK1pOGCEa7Trr78emZmZ5jp//etf8fDDD+Ozzz7DypUrzRC67OzsSm+zc+fOuPDCC82BY/6OdZtsc0OcuswD1yyLZp9EZhGy3Jn9sm+44Qa88cYbptR5/vz5+Pe//23O01/+8hcT0LzpppvM0BMeCOf6M5gUJAwiptEf0s3Zp2D/gSIsreQPjIiIiLhLjvfu3Wum1rHcxHLnnXeaI8e8nD0LOUzk5JNPrtPTxtKStLS0Sv+d9/Xll1+aPoQMDHLxyODg1KlTPY5yf/fdd+aItf10xBFHaJeK1AKzzJ79dhk+mbvVlKVK/XhrxmrkHixynT/vqJ6mv7o/tLGVG9N2DS+RMLE3twCZe9yVAwM6evYbXrM9fEqOIwn7UbNPNLMOOZiE2XpcP7InIScPEycIsz0M127sAciA3pQpU6q8XZY78wAxA4pc711++eWuyhGWE99777249dZb0apVK1x77bXm8n/+85+46667zJRjbgcnLLP82GpJw23kZGQGHlkJwwEq7LEdTFGOyjozRqj9+/eb9FI2q7ReIIHAVNSdO3eaFFeWRlXm5yVb8MhnC83PJx/aGVeN7xewbZLA708JD9qfkUX7s2a4QOLRTy5OEhNDb+oelyUs+WCQra59rhqKqvZpfa17Qpn9OWA2qj7Pw9v9H83Dryu2m58vHtsTZx3RI9ibFPE27MzBVf/9FaVlXxs7pCXjhSuPQmxMtF8+u/m3/qRHvkdBkTPoe+phXXDFOGfrBgltDX0N9uuKbbj/o/mu83eeNtTj/CVH98aZh3dDKK8XtO4KTwe99qX9vcgsxrqs/RreOznEjOjRErHs+gtg1qodlU7TERERERFp6OyVN/PXq2w/0Pjd5MUflrsChHTlcX3rFCD0xiChPZtw+958v922SCAt2+z+e8Tv9Id2b4lWTd0DLtYqk1DCkIKEQZacGIdBndNc5RPr1OBURERERKScopJS7M0rcJ3fkOUeoiGB8ceanR7B2EO7p2N4d2e7JH+y9yXclq3BJRJ+Q0t6tG2KhLgYdG/tztxaE0YTjkUsChKGgJG9Wrt+nrnKWT4hIiIiIiJue3IOejwd+/ILTU8wCVxQ9r8/rHCdj4mOClgZsH3CMTMJVV0loS6/oBjrbJmC/Ts4+xF2b9PUYwiPvZenSDhQkDAEjOrVyvXz76vqNolRRERERCQS7fIKElLGzpygbEtD8PmfG7DFNpThxOGd0aFFSkDuq02qu0Qzv7DYDHUUCWUrtuxFqa1TWP+yoSU9bEFCUqWghBsFCUNAWuNE9G6X6lrobLV9GIuIiAQCGxxLZFDGjTQUu/aXDxJu2Lk/KNsS6Q4WleCdX9e4zjdJisO5RwZuSIy93NjKwBIJl1Jj6tuhmfl/99aeQcJQ7EuodUP4Kw3gOj42YLcsNTKqV2us3JLtGmBy2siuegZFRMTv4uPjzQTCrVu3Ij093ZwPpSnCmrJX8+crKyvL7MO4uLgA7RWR0LBbmYT1ZsnG3cgrKHadP390TzROCtzfGPvgEqvk2EqiEAn1oSWd0lPQJCne/NwsJQFpjROwO6cg5PoScp3A9QLXDVwD8metu8KLw+FAYWGh2Ydcz3Md728KEoZQyfGrP680P89atV1BQhERCQguKLp06YJt27aZQGEoLn54dJTbGUrBy1DG56l9+/aIiYkJ9qaI1Hu58QaVGwfEPNuwkuioKBw9oB0CqZVXkHBbtjIJJbT7da7M3Fuu1NjCbMLdOTvNz2tDKEjIdQLXC5mZmdiwYYO5TOuu8NSoUSN07NjRrJf9TUHCEMH+Hh3SkrF5dx6Wb96L7LwCpCYnBHuzREQkAvGoIxcWxcXFKCkpQShhgHD37t1IS0sLyMInEjEzQAFCaajlxhuzclDqcJhAlvjPvHVZrp+Z0ZeSGNhMZU6FtWdfMZNQJFQx8FdQXFpuaImFfQk5GZwyd+fhQGExkuJDI/SSkpKCHj16oKjI2fdT667wExMTg9jY2IAdTA+NV6oYo3q3xvsz14H9T39fvQMThnTUMyMiIgFhlaeGWokqF6vcpsTERAUJRaTaTEJ+UWf/unbNk/Vs+UnW/gPYtCvXdX5Y1xb18ty2Tm3kChIqk1BC2dLNnv0IK8oktDjKhpd4XyfYQSbr4KLWXeJNh+hDrC+hhX0JRURERESk8p6EpJJj/5pvKzWmYd3S6+Ul2MY2vESZhBLKlm5ylxq3bJpkTnbd2zQJ+eElIpVRkDCE9Gzb1KTZ04L1u5BvaxYsIiIiItJQsW+WlWXmLUN9CQNWapySGGu+o9QH+/ASZjOy75tIqGF7g+W2TMJ+ZVON7Vo0TkRqsnugxNptmsIu4UNBwhDCXipWNiE/FOfaPqBFREREJHRtz87HU18trrBvntTdvvzCSoNGG3bqC7i/lJQ6MD/DnUk4uHMLxNRTf9jWtkzCUgewc9+BerlfkZrYvCsX+w84+/lRRWXEbOliLzkOpQnHItVRkDDEjOzVyvUzpxyLiIiISOg6WFSCZ79dikuf+wXfLtiMd39bE+xNikjewdf4GHfDdpUb+w/LInNsAZD6KjX2LjcmlRxLKFq6yasfodfQEkv31u6S4027csxnhUg4UJAwxAzqlIbkBOc8mT/X7FSavYiIiEgIi4+NxorMvShm6hOA7xZsVnCjHoaW9Gnb2PXzlj35KCzWF3B/lxrX59ASa3CJnYaXSChattndj5BTvzump1R4ve5t3JmE/HhQxrOEi6AGCWfMmIHJkyejbdu2JiX3s88+q/L6v/32Gw4//HCkpaUhKSkJvXv3xpNPPolIEhsTjRE9Wpqf8wqKsXjD7mBvkoiIiIhU0S7mgjE9XecZLHz7V2UTBnpoyYAO9i/gDmzKck/jldqbZxta0j4tGa28AneB1DwlwQTdLcoklFDPJOzfoZn5DKhID1uQkNaoL6GEiaAGCfPy8jBo0CA899xzPl0/OTkZ1157rQkurlixAnfeeac5/fe//0XkTjlWybGIiIhIKDu0e0v0aZ/qOv/T4kzTt0oCU24cGx2FPm3cmYSk4SV1l1dQZLJiLcO61l+pMTFpxJ5NuG1vfr3ev0h12Cdzh61XZr8K+hFaWjVNMpmGlrXqSyhhIqhBwgkTJuD+++/HlClTfLr+kCFDcPbZZ6Nfv37o3LkzzjvvPIwfPx6//vorIskh3dMRF+PcNbNW7TBHR0VEREQkNDG4cdGYXh6lZW9NXx3UbYrkcuPmjRPROjURsfa+hFk5QdqyyLFow24zuMQyrFv9lRpX1JeQw4BEQsky21TjyoaWeAwvadPEo9+nSDhwNr8LUwsWLMCsWbNMoLEyBQUF5mTZv985/ay0tNScAoW37XA4anUfCbHRGNIlDX+uzcKe3AKszNyL3u3cR6el/tVlf0ro0f6MLNqfkUX70//Pp9SPwV1aYHDnNCwsaxUzffk2nHXEfnRt5f6SKP4pN27ROMFkE3ZMS8H6nc7goDIJ626+rdSYz+/ATmkIZpBw6958s/5msEUk1EqNWRrvXVLsrUfrpliYsds1YIm9U+NjYwK+nSINLkjYvn17ZGVlobi4GP/4xz9w2WWXVXrdhx56CPfee2+5y/n7Bw969jbx96J837595oMtOrrmCZv92zTCn2udP/+4IAPN49r5fyOl3vanhBbtz8ii/RlZtD/9KydH2VX16cKxvbDwtVmu82/8shr3nnlIvW5DpMqylRunNU40/+/csrErSLix7P9Se3NtQ0v6dmiGpPj6/6poLzfOLyhGzsEiNEmKr/ftEKluaAmTeKzqv8rYh5ewX+3GrNxqA4siwRaWQUKWF+fm5mL27Nm49dZb0b17d1OGXJHbbrsNN954o0cmYYcOHZCeno4mTZoE9EsOj3rxfmoTVBqX3BSv/7bRlKssyszBtSc4h5lIcNR1f0po0f6MLNqfkUX7078SE53BFKkffds3w6E9WuLPNTvN+dmrd2DllmxVhPg9k9AdJLSXI+ccKELjJHcPMPHd1j15Hj0AD+lWv/0IK8oktIaXKEgooYB/X5gNaOnfofJSY3smod2abfsUJJSQF5ZBwi5dupj/DxgwADt27DDZhJUFCRMSEszJGwM9gQ72MKhU2/tp3jgJfTs0NynNmbvzkLknHx1bVDxeXepHXfanhB7tz8ii/RlZtD/9R59Z9e/C0T1dQUJ685dVePDcEUHYkshxoLAYeQXF5TMJ0z2Hl2zYuR8DglAiGwnmZ7hLjWloPQ8tqSiTkLZnH0DPtmq7JKHRj9DhYz9CS5vmjdAoIdZkxZL6Eko4iI6EjAN7z8FIcnivVq6fZ63UlGMRERGRUMfysiP7tHadn7d+F5ZsdPakkrpPNqYWTaxMQs8D6OpLWHvzbKXGTRvFo1vr4PTSbJ2a5HFeE44lFEuNo6OA3raJ9pWJ5vAS23uJmYQioS6oQUKWDC9cuNCcKCMjw/y8adMmV6nwBRdc4Lr+c889hy+//BJr1qwxp1deeQWPPfaYmXIciUb2ci8wOeVYRERERELf+aN7wj5q4fVfVpu+xlL3UmNKa5zgKjtOTnAXRilIWDvFJaWugTs0tGsLE9wIhsT4WDRLdleBacKxhOLQEg6kSk7wrbVBd1vJccaOHPN+EwllQQ0Szp07F0OGDDEnYu9A/nz33Xeb89u2bXMFDK2sQQYOBw8ejEMOOcQEDR955BHcd999iETsydGlrNfKqq3Z5Y6iioiIiEjo6ZTeGEcPaOfx5ZIZhVI77DdoZ/UkZGsCe1/CjVkaXlIb7JtplUPSsCCVGltaN3NnEyqTUEJBQVEJVm/NrlGpscWeSVhUUopNu3L9vn0iEdOTcMyYMVUeVX399dc9zl933XXm1JAc3ru166jo76u3Y/IhnYO9SSIiIiJSjfOO6oFflm1FCafQcdLxtFUY1rWFCWxJzXgfKG+ekoDsPc4v2jygbpUBcqgAv1voOa6ZeevdpcZWJmEwtUlthBWZzoCMMgklFDBAyOnENRlaYvGeZsySY2YiioSqsO9JGOlG2fsSquRYREREJCy0bZ6M8YM7uM6v3rYPv2stV+dMQvbLi4+NcZ3v3NL9ZZvDTbJUeVNj89a5s1wZdLUGwwRLa9uE4x3ZB1BSqvJMCa4ltlJj6texmc+/2y4tBQlx7r9ZGl4ioU5BwhDHowytmjpT7hdt2I3cg0XB3iQRERER8cHZR3RHXIx7uf3GL6tRqt6EdepJaJUaW6zWPJaMnfv12qyB/QcKPcooh3ULbqmx1XLJwvdL1j61XJLQGVrStnkjNE/xPZAeEx2FbrbMwbXb9DdKQpuChCGO5RKjejsHmLBc5c81O4O9SSIiIiL1ZsaMGZg8eTLatm1r1kWfffZZ2Dz7LZsmYdKwjq7zG7JyMH3Z1qBuU7iXG1uTje39H+1Yciy+W5ixG44QKjW2yo3ttmXnB21bRPgdfHnm3lqVGldUcrxux35XGwqRUKQgYdiVHG8P6raIiIiI1Ke8vDwMGjTIDKwLR2cd3t2j1Oyt6WtUPlmHcmPvUtjGSXEegUNNOK6Zeevc/QjjY6MxoAYDGeqj3Jg0vESCKWPHfo/BPjUZWmLp3qaJxxCUzN0aXiKhS0HCMNCvQzPTf4XmrM1CYXFJsDdJREREpF5MmDAB999/P6ZMmRKWz3izlAScPNw9eG7Lnjz8uHhLULcpnBSXlGJvbkGl5cbeJcfKJPQdh7zYh5YM6JTm0e8xWBgItpfpb9+rTEIJnmWbPfsR1iaTsHtrz+Ela7ftq/N2iUTkdGPxTUx0NEb0aImpizJxsKgE89fvwmE93dmFIiIiIuJUUFBgTpb9+539n0pLS82JgRH+vz6delhnfDlvoysb5e3pqzG6b+uQCMiEut05BzzKYdMaJ5Tbj53SU8yBdNq8KxeFRcWItQWZpGKbduV6DHoZ2iWt3t4b1b0XW6UmIXN3nvl52968en/Pim+C9Tc1WENLmiXHo3VqYo0fb4e0RiZTt7C41DXheGz/tggFDWEfNgSltv1Y132pIGGYGNWrtQkSEifjKUgoIiIiUt5DDz2Ee++9t9zlWVlZyM/Px759+8xCOjq6foNI4/un49N528zPO/cfxEe/rsCx/VrW6zaEo7U7PMvyYkoOYufOnR77sXmCO4xYXOrA4jWb0b65c/CfVG7G0h0e5zs3jTbPbX3gl9iq3ovNG8Ugc7fz581Z++ttu8S/+zHc8XEt3uCe/t29ZSPzWVIb7ZslYn2WMyt2+aZdIfOajvR9WBOlpQ7MWL0LWTmFGNevJVIbxSFc2Pcj27TUhYKEYYJNhNnPhj0Mfl+9A9eXOsykJBERERFxu+2223DjjTd6ZBJ26NAB6enpSElJMcNP+HN9fxk6d2wz/LR8F/YfKDLnv1q0E6cc0QeJtn6FUt7qPZ4ZEd07tELLtGSP/TjIkQj8ssF1nf0lcWjZUgHY6qzauckjQ3NIr47mea2vL7RVvRc7tczC4s3OLOBdeUXanyGquv0Y7rbuzcO+A+5+hEO7t6n1a7F3h52uIOGmPQfQgs9ZPb3fGvI+rImfl2zBa786/y7+mbEPT140Es1TEhAO7PsxN7duPS8VJAwTDBAe0i0dM1dux778QjNhKRQaC4uIiIiEkoSEBHPyxi8/PHERbf1cnxonJeCMUd3w8k8rzfm9eQX4ev4mnD6yW71uR7jZneMuHaeWTRuV248d0xubL9ulDmdG4cas3Ab/Zbc67HG+ZGNZqh6AYV3TERNTvwHrqt6LbZonu37OOVCE/MISpCSGT1ZPQxKsv6n1YXlmtsd59u2s7ePs0SaVebHm5wOFJdiRfRDt0tyv82CK5H1YEws2uP8m7tx3APe8PxePXTgSSfHhETbz135s2K+CMKMpxyIiIiLha/Lwzh5ZCR/MXOcxNVPK222bbMysy0YJ5b+ssbdje9uX7Q1ZmhxaneWb96KgrD+aFSQMJW1SPScca3iJBMOyTXtdPyfFx6BrK/eQpJrq0cZzeAn7Ekro9Wm1W7t9Px78ZAFKGli/RgUJw8iIHq1cKcmzVm439eYiIiIikYxlMwsXLjQnysjIMD9v2uQulQwXDHKddUR313mWHn/6R0ZQtynU7bIFCTnZuLJy2M4eE46dZapSubnr3H3V+IwO6doipJ6u1s08g4TbsjXhWOrfUtvQkr7tm5mBorXFAUuxtnZha7crSBhKGFvJ3FW+l9+fa3biue+WNajYi4KEYaRxUhwGdnaWGG/PPoCMnTnB3iQRERGRgJo7dy6GDBliTsR+g/z57rvvDstnfsKQDkhvkug6/9Hs9dh/oDCo2xTKdtmm77awPW/eutiChFwnK0OzavPX7/LIcGraKB6hpLUyCSXI9uYWIHOPO2jUv46tvpjxbD+YsUZBwpA7IJVfWHFm/9fzNuGDWevRUChIGGYO79Xa9fOsVZ4TyUREREQizZgxY8wRfO/T66+/jnDEL4rnHtXDdZ7BrI9+bzhfPuqSSZjWuPIgof3LN23M0sH0qoIf63bs9xiQGGpYVm4PXCqTUOrbss3uLELq16Hu8wC620qO127b36Cy08Kt1Pj0kV09Bsu8+vNKTFu6BQ2BgoRh5rCerVw/s+RYRERERMLLuIHt0ba5u5zysz83IDvPc0CHOMu/dnuVG1emS8smHudVcVO5+evdpcbE4YihqI2t5Fg9CaW+Ld3s7kfIMuHe7Th4pG66t3YHCXMPFmFH9oE636b4x2avIOGph3XFdRP7e1z2+BeLsdg28ClSKUgYZlo2TULPsiMQPAK4Xf05RERERMJKbEw0zj+qp+t8QVEJ3pu5LqjbFIo41bbQNlyjqnLjVqlJpuejZYPa8lRqnq3UmMMYerdvhlBkLzlWJqEEsx9hj7ZNkWD7+1JbPdp4HsxQyXFoZhKmJMYhNTkeE4d2xJmHd3NdXlRSins/mItNEZ6priBhGBrVWyXHIiIiIuFsdL+2ppG95au5G5G1X1kllZUaV5dJyLKwTum24SUR/iWutkodDo9+hIM6t0BcTHTIZxIy46qkVKWZUj/YBmKdrWdgfz+UGlsZz/YS1rWacBySmYQdW6S4hmRdNLYXxvZv6/q33IPFuPPdOdiT6/n5FElC8xNBqjSql7vk+PdVKjkWERERCTcx0VG4YHRPjwyFd35dG9RtCjX2UmNKqyKT0Ht4CTMJ1e+rvIwdOdhrK20fFoL9CCsKEjJAuEtBdKknK7bshT0mXdehJRZmI9oPDq3ZrknsoZhJ2LGFex8xqHvj5IEY2Mn9Gtix7wDuencODlQy6CTcKUgYhviibdc82ZUGvS9fE/FEREREws3hvVuje2t3+dn3Czdj2978oG5TKMmyTTauLpPQe3gJ18f2YJg4zfPqRzisa2j2I6xowrFKjiUYpcbU148l+fa+hMwk1MGM4Nt/oBDZee6YSgdbkNAaOHb36Yd4BA/Xbt+PBz+ej5JSd0uMSKEgYRhi6quVTcgjHLNXa8qxiIiISDiu6VjKZM+WenvG6qBuU6hmEjKbIzU5wedMQtLwkqqDhK1TkzwG6IRyJiFpeInUl2W2oSXM/Gtim7RdV91tfQl5MMO7rYIEf2hJR68gITVOisP9Zw9H8xT359Cfa7Pw7LfLIi7QqyBhmFJfQhEREZHwx8my/Tq4s1R+XrIl4pui+8r+5bl54wRTou1rJiFpeImng4XFWLbJHfwY2jXd1XcrFKU1TjRTZS3KspX6wNYPKzP3+r3U2NKjbAipZY36EoZUqXFlQUJqldoI95013GNI1jfzN+GDWZE1eExBwjDFEexWFHv++izzoS8iIiIi4YVBmgvHuLMJWSXy1ow1Qd2mULHLVm6cXk2pMTHTkBMpLQoSelqyaY8JgNgD1KGMQWF+Kbdsz9ZgHwk8DiwpsE1V99fQEkvXVk1gD82v3aa+hKGUSZgQG42WqUlVBnnvPG2oxwCaV39eZQ7wRQoFCcMUX5SH9XSWHBcWl2LuOs/+IiIiIiISHgZ1TsOQLu4BEjOWb/OYrNlQ2cuNmVXmC3s2oYKEnubZphrzu8TgzmkIda1tJcfKJJT6Cqbb+TuTMCk+Fu3TnPMFaK3+1odUkLB9WopHALAiw7u3xHUT+3tc9sSXi7F4425EAgUJI2TK8axV6ksoIiIiEq4uGuuedExv/qLehPZy4xbVTDa2dGnp7ve1MSvH9HkUp3m2pII+7VORnBgX8k8N+yZatmdrqI8E3lJbSX7Lpknm5G/2kmMFCUOr3LhDJaXG3iYO7YizDu/mOs8s7Xs/mGs+d8KdgoRhbHCXFmiUEGt+/mPNThTbygdEREREJHz0btcMh/Vo6To/e81OrLD1xWpoCopKkHOgyOfJxhUNL2HJoIZdOO3cd8DjizD7EYaDNrZyYw55yC9QiyUJnFKHA8s3uzMJ7f1i/am7LUi4O6cAe3I1vCSYnzU7bK0MOvgYJCQOHhvbv63rfO7BYtz17hyPLPhwpCBhGIuLicah3Z2LydyDReVSo0VEREQkfFwwxjOb8I0GnE3oPfGzNuXGlLFT/b6sHuZ2w7q6y9vDpdyYlE0ogS473W87OOHvUmNL99aew0vUlzB4MnfnwuHD0JLKegrfOHkgBnZyv0527DuAu9+bgwNhPDNCQcIwN9Kj5Hh7ULdFRERERGqvW+umOKpvG9f5BRm7sGhDZPQ4qsvQkpqUG3dqkeIxFGBDlufUyobK3o8wJTEWPdumItwyCUmZoRJIS737Efp5aImle2t3WwRSyXHoTzauTHxsDO4+/RCP31u7fT8e/Hg+SkrDs9JTQcIwN7x7uskotPoSOhzquyIiIiISrs4f3RPRtijXG7+sapDrO+9yLV/LjRPjYz2yzzYok9D0ZZxvCxJySA4nB4eDNl6ZhNvUl1ACaNlmd4uHlMQ4dEyvWcDIV+wH2ra5+7W9ZpsGVYVCkJB/Ftva9ouvGifF4f6zh6N5SoLrsj/XZuHf3ywNy89vBQnDXHJCHAZ3SXMdcdUfGBEREZHwxWyEYwa09/jSOtc2cKKhqG25sXdfwoyd4d9Evq74/YCtiSzDuoVHP0IrmMIv4BZNOJb6yiTs36FZtVNu/VVyzMwzCf5k4zbNkk1mYG20Sm2Ef541HIlx7t//dsFmvD9zHcKNgoQRYFSv1q6fNeVYREREJLyde1QPj0yv16c1vGxCe7kxg0QJti9eNelLuHVPnmlM35B59yMc2iU8+hFWVHKsnoQSyOE+7Cdn6RegfoQVBQl53xzMI+Ex2biqgTR3njbUI7j82rRV+HnJFoQTBQkjwGE9W7p6r8xcqb6EIiIiIuGMJZbHD+ngkWXS0NZ49kxCX0uNLV1auvt9lTrK95xqaOyZqO3Tkk3GSzixl48rk1ACZZltqnEgh5ZYetgmHNNalRzXO/YM3LI7r9b9CCsyvHtLXDexv8dlj3+xCAs3uFs+hDoFCSNA85RE9GnvHM/ORRAn9IiIiIhI+Dr7iO6uvtP05vTVprdcQ+xJ6OvQEktnrz5iGxpwyXFeQRFWZGa7zh8SRqXGFWUS7sg+gNIGllUr9V9qHB8bXS6I528aXhJ8POhQbPtc9UeQkCYO7Wg+wy28j/s+mBc2n0UKEkaIUb3dU45/X7UjqNsiIiIiInWT3iQJkw/p5Dq/MSsX05dtbZDlxjXpR0jt0pI9AqwbssLji1kgLMrY7RFUG9o1vEqNvTMJi0pKyw21EfH30JLe7VI9/oZUie+vYttrsrgAmH4TsH1Olb/WpFE8WjVNcp1fs019Ceubd5Z5XcuN7S4c0xPHDGjnOp9XUIy73psTFn+/FCSMEKN6uvsSzlzVsMpRRERERCLRmYd382iC/vHs9Q2iNyEzJvfkFrjOp9cwSBgTHe2REdKQh5fMs/UjjI2OwqBOzoGH4cR7wvH2vflB2xaJTDkHijyyvPp3qEGp8ZKXgLeGAjsXAblbgXcOBeY+BnxzHlDkLmWtrIedZe12TTgO5tAS6tAi2W+3HRUVhRsmD8SgzmkevSfvfm8ODhQWI5QpSBgheMS0U1lpxcrM7LCIUIuIiIhI5VKTEzBxWEeP3oRLbCVxkSo7r8Aj+y2thuXG3sNLNuxsuBk689bv8hjEkBgfi3AuN6Zt2QoSiv/7ETpq049w93Jg2t+APSuAd0YARflATILz3/auBqb/3eeSY5a+2qeQS/1mErL3bXKCe5K6PzAb9e7Th3kctOLn+AMfzzf9EEOVgoQR5PCyKcf8Azd7tUqORURERMLdScM7wzboGB/PzkCky7KVGtdmcAl1sQUJd+cUYP+Bhjc5lJOd7YM+hnUNv36ElN400WNaaCQOL5m/fhfemr4aW/ZUnXkmgS815t/b3u1Tq/8llhh/dRZQXDYRuf8lQLPuwIS3gdiywPaiF4B1X/k+vETZhGE72bgyKYlxeOCcQ9E8JcHjb1goT7NWkDCCjOrtLjmepb6EIiIiImGvdWojHG5b4/2xekfEBxK8K2JqEyS0ZxJSuDSMD1QWIQ0Lw36EVvl4q9SkiC03XrklG3e88yfenrHG/L+4JHQzjBrC0JKurZr4llE242Zg1xLnz2n9gNGPO39u3hMY+6T7et9fAuRVnMDTvbX3hOOGm/Vc39i6I3OXfycbV6Zl0yT886zhSIqPMVmqT148ygyfDVUKEkYQpiunl5VjLMzYZaaZiYiIiEh4mzKii+tnVox89qefsgk3/gh8PAH442FnmVyI2OUVJKxruXGDDRKuc/cjbNooHl1tpY3hGCyP1HLjL+ZscJXXM8OoIffQDIaCohKs3ppds1LjdV8CC/7t/Dk2ETjhPSDOHcjGgMuBrpOdPx/IAqZe5hxw4qVZSoLHQZA129SXsD4/Z/JtvQEDlUlo7z/56AUj8dC5h6JJUjxCmYKEEYTNMUeVlRxzzPacNe6FgYiIiIiEp77tm5lpm5apCzNNo/06yd8JfHEKsOE74LfbgFd7AEteBUpLEEqTjeNjo9E4seZ9ovjFOyXR3X+voQVemI22aMNujyxCe8luOA8v2b63rLwzAjCp47cV2zwuW6eS03rFACG/O/s8tITDSb672H2eGYQt+nteh++18S8DjVo6z6//Clj832r7EqrcOHiTjQOZSWgvL4+PdQ8jC1UKEkaYUb1auX6epSnHIiIiIhFxINieTXiwqATfLthUtxuddQ9QmOP5xXfqpe7smBApN27RJNE8/pri73Ru2aTBZhKu2JLtkSUzNEz7EVaUSbg3rwAHQ3w6qK9mLN+GgmLP8mIONpD64z0Mql/HZpVfmQdRvj0fOFgWgO92EjDoqoqvywDh+Ffd53+5AdizqsoJx1t25yG/IDJe2+E22bg+goThQkHCCDOgU3PTHJPmrM1CYXHwjwaLiIiISN0c2ae1q60MfT5nQ+17l3Eip5XVEpfiLotLSgcGXBpS5ca16UdY0fCSDVk5pgdVQzHfVmpMQ8O0H2FFmYS0PTsysgmZFextnYKEQRta0rZ5o6p7xc15FNj0s/PnlHbA+FecWYOV6TrJHUTkgJO5ZX0LKxlewr9Q63YoSFzfmYSMn6Qmh3YJcH1SkDDCsLHvYT2dac08emgvMxARERGR8F3jcdKxvST3txXba3dj028CHGUBxkNvBaZ8AZw5Axj3IhDv2csPqz4E9mUErdw4rQ5BQntfQmbn7NwXGYElX8xdn+URLK3L8xiKQcJImHDMTKblme4AlWX9jv0osZW/SuDwebbvg2pLjaOigSiWi0YBE98GktKqv5PRjzkHm4y8BzjmOR+Gl6gvYX1nEjKLsDYZ65FKQcIIZPUlJE05FhEREYkME4Z2RGKcu5/RJ39k1Dw7bsMPQMY3zp9T2gPDbnD+3P5IoMcUz+vmZALfXQC81hv45UbggGdZXiDw8fgrk7BzumfAs6H0JdyfX4g1W92BhkO6hXepsXe5caQML/lhUfksQqudwNYIn2AeKjJ27Pco7612aMmhNwNn/eacXtxhjG93EtcIOG8eMOofQEz5/qppjRM8stjUl7D+MwlVauxJQcIINKxbOhJinbv291U7XNOyRERERCR8sSRq/OAOrvOrtmZXmIlUpaxFZZkwrGF+yPkFtjJ/PgIUHwRKCoF5TwKvdHOW2/GyAMkrKDbTRu09CWvLe8LxxqyGESRckLHLlC1GSj9CapwU5zGIZnuYZxIyg+3HJe4godUuyqKS4/qxbLPngY9qMwmp7WHA0L/W7I5iEyr9J2aw2bMJ125TuXGg7T9QiOy8wnqbbBxuFCSMQDzCbC0G2Nh3RU0XjyIiIiISkk4+tDML3Vw+mV3DUuDhfwcuXAIMuxHoc07V1z3ifmDE7UBsWaCuIBuYcTPwai9g+VvukuUAlRpTXcpkGXix93FsKJmE82ylxkwc6F/VIIYwzSYM90zC+euzsDunwHX+rCO6ISba/c5WNln9WLLJ/T25WXKC6UlYzsEAfJfeMQ+YeU+FfQk37cox2aQSOBpaUjUFCSPU4b3dJcfMJhQRERGR8Ne2eTIO69nKdX7Wqu01z6pK6wOMedzZX6sqCU2BIx4ALlkD9L/E2YeLcjYB314AvDUM2Pgj/Mleakz2IF+dh5c0gCAhy7Xnrd/lOj+gUxriY90l6pHSlzDcMwm/tw0siY2OwnGDOqCTrTxewyvq571izyTs16FZ+b50e9cAL3cFfv+nc7KxP8x/GnjnMGD2fcDqj8xF3Vu7J7GzHSXLoCVwFCSsmoKEEWpEj5awDkbNXLW9QU1zExEREYlkpx7WxeMLJScdB1Tj9s4pnhcsArpMdF+etRD4aByw8j2/3dVuryBhXQdudG7ZxOOLYa0nQodRny17NuawMJ9qXFkm4fbs/LBtqcRSx9mr3UkcDPo3bRSPbrZAEcuN9f0tsDj8Zk9uQeX9CNlm4euznRnUs+4G5j7mnztOSAVKy/og/nAlkLMF3W2ZhKRM0vrrRxgfG430pkmegeENU/0XFA5DQQ0SzpgxA5MnT0bbtm1N1P6zzz6r8vqffPIJxo0bh/T0dDRp0gQjR47E999/X2/bG06aNIp3/aHbuicfG7PcbwQRERERCV9c49kzT75bsBl5BUWV/0LOFmdPwWL3F+JaSR8AnPI1cPpPQMuhzssadwC6nQR/ybIFuHjAu3lK5b28appJWFzqQObuyB4IYc8itHqVRwp7JmFhcSn22gI84WTa0q0osgWrjxvc3vzf/p7el19YLqtW/Gupdz9C7yDhb3c6y4KpWU9g8DX+ueO+FwA9T3P+fHAP8P3FaNUkwfTdtKzRhON6yyRsn5biLvXfvxF4ayjw8XjgsxOBgoaZ0RnUIGFeXh4GDRqE554rPwq8sqAig4TffPMN5s2bh7Fjx5og44IFCwK+reFecsxSFBEREREJfzy4fsoIdzZhfmExvl+wufJfmHmnczrx632BHX5YN3c8GjhvDjDxHWDsM0CcLQuD1n0JHMyucyZhs5QExETX7euKvYSzIZQcs9edfTJ0JE3tbG0LElqZYOFo6kL3e5VBcGv6dDfb8ArS8JLAWmbrR5gUH4OurWx/K5hJNvdR58/RccCkd4F4P72XWNJ87AtASlvn+Y0/IGrhsxpeEgqTjfesBKLLBiRlfAO8OwrIXo+GJqhBwgkTJuD+++/HlClTfLr+U089hZtvvhnDhw9Hjx498OCDD5r/f/nllwHf1nA00qNfjfoSioiISO0VFxfjxx9/xIsvvoicHGegZevWrcjNVbVCMBzVr61Hlt1nczaYianlMCi47A131kqTTv7ZAPYz7HM20ONkz8v3rgW+OMU5CXn2/cDS14DlbwMr3wfWfOIMIGZ8587Qsdu3AaV716Bl1E40j9qNTskHnMHGojxn6V8tyks7tEhGtK3PWMbOyM0MKSwuweINu13nh3ZtUb7HWiip4f60lxuHa5Bw/Y79WLvd/Ro8ZkA7VyDcI0hlSk4j97UaCpZucmcS9m3fzH1AIn+ns+eq5ciHgVZlmdP+kpQGjH/dfX7GLRje1P19fUNWjnk/i/8VFJVgR/aBiicbdx4PTHrHfX73MuCdEUDmrw1qV7jnyIeh0tJSs0ht3rzyUeUFBQXmZNm/f7/rd3kK5Laxj0Qg76M6bPTcrVUT0/iWKcs7svOQ3sTrSK+Ezf4U/9H+jCzan5FF+9P/z6c/bNy4Eccffzw2bdpk1lWs7GjcuDEeeeQRc/6FF17wy/2I7+JionHi8M54fdoqc55felg5cmSfNp5BmOn/xx+c5w+7C0iqfN3sF+zdxX5bDEjOvKvy63UYA5wxzfOyz6fgBvY5tJarXLZ7FxxFxQBHPAh0tH2JrwKHdrRPS3ZljkRyJuGyzXtRUFwaGqXGWYuB7XOA/CzgQNnJ9fMuROVnIT02BRh1NzDEtzLOlk2TTAm6FQvfEYYTjqcucg8soeMGOUuNKTkhzkzXZasoWrd9X71vX0PBUvXMPXnlS405sf27i4D8soBd5+OBYX8LzEZ0HgcM/Rsw/ymgpADjt92G1/AAihBnDvjwb1XPtqmBue8GLHN3rvWJaJTLtu4yAbhkNfDpZGDvKvP3Ch8e48z+HMABXpEvrIOEjz32mDl6fcYZZ1R6nYceegj33ntvucuzsrJw8ODBgC7K9+3bZwJL0XUsk6iLQe1TXNOxps5bh3H9WgZtW8JZqOxP8Q/tz8ii/RlZtD/9y8r4q6u//vWvOOSQQ7Bo0SKkpaW5Lmc1yOWXX+6X+5CamzS0I979dY0rMPTpHxmeQcL1XwGbywJxqd3811OrKkc+5AzkrXi76uuxhM9baRV9FS0OZtd4ZccxGMqAVOvhzlK+CvoSWkHCjKzIDRLOW+cuNeazMLRLHYeWsHF/9logb7tnkK9c4G8XcOk6z7JzTm2d/c9Kb5rbF1OcD8f0vwPdJgNNOvoUGGfCw459ziygbWEWJGQfwp+XbHGd790uFR29yuG7tWpqCxIqkzBQ7FONqV+HsiDh/GeAjG+dPzdqBRz/evVT4Ov693LjDyZjLSVnOS6IewuvFF3iyiRVkDCwpcZUYUuGZj2Ac2YDX50JbOQQkyJg6qXOzMKj/gVER8bE+IgLEr7zzjsm+Pf555+jZcvKA1+33XYbbrzxRo9Mwg4dOriGnwTySw7T+3k/wQwqHTs0CZ/M22p+XrI1H+eOVZAwnPen+If2Z2TR/ows2p/+lZhYt8mwll9//RWzZs1CfHy8x+WdO3fGli3uL71S/4Pqjh3UHl/P2+TKJFu5JdsEH1BSBEy/yX3lIx8BYhPqYaM6ARPfAkbcBmz7w1kmzC9YPHGbrJ+bdi33qyVdTsAvO9IRixLEoBhdWiSiXWp8+d9NsQVCadtsZ++o5n2AgZc7BwOwnK9M55aNMX35NlfGZX5BMRolhO3XIJ+GlvRo29S8Pupk3RfO0nFfMGAYZwv0JVWSxciAS2IaHCUFiCrcD8ek9xDlQ4DQ3pfQFSQMs3LjP9fsNANJKsoitHB4ya8ryl6r+w6YSchNkuq4H6WcpZvd/Qhjo6OcfzPZmmHGze4rTXgDSHa37wqI2ERg4v+Adw41fyunxH6Jz4tPxC5HCw0vqYcgITOTmb2L9V8DLYe4+0RSYqpzUBf7+S74t7kI855wZhee9HlEBwrD8tPxvffew2WXXYYPP/wQxx57bJXXTUhIMCdvDPQEOtjDoFJ93E9VurZqYiaB8UN0ycY9yC0o1gdNGO9P8R/tz8ii/RlZtD/9x1+fWQzelpSU74+UmZlpyo6lTGGOczpii/719pRMObSLK0hoZRPedsoQYPF/nV9mqN0RQA8fgz3+ktbXeaqBnQPvxr9muEuQbxo+CO0Glg+kgGX0O3e6zy9+yfn/PSucX+h+vRXocSow4HJT1swgoR37fbEHWSTZk3vQ9LuzDOvqQ6lx7lZgy0xg60zn/4ffDPQ63f3vbUdVfxsx8c6AYFFu+eE2414Eklo4/52nRulAYjMTKHSUlmLXmt+Q1u2IGj3ONqmNsAjOvovbwyyT0D6wJD42GmP62QISZbrZJhzT+u37MbiuGaFSzjJbP0IG1BPiYpwHUTjFmNliw/7P2Z+uPrQc5GyhsORlPFd6C3Ztde7vtZpwHPDJxm2aJSO+aB/w1VlAaSEw6Gpg7JPuK3OIydHPAGn9gJ+vdbbSaD0iogOEYRkkfPfdd3HJJZeYQOGkSZOCvTlh8UVrVK9W+Hh2BkodDvyxeifGVXDUSkRERKQyxx13nBkg99///te1vmDLl3vuuQcTJ07UE0csu/xkgjNIeOavQFrvenle2HT90O7p+HOts9R0xvJtuOyI1kifdY/7SqMfr7AMN9Tssk02phZNfMyEZQ+pfeuAzBnO88xeXPmu85TaHf27X4im6IJ9cPb3Yq+vSAsSzrdlEVbYj5C91nYtKwsI/uYMCu7f4HmdzOmeQUJmUQ26CohLcQb4rICf+bnsfHzjil9bLfo5T1Uoadqz/IUMbvc4rdLemfYJx7tzCswQAhPgCYMgrvUepcN7t0ZyYvmS++5eE45ZcqogoX8dKCz2GArT3yo15kGNc+cA858GDnFXItaLYTeY91rSLxuArRnmooydOSguKUVsjBJUApVJaIaWLHjGfZCjpJJ2dIOudAaQV38IHHYnIl2NXnH/+te/cOCAexLMzJkzPYaCsO/N1Vdf7fPtcXG5cOFCc6KMjAzzM5tiW6XCF1xwgUeJMc8//vjjGDFiBLZv325O7BUnlRvVq7Xr599XbddTJSIiIjXCtRfXfX379jU9nc855xxXqTGHlwhrsm91Tuxlf7aPjwP2u7OGAu2Uw9yluzwovO27O4GDZVNue58NtDk0LHbR7v1eQcLGPgYJGdg6czpw0QpnBlCiu9SYPfWazL0L7yRdiJNiv4jY4SX2IGFSfAz6sHzS8uvtwHPNgTcHAj9eBaz4X/kAIVnDGuyO/Q8w+l/A8JuA/hcD3U4A2owAUrsCCU38G3ye9xTww5XA+0cCOZ4DPuyZhHbhMrzkpyVbzHvTctygDhVer1lKgsfU8gY7vGTjj8DUK4C9a/x+08sz93rsC9fQEmJfzRG3OjNk6xPL8OMaoYctSMwelhuzvDJ0pU5KSkuxZbd7YE03HitiUNjKGhx+S+W/3HGs8++h99+8PasbdpCQQTt7A+wJEyZ49KHJz8/Hiy++6PPtzZ07F0OGDDEnYu9A/nz33Xeb89u2bXMFDIlHr4uLi3HNNdegTZs2rhObaUvl+rRvhtRk5x+6ueuycLBI49RFRETEd+3btzdDS26//XbccMMNZr328MMPY8GCBVX2hm5QxjwBpA92/pyz2RkozPfM7gqUwZ3TzHAOy8odB+GIjgdiEpyN8cNEVo47GaFGQUILszfHPAZcuQWY9B7Q8RjXP8VGlSCjtLP5OWNnZA2EYMBj3bqVOCLmN1wa9xoGd2rumX3EL78F+8r3Qms/GhhxOzDla+CaPcDkDxE0BfuBOWUHHHYvd/aY3L2iykzCcBlewsGHUxdmekxpHtzFFsiuoC+hxZ7x1mCs+xL4+HhgyUtAgmdmpQke59taDdTCsk3ufoSp2Iu+bTxfU8Fk7Xv2ZD0n9l1kZiwL9iZFFLZgK7bGowMYmf8RUJDtPNPnfKCp8zPCZ5t/Ad7o52xzwUFPDbHcmH/gqjpfU2PGjKnyNl5//XWP87/88kud7q+hiomOwmE9W+G7BZvN9Lv567M8sgtFREREqhMbG4vzzjtPT1Rl+GX21O+A945wToTdsxL4dCJw+k/OkswAYvn3lBFd8MSXi835Vw6eg9QxV+G49C3OQSJhguWjlpTEWCTG17IzEnuL9T7Tedq7Flj6CjKXTMXi/AGuTEJ+B4la+Q6w6kPnsJPOx/unzxS/2xTnAweznT2umnbx/Pc1nzpfHyyJNkNdyv5vDXixLut0HNDvAs/bfXdk2QAX++8UwlF4AC9iL1CWgDatzZWe99nucKBRS2dvyraHO8+zSX99Z0tVhVmJZ/0GfDweyF7nDLTzvTTlK6DtSNfV2GvdbnsYDC9ZtXWfR4njsQPbIbqyDMyZd+OwlL74E86/GZm7c02CR2IYlFT7xaafgS9Pd04xb9wB4MEOuz8eBBY9D6QPAjqNAzodC7Q70mTh+Wpp2WTjaJTgwZSH0eSzJ4BJ7wLNeyHY2qWloH3cHtwUfT96x6zGloWrgBGzI74HXn2xvw8TcBBdMl92Z3IeemvNbix/F/DFqc4+hfOedH7m83XkHdgOQ2HXk1Bqh30JGSSkWat2KEgoIiIiPnvzzTer/Hd7e5gGjT3cTpsKvHs4kLcN2D4H+PwUZ6AjwJOFx/Zvi1d/XonsPOf01HcXF+HYq8+oWdlQkO2ylRun1TSLsDLNuptsyrkJVwDfLzcX7T9QhD25BUhb+LyzR9+6z50Bif6XOE8MqDG7hIE+/p8BNfv+YyBj1fvAwb3Of7dO1vUZ7KMWA4ALnYFblyUvAxnfVL/dic09g4QMKrGcnV9IvXiHDwbHc2DNOPcFDKb8ZXvo96VM7QacNRP4ZCKwcz5wcA/w4THODMeuzl70TZLi0Cg+FvmFzudhW7Zn9mkomrrIs/WAq9SYgV/7PuHfi9n/BB9px4R+eL/oNMwpPQQbdu5H73aR1UOzQpyG/tmJQEnZwYJmvcofYNn4g/P/WYucp7mPOYPdDH4zYNjxWKDVsEqDaizhXZnpzCQ8L+5ddCtdBjAxkYHJCxY6g0VBTu5p1bIVUnc7s9va5c8F5j4OHGqbuix+GVoyMfY7xBaUteXodSbQvIIeqVVp1MKZqf/TNc6/yxnfAu+MBKZ86fxbFsYUJGwghnRpYfqTHCgswR+rd5h6/BhN6RUREREfeLd2KSoqMm1m4uPj0ahRIwUJ7Zg5dur3wPtHOQNGm34Evj3fmWEQwGyQ+NgYTB7WCW/NcPbw2ron3wysG9mrFcLFLlu5cYsmSX697c7pnsGGzVs2I83el4+Za7/f6zx5u3SdswefhRkjHLBRHauMzS66/LCKCjFLsNzvMqsqyhkU4cmUlMdjd14pthalYllpH2xLGoIbhnhl/AY58FHjQPuZvwCfTwE2/QQUHwA+Owk47mWg/0Uma5Ylx9YkZ5YPhjIOVvll6VbX+YGdmqMNtgBf3gGkD/QcgmB7TQ2IWWZO60s7Y9+ivwFtrnGWjUeqrCXOwU9FZf3iuk4GTvzY828mAzGcWs6/qTvmM8rqfq9snuY84Q4gIRU45lmgz7nl7oY9HllZNyB6Cc6Ofd95IZ/X414KmfdJ+7bt8Oi2G/Fowm2IjnLAMfNORDHQ38rZok3qnkkYh0KcGf+p+x9G3FG7Gxx4BZDaA/jyNOdBjT0rgP+NcL52O4wO211V4780L7/8MlJSUszP7A/IkuAWLZxjuu39CiW0cOF4SLeW+HXFNnP0dOmmvRjUufJeGCIiIiKWvXvdPZwsa9aswVVXXYWbbrpJT5S39AHOPm8fHesMcnAiIjPDjn0+cNlcaz7FqViAT2OGILfEWaL3yR/rwypIaC83btHYv5mXnW09G2lddgwGX77Bmf3B4Ayz+zgB2JdgH4MQHqKAxFTn5dYpsRmQ3Kb8bXEASN/zywX6yv1c7j4AXJ9b7vVzsLAYFz72g8mQokkDOzq3JZwxe4zvn+8udGZssvT0+4udg1WG34w2qUmuIGGolxvPWrUdeQXOrMfm2IPr498DXnvX+Zj42ht4pXNaNB3zHNB2FBxz/oUoBqIZK4veAKz4G7DlSedQngGX1qi0NiywJcBH45yZudRhLDD5AyDGK6DOYN5RD7tLPRkUZMCQ2YX7nBOBXe/X5LblexlumYkV27ugMfbjlvjHTQDOGHWfcxhPiOjRpik+L+2PD4pPxVlxHyGKmcnfnAucN885WEXqHCQcH/sDmqEsi7DHKdVOYq8SB5qcwyzYyc4DSBwaxs/+Y54HBl6GiA8SduzYES+99JLrfOvWrfHWW2+Vu46Ebskxg4TWB5aChCIiIlJbPXr0MMNL2Kdw5UrnF1qxaTfKmU3A8jlmwBzIcv7f+4uvPxQfNI3Tk/ZvwBuN0nF5zuPYg+ZYvHEP1m7bh+5tmobF8I3dOQEoNy6TmpyAZskJ2JvnDERmcMIxgw7dJjtPDCIsfQ3Y8J3zcivQZ/3frvN44PwF7n9nUMvXLCT2A6ytCgLM3MdWgJCGdS0LOIU7lndPesdZ+r3g387LFjxrMnfsw0s4uMT0lwzRUuqpizLRCHk4I+5jTIn9HIlb3IFwE+xjUMEKEjI43P9iRPW7EK+//C+M2Ps6+sSwdJw18huBadcDs+8Dxj4N9DkHEYHvOwZUrMnaDNad/LlzsE51pZ6cas4TZa93BgsZNNw62wRbPaz9HPj5Wkxha4akpkiN2ucOSA4PrVJea3jJW0XnYljMAvSIXufMUPv1FuDoZ4K9eWGLfycydzkzVY+J+bnuWYTebS3OmQ18dSaw4XvnZ/0PlwO7lwGjHwu7npI1ChJu2GBLyZewM6JHS8RGR5mJPr+v2oG/HNc3ZD9QRUREJDyGmWzd6i6lEy9dJgDHvwFkTgeO+U/gvigwiFJWOhuT3h97ctxBrU/+yMDNJ5dNXQ5h2XkFKLFNnUz3c7mxlU24N6PANbzEQ+P2wMi7nKfqJKU5TyFg3vos188chsFJ1xGDgVcGxJiRyd5zp31vArNtmu3zKOdlH85mKYHt+VkbO3fvRZdNL+G2pA/QJMr2eotLcWaUDruh4qFGUdE42OlE/G17fwyIXoqz4z/CsOh5zn87sAtIipBAMCduM4OQAVCrh+eUb2o36IntAFKvBAZdWb7XIyuSN/zg6t1pBQgLYpshYcJbIRfA6ZiegvjYaBQWx+GRgr/jhUZ/Q6yjwPl3vstEoMvxwd7EsLQr56Crl+nNBQ/hwUHLMDBpM9BqqH/uIKGps//w9L8D8592Xrb4RWe2cFpvhJPQKLyXepGcGOfKHtyx7wDWbXem6YuIiIhU5YsvvvA4ff7553jhhRdMFuHhh9chM8pHzz33HDp37ozExESMGDECf/75Z/jsMGb8jHsxcF9EWXb3xwNlZ6KQdNxTGGrLJpu+bKtHhl44lBpTmp/LjamLreR4Y1aOR1AyHG3Pzse3ZYMJqU/7VLPejygM9oy4Dbh4FZDW11zUOtWz3JbZhCFn5XtIfqcfroh/xRUgLI2KA4ZcD1y2Dhh5d5XBsG6tmE0WhSWlA3D7wXuxZdJMoPfZQOvhzgEdduzPt9s5lCesxDcBup3o/Dm1u3PoU1Lzut9uBUkwHxeegA+KTsHaUmdv0SLEI/aEt4DG7RBqODegS0tnNuFmRwd80eha9z+y7HhPWXap1HqyMfe/Y+BVzs9mf4qOBcY+VfaZHwcwCB1mAcIaBwl///13fPXVV+Wm3XXp0gUtW7bEFVdcgYICzw94CS2jerd2/cwpxyIiIiLVOfnkkz1Op5xyCv7xj39g4MCBePXVVwP6BL7//vu48cYbcc8992D+/PkYNGgQxo8fj507OZIyTGWvAzZM9c9tcdAGM3KIk3nTB+KUEV1c/8wKki/nlmXqhMlkY2rh53Jj776EhcWl2La3bEhCmJbOPfXVEpNJZzluUHtELJaXlmnTrBFiUIzL4l5FU2SHZF9Cx56VSC50tnkqdURhVvQ4RF2yEjj6aWcJtY8lp5YVhR2d5dec/uwdBJv2V+D1fsBnJwNbf0fY4OM48mFntujpPwLJ7u+p/vT9ws14ZXUbvFJ0Ca45+Awui/oYuReuR0w358TsUNSjjXv/v7nvGDi6nuA8w+EYVpaa1HqyMXVs4ZyzERADr3AOvOp5KsJRjcqN77vvPowZMwYnnOB8kS5ZsgSXXnopLrroIvTp0wePPvoo2rZtaxaNEppG9myFf3+z1NWX8IIxNRz1LSIiIg1OaWklAx3qwRNPPIHLL78cF198sTnPDMavv/7aBCdvvfXWctfnAWv7Qev9+/e7HgNPDK4E8/EgaxGiOMWzYB8cp3wHtD+y9re1ZyWiFj3PsRlwxCXDMfJePlAM7ZKGDi2Ssbms/9JX8zbijFFdkRgXWmV1dln7PAM9zVMSKt1Ptd2PHVske5zP2LEfbW397cIJe90tyNjlOt+/QzMcO7BdcF/bNVCX92KLlHjcEP8MxsX+jFExv2POtrdQ2q+CITH1iUNvbH0pl6dfhHaOp7GmtBteK7wQR44+Hoc16QyHj4+3XfNGiIuJdvWbZG/Ro/u3BaJizHvcZessRG/5zfnzus/NydHuKDhYztx5QuAGJZXxy9/UwWWZcgF47XLAzbPfOr/7WiX5151yFJo2bx7S7xVnJqnTgcJSZB76Atrvn2h6mjrGPOXX5yokPhfrwcad+5AetRNZjpZISYxDk6TYwD7mlHbl9lPUjJvhaNEf6HuB3+/Ovh/r+rhqFCRcuHAh/vnPf7rOv/fee6bkwxpm0qFDB3OUV0HC0MUm0L3bpWLllmzTsHnrnjy0be65YBIREREJBYWFhZg3bx5uu+0212XR0dE49thjTYVLRR566CHce++95S7PyspCfn4+9u3bZxbSvJ1gaPLHE2hU1qTf8dmJ2DPuExQ3q91kxdRfbkAip6QCyO1zNfLyY4B8Z4blMb3T8PpvziBhzoEifD5rJcb2Cd1eZpt27HH9HBcThYM5e1GQW3GAg1+AarMfG6HUGVAtO780Yzt6NA+/7kvZ+UV4cepyj+fr/MPaYleWuz9hqKvtPqToAzswNHax+bld9DY0XXYG9rR/H8XN+6O+xWavRMrCh1DctAdyh9zpuvyz+dux5MC/kY1mJk53XZuEGmc/t2+WiIxdzuD5is27Kvz9KEdrJA29B8kr/ouYA87MxagtM8ypKLUP8vpeg4OdTnKWQQZ7P5YWo8kfN+FAt7NR1PJQBNqBwhLc9+kKkzVsOW14W7RKLA75TPTmCe4MYZq3YR/ij/4YjthkYJf7b2Ww34vhJGXzl3g98V78XDIWMxtfZNYE9SlpzVtosuAZ7D3mAxQF4PVn3495eXXLkq/RX4u9e/eiVatWrvPTp0/HhAkTXOeHDx+OzZvdfTEkNI3q1doECen31Ttw6mHO3gwiIiIiFpb41iTbLxB27dqFkpISj/Un8XxlE5UZULRvOzMJeSA7PT0dKSkpZmgbfw7al6GJL8Hx+U5EbZyK6KL9SPvlXDjO/BVI7Vaz29k8DdFbnCXLjpR2SD7qLiTHuQ/8ntwsDZ/M24b9B4rM+Z9W7MbpR/U1mTShKL9ku8dBbe997v1lqLb7kaWqW8vKU7PyS03LpHDCL4AvfLQA+YXuIML5o3tiYM+OCCd12YdASzzY7L84f/df0Sl6M1JK9yD5p1PgOPFT57Ta+rB/E6J+/wew/E1EwQHH9hloNOpmMwDnQGEx5mQsxEE4BwgN7dICvbvWvBS8V/udriBh5t6D5rkqP3SyJdDubuCIW1G68n+ImvMoovY6e9bFZa9A6qxr4Vj6GBzD/s85QMHPvVF93o+OUkRNvQxR699D0qYv4ODk907H+XVbPO7O4cCDnyzEjv0FHkM8Lzx2QMj+DbRrllaK2OiVpl0E7cwH0ttV8J2drSY4MCNo78Uw4XDg6Pw3ERtVguNif0Ru0yn1/rc/auVOOI59Hs36lfXh9DP7fszN9SytDmiQkB/WGRkZZqHFI7vsC2M/UpuTk4O4uAhrlhuBRvVqhVd/di6sZ67criChiIiIlLNgwQKfnpXyX1qDKyEhwZy88csPT9xe6+egiE4ETvoE+PBYYNtsROXvQNQnxwNn/QaktPG9tHHGTa6zUUc8gKgEzyEISQnROGFYJ7zz21pzfvPuPCzI2I3h3UMzKLY7192TsEWTpGr3T233I4eXWEHCjTtzw+5L8Yzl28xBfkuPNk1x2siuYfc4qC7vxaQWXfF/Wx7BfQn3om/MKkQV5iDq04nAxP8BPU9DwBzYDfzxILDwOaDEHYCKSkpDVM5GoGlHzFy1AwftvSIHd6jVY+zepimw0JmAk3uwGFn7C9C6svJ4/l0ZcCnQ/2Jg7RfAnIeBbX84t23/BkSt+RAYck1Ayo+r3Y+cNvzL/wHL33Bev7TI+bkRwNfsp39k4LeV7gMPrVKTcNNJgxEbE7otF+wSoqNND9W1ZYNGOXC03PO7cyHw6SRgxB3A4KvrdH9B/1wMsPwVn6MLnJ+Fq0u6o7TT+Pp/rGMeNZnsgeSv/VijIOHEiRNN75dHHnkEn332GRo1aoQjj3T3UVm8eDG6davhUVCpdx1apKBDWrJZLC7fvBfZeQVITfb/BDkREREJX9OmTQv2JqBFixaIiYnBjh2ew9Z4vnXrwDS5rxfM+JvyNfD+UcDuZcC+9cDH44EzpwOJzuyjKpUWA92nAHtXA816An3Pr/BqJxzSCR/MWufKRvl4dkboBgltg0sCMbTE0rllExPEoa1788zgj4QQ7tVotz+/EM995+6vFhMdhRtOGGimoTY0zAjNQRPcWvAA7kh4BCNi5gAlhcCXZwDHPFt90ITXLcgGDu51/t/8nA0U5zsDbXZzHgVWfwQU7AVyMoHiA+5/S0gFRtzu7KsXl2Qumrow0/XPKYmxJkGjNryHl6zdvq/yIKGFfRF7nAx0PwnInAHMeQTI+BY44iHPACEDd+b69XCQZ9Y9wIJn3Ns36V2g8/iA3d3yzL146ccVrvPs7XjXacPQOCm8kpkYJLaChNz3zI50HZTj6/C9I4GiXODn64AmnYCuoTuIJaj4Wp/tbpn3TvGZmJAewKElEaBGnyjsRxgbG4vRo0ebPoT//e9/ER8f7/p3NpA+7rjApQ2L/6cc8+Nhtu1opIiIiEio4Dpz2LBh+OmnnzxKanh+5MiRCGtJzYFTv3d+uaNdS4BPJwNFPkxqjYkHRt4FXLIGmPCmx8AEO5btjuGwgzIcdMFhHaFoV447SJjWOHAHr+0Tjhk73eQ18TKUvfjDcmTnFbrOnzmqG7p5BZIaitapzmBZARJxb8EdyOt2btm/OICfrgGWv+X5C5unOycAv9gOeLoR8FQC8Hwr4LXewDuHAR8fD3x9FjD1cncAzcKAzPY/gb1r3AHC2ERg+C3AZesBDgkpCxCy3/uSTe6ecWP7t0N8bO2C0F1aNUG0LYbHbDKfMZjUYTRwyjfOvxPtRnn++4bvgf8dCmz4ofzj9ae5j3sEaDD+1YBOfN2XX4gHPp6PkrIDI/SX8X1Nxm24sW8zM0m3Z9uC043buwPhzCz/6kxgx/wgbGUY2PQzGu2ZZ37MKO2E2SUj0LGFZ+a91CGTkEdzZ8yYYRoisqcLj+zaffjhh2jcWE94uPQlfH/mOvMzj6YePyS8+piIiIhI/Zo7dy4++OADbNq0ybSdsfvkk08Cdr/sL3jhhRfikEMOwaGHHoqnnnrKNOW2ph2HtcbtgNN+AN49HDiQBWydCXx1BnDSZ74NGmB5cjUlyqeM6IIfF29xnf/0zwzcOHkQQkleQZEZMmAvN66PICFt2JkTFgGEOWt3euzHji1ScPaR3dFQMZPQUoJYLO/3OIandQT+fMh5YWJzz19gIGW3e9hLpTgIqCgPiE/xzBZkoSB7vzHTl730DrvL+f718sMidxYhHTeo5r0ILZxG3j4txRXIXlvbAH+z7uWfi99uB3YuAD4+ztnH8YgHgbaHwa8WvwRM/7v7/NhngH4XIlBKHQ488tlC7LJlJY/t3xaThobn99zurT3/LnHCtf11jyMfAvZlAKs/dL5mPz0BOGc20CQ8H2/A/HG/68d3i85EXGwsWjYN3GdMgwsSXnLJJT5djxmFEtp6tm1qSjl41HbB+l3ILyhGo4TATL0SERGR8Pbee+/hggsuwPjx4zF16lRTObJ69WpT9jtlypSA3veZZ55pphDefffd2L59OwYPHozvvvuuysEWYaVZD+DU74APxgCFOUDrEUCU/8pfu7VuioGdmmPxRmd2089LtuKSo3uHVKsZe6lxoMuN2zVvZMoPi0qcE08zdoZmZqUd1+lPf73EdZ7JZTdMHljrDLVI4BEsAbCNWVZHPggktwam/c1ZPmzH4F5cijPgl5gKJDSz/Ww78XreAfrD7gBG3l3t0A9mr/2w2B0k7JzeuM4BaGaKWkHCddv3wS84AZ2BQsvmacC7I4FuJwFH3A+08MOU6JXvAT9c6T5/+D+BodchkN79dS3mrcvyCKT/ddKAkOubW5P+qRyywuAnrdm+D0f2tR0UYgb58W8AuVuArbOAvG3OHoXsb1vHYSYRI/M3YPMv5sfNpe3wa8nh6NwqxbRqkMrVKCr0+uuvo1OnThgyZIipiZfwxT84I3u1wpdzN5pF0tx1WTjK/kdHREREpMyDDz6IJ598Etdcc42pGnn66afRpUsXXHnllWjTJvDrh2uvvdacIlarocDJXwC7VwCDr6r8ej9e4+xnyB5oDG746NTDurqChFz3fTV3I84b3ROhIstWahzocmP27+uUnuLq9cVMwlDHgYNZtkDqSYd2Rt/2PvSujGBNG8WbTDtrQMi27LIy/aHXOweXlLozU42Wg4Hra7mvWd7vg4Ubdnnsp+MGt69zgIpBwmlLt5qfd+cU+KeXPAOp588HVn0AzLwLyHYOdMC6z4F1XwB9zgVG3QukVjBN1xcsz/6OGYNl8YJD/u4crhFA89fvwlvTV7vOs8/onacNRVJ8+CbB8DHwb1VG2d8oZhKWwzL3kz53Bnm5H3ctBb44zVlmHhPcHoxb9uThfzPWmIM+54/paQ7OBDOL8L2iM1CKGBM8lqrVaE9dddVVptSYE47Hjh2LV155BZ9++mm5k4QHBgkts1a5pz+JiIiI2K1btw6TJk1y9QlkuS+//N5www2mR7X4QYcxVQcIdy4CFj0PzH0U+N8h5YMgVTi0R0u0a57sOv/F3I0oLPb99wNtt1eQMD2A5cbeJcfWF/BQxf52PKhvn9J68dheaOj498eeTbi9bGK1kdIWaNKh3rfJPrCEmUrHDChfjlznktOa9CWsCrPQep8FXLQcGPei8zkzHMCKt4HXejkPSnBIUk2xX97415wZ0QOvAI76V0AHpLC8+OFPF1ghSeNvkwagU3r4t0EzE65t+77CRK1GLZxBwcQ05/lNPwI//iWwvSarwc+XO975Ez8t2YL3Z63D539uqP+N2D7X2XuTBxFKW2FayWjXEFfxY5Dwueeew7Zt23DzzTfjyy+/RIcOHXDGGWfg+++/V2ZhGBrUKQ3JZSXGf67Z6Sq7EBEREbFr1qwZcnKcwZR27dph6VLnhNXs7Gzk5/swaENqZ+OPwOKXnV/2TG+vsi99A6+stvTRu4Lk5EM7ezT3t7KTQoG9hxhDCc1TEuotSLgnt8BMDQ5FnLz85JeLPS7726SBSAzj7Ch/alU2vIS22YOEQZB7sMgj6eLQ7i39UtLvPZjGbyXHFmabMZB3yVrgqEfdvRwZHMzZ6Ftv1Ir0OQc453fgmP8ENEBYXFKKBz+Zb/6mWSYN64ij/RCgDQU9bPufj9GeqVqubcXJnwMxZa+5pa+6ymyD4bM/N3i8J6cvC8LnTfogMyinsHFXvF90uuldSsokrF6Ncz4TEhJw9tln44cffsDy5cvRr18/XH311ejcuTNyc8NnOpgAsTHRGNGjpXkq8gqKsXjDbj0tIiIi4mIFA4866iiz9qPTTz8df/3rX3H55ZebNeExxxyjZywQVn/k7C/1wxXAT9c6s0OoSWdgSM17e40b1B4pie4v/J/+kREyB/ntk40ZWOEaNZDYK85uQ1ZoZhO+PWONKdmzHD+4A4Z2bRHUbQol9kzCHdkHgvp6/mXZVhQWl3qUGvtDk6R4jyELfsskrKhsdfjfndOaD7sbiG8CHO4u1TT4/HJARkUqurz18BodzKiN16atwrLNe13n2QPyL8f1RaSwZxLS2qqCxO0Od/YoZGCX2aEdxyIY9uYWmP6QlIxcxKIIq7ftK5cxHnAMgPe/GLMO+xlTS451XawgYfXq9AkcHR1tUr35B7mkJHRKFqRmU44tKjkWERERu4EDB2LEiBEYMGCACQ7SHXfcYSYOc2jJqaeeatrPSIBKpUqYHeMAFv3HfflRjwCxNR/swd5cE4d28iizXZCxO+QGl7RoErihJZYuLT2zs0Kx5HjNtn346Pf1rvPMrrx8XJ+gblOoaZPqDp7lFxZj/4GioG2LvdQ4NTneZBL6S7dW7tfrukAFCS0ceHH4vcCVW5x9HO3WfAy83BWY/2+guMB1cUzOBkS93hdYUr/DS2et3O7xHuFBkDtPHRpRA32476O8/i5UqfeZwCVrnNmhQfLm9NVILdqIW+IfxUdJZ+ONxEvRHHvw59qdQdmejXsOurIIOa+kbXPPoUfihyBhQUEB3n33XYwbNw49e/bEkiVL8Oyzz2LTpk1ISVF9d7g5pHu6q4no76t3uKYniYiIiEyfPt1UjTz00EPo06cPLrzwQsycORO33norvvjiCzz++OOmFFkC4MiHgP6Xel7WZiTQ0xmsrY0Th3fymOr4yR/uL9ihkkmYFsDJxu77SEBKorupf6gNL2EJ5eNfLPJYl183sb/HNgvQ2nvCcZBKjjdm5WDVVvc0ZZa6+jMbtrut5JSZpZx2HXDxXt/rWX78253OycjTrnf2LFz2BrB/E5r9dAaicjOBqZcCK94J/LaV7evHvljkcdnfTxxc7jUR7thawN5Dz6dM0qbu1hIutektWQsb1y5B76X/h5cTr8LRsdMRHeXAotKB2INm+GN1PQUJ7ZO7OdW4bDo4tWmWHFFB5ECp0V8vlhVzgt3DDz+ME044AZs3b8aHH36IiRMnmqxCCT88qmyVLXBi1mrbB5yIiIg0bEceeSReffVV05P63//+NzZs2IDRo0ebA8WPPPIItm/X4LOAYR+vcS8A3aeUnY8Gxjxep/5eHAhyZB/3NOo5a7OwKQRKbe1BwhYBnGxsYSVUF4/hJQHOzqqhD2at88huPKpvG4/qH3FqY+tJWG54ST2ausidRUjjB/l3aEo3r+El63cE4fVamAOkD3Sf378R+O4iRL3WA7F5m52XtegPdB4f+E0pLsH9H80z7bIsp4/s6jGUM5LYg8QVTjiuzqoPgTcHOQO8gZKzBY4fr0a7z4difOwPiIlyB+peK7rAdJudn7GrfgZmfX8p8NXZzknPnONiCxJqaIlvahTZe+GFF9CkSRN07drVHFm+4oorcMopp5Q7SXg5vLet5HjljqBui4iIiISe5ORkXHzxxWb9t3r1alN6zIF2HTt2xIknnhjszYtc7C11wnvOSaGn/wy0HVnnmzzlsC4e5z8NxtRJGw7Oy85zDx1oEeDJxhUNL9m4Mzdk+jMyK+2dsn5e1DgpDtcc3y+o2xQOg0toW3Z+ULI+f1q8xXW+Z5umHq8tfwj48BJfJDYDJn8AnDcP6Hy86+Kosgw1R9NuwKlTgaSyCbsB9Pz3yz0y6vp3bI6Lj47cid/2voQctFSjAzvM9vzqDGD3cuCzE4GiA/7dOAYef7kReKUbohY9j1g4Xw+5jmT81OQv+PXYVchytHQNYlrEGQhrPw9cxmn2OmD5W8Cq94D3x6Ck8AC27Hb3y1Q/wgAECS+44AKMHTsWqampaNq0aaUnCS8cXmJVnqgvoYiIiFSle/fuuP3223HnnXeicePG+Prrr/WEBVJMPND/IqDDaL/cXK+2qejXwV0i/tPiTI/JoPVtj1cz+xb1UG5M9kAO+9nt3OfnL8+1UFLqwJNfLTaBU8tVx/X1y5TcSJQQF2NKx4OZSTh3XRb25hX4fWCJXXqTRDRJigv88BJftBoKnPotcOZ0oO0oc1FJcns4GCBMcWcpB8qPizPxzfxNHv0fbz9lCGIiuKqxb3vPlh5Pfb3E/K3wScdjgJSySc/b/gC+Pa9cOW6tHdwLvNIDmPckUOJ8D+Q7kvBO0Zm4tOAV9D7lMQzp1cWjxcXSZQtNBiq+ORf47mKg0M+Db/98BHCUZSsO/Su25ZSg2PZcKUjomxrNNH/99ddrcnUJE1x49O3QHEs37cHm3XkmJVdvIBEREfE2Y8YMU3788ccfm1YzZ5xxBi691KtvnoS8U0Z0cU0ELSguxdfzNuKcI3sEvdS4vnoSkr3cmFje652ZVt++mLMBKzLdrX8O7Z5u+ttJ5VqnNjItk4KVSTh14Wb3gOCYaIzp1y4g5fEsOV6Qsat+hpf4ov1RwFm/oXTvGmTlx6FlU/dQpEBh79BnvnGWkBJjT7dNGVJvfzOCpXe7VAzunIaFzMJjcuDmvfjszwyceljX6n+5cXtgytfAe0cARbnAmk+A6TcDYx7zT3Zp95OB5W+iJDoBnxRMwodFp2IfmpqM9XZpya5MT5NByFZn6z8ESsv+xi17Hdg6C5j0HtBqSN23Z/9m520Sp3MPuQ6bNnoGIVVu7JvIDblLjRxu6+Hw+yr1FxIRERGnrVu34sEHHzR9CMeMGYO1a9fimWeeMZe/9NJLOOyww/RUhZmRvVqjtW0y7JdzN3pkr9WnXbbJxvU13Zg6p3sGCYM9vIRZcK9NW+U6nxQfg+smDjABIqlcG9ugivrOJMzOK8DsNe4+b6N6tTLl4YHuS8eS9GC9Xz3wtZnaHYgNfIsADmv550fzTMmq5fzRPTG4i7O3fiTj34AbJg80fxMsr09b5dFrr0otBwGTPwSiyn5/3uPAgudqthFFecCCZ4ESr6zzkfegoP9VuKLoVbxcdIkJEDLr9VzbQSdWLVpeyz0ZO0a+CMSVDWPZuxp49zBg3lNAXVs+zH0UKC2bcD7kOiAx1WNoCXVo4QxcStUUJBTXYtEyU30JRUREBMCECRPQqVMnM7RkypQpWLFiBX777TfTn5B9CiU8sfzrpEO7ePS5WrzRmelR33YHqdw4OZHZT+7ghn1QSH1jP0SWENoDIJce08dj+6T64SVZ+w/Ua/Ds56VbPco+jxvs34EllfUlZPnkxhCbyB3o98fTXy9Bpq233PDu6TjriO5oSBmzV4zr6zpfWOycgF5S6uPrvcvxwLH/cZ/nhOp1X1X/e8UHgflPAy93A36+Dlj6mue/p3bFS6VXI7PA3XLugjE9PSaxH9bDc6DMTyVjgfPnA62GOS9g4PGXG4DPJgP5WaiVvO3AkpecP8c2Aob+zfxoD6TysyU5QRPifaEgobiOwlllF6u2Zpc7qisiIiINT1xcHD766CNkZmaaaca9ekVuc/iG5livMlZ7mWt9yrIFCRvFx6JRQo26IfmtL2EwMwk5HdcqJaUBHZtj0rCOQduecNLalknIeF199ZZk4MpeaswM2CEBzGrznnC8LhgTjoPkq3kb8cuyrR49Gm8+aTCiG1iW7YQhHTCsW7rr/Mot2fjo9wzfb2DgFcChtzp/Zl/Cr84Edsyr+LoM3C16AXilOzDtb0B+2XDTPx8CSks8/m6yXYWFLcsmDvX828Wy4/bN3QcV/2D2bbMewNmzgGH/577i+q+dU5g3/Ywam/uEM6BJg64CGjnfi5psXDsKEkqFU45/X60pxyIiIg3dF198gZNOOgkxMe4yJ4kMTRrFe3xxW7nF2aOwvu22HZiur1JjSxdbyfHm3blBKeFkJuWLU5e7zsfHRuOGEwY2uACIP8qN67PkmMND7NmnDLrbBzT4W7vmyWZQi/v+gzDhOAiYvPLC9+73R2x0FO48baj5+9XQmLLjEwYg2XYg5a3pq2t2gOOIB4BeZzp/Ls4HvjjVs4SY06qXvg682gv48Sog1z25Gz1PB075BoiOcQXK//vDchOct1x5XN8Kh8iM6OkuOV61JRt7cwucQ7nYG5G3mVQW/MzbBnx1Vs0GmuTvAhaVZUnGJACH/J9r+zJ3abJxbShIKB59NCyaciwiIiIS2Xq3T/XIJOSXqmAOLqnvAQT2TEKWjW6xlTPWBz7f//5mKfIKil2XXTC6p6vhv/hWhmlXX8NLpi5yZxHScYMCV2pMDEB2beV+vYbE8JIA23+gEA98NN9jOu0V4/qgdzvPab8NSXqTJPxlvLvsmAc2HvtiEYp9PcARFQ0c/zrQ9nDn4JEJbzqDdcwsXPku8Ho/4PuLgf0b3L/T7UTg/AXA5A+ANPd9z1mbhXnrd3kMWjrEluloN8JWcsy9+edady9PdJkAXLgY6DTOef64l4H4sp6FvljwtLNnIg24zDVlm58tnFxv0dAS3ylIKC5dWzVBq7Im1pxAlHuwrPGniIiIiEQc+5dtrvu27KnfIJl3kLC++hFWFCSkjJ31G3iZsXybR/VOjzbOqaDiu+YpCSb7sj4zCQuLSzBtqbv8tV+HZvUS2O1uKzlev2M/SoMQ1K8vfGyPfr4IO2zl40f2aYMTh3dGQzduYHuPYSBrtu3DB7PW+X4DsYnASZ8BZ//unFLNoFD+VkQxOMhBIpZOxwHn/AGc/DnQcrDHTTAo+eIP7gxPZj5fbuuZ6I3vkZTEWM+SY7vk1sCp3wGn/QB0P9Hz35jdWJWWQ4C0fkB0HDD8ZtfF3oNdWAotvlGQUDxSmEeVDTDh0dQ/vd+8IiIiIhIx+rRzZxJaPa7qO5NuT05B0MqNmVliLxGtz+El+/ML8dx3y1znuR03Th5YYameVP39xZ5NuK0egoSzV+9EzgF3MsX4AA4sqWx4yYHCEmwNQlC/vnw4a53Hd1G2RrhhsqZ9W6/5v04a4DEc5H8z1tQsu5Q9+5q7ewyXJrcH+l/qPNPuSODM6cBp3wNtDq20T6R9kMzkQzpVGYSLjYnGId3cgc3567NMsL1clmOnYz0vYyD863OBqZe7swW99TjFmYl4zmygibsfovdkYwUJfadPIfGgkmMRERGRhqFLq8ZIsGVh1XeQcF9+oUcfwPouN46LiUZ7WwZYfQ4veWHqcvP4LWce3s1U9Ujd+hJur4dyY3upMfsEMsOtPtgzCa2+iJFoY1YOXp+2ynWef6PYh1CTaeHxt/Ka4/u5zrMkm2XHdemr6hhxO3DqVGeAsCzDsLIy8Lemr3GdZ7DyvKN6VHv79uxHBrmXbNxT/UYteQVY/QGw5GXg7UOAnYsqvh4DjK2GelxkDxJyG1OTG14fy9pSkFDKpQI3LWsEyz4D5SL8IiIiIhIRmLXWs629L2H9Di/ZZRtaEoxyY+rSskm9BwmZIfXTki0eGS5nH9G9Xu470oOEW/fmB7S3Jl+z89Zluc4f1adNvU3k7pTumfkaqX0JmaVmH4Zx3cQB6KIAejlj+7fF4baZAixBf+dXd/CuxlLaAZ3HMVWxyqsxa9Heluz80T18GiRzSPd0j4FMs9f4MCg1LgmIKzuQs2cl8M6hwPx/OzMMq2EvN+bfWGZgim8UJJRyi0Uryn+wqATzbc1IRURERCSy9LaVHK/fkWPWf8HoRxiMcmMr8GJh/7O8gsD25ObtP/PNEtd5fm1lmXF8rCaI15a93Di/oBg5Aeyr/tOSTI8A1rhB7VFf+Bqxl0yui8AJx/z787MtgN63fbN6fY7DCYNeDKA2SXKXHb/32zrTozBQGHj7Ys5G13lmYp8wrJNPv9skKd4kJNn7ElYb0O9zLnDefGffQeIk5mnXA5+dBHx3EbB9bqW/utk22bhDCw2DqgkFCaUcqy8h/b7Khwi/iIiIiISlPu2beQwLCOQXTG+7c0Irk5A2Znn2sfK3V39aiSxbBuWUEV089oHULZMwkMNLGNCYujDTdb51ahIGdGqO+mQvOWa5cTAmkgfSr8u3Ifege1DFxKHuHnNSXrOUBBMo9Bz4sjBg1YAv/bjCY2AOp02z36Cv7CXHO7IP+Pb3tnlP55CVYTe4L1v/JbDsDeB/w4E/Hir3K+wZujfP3e9Wk41rRkFCKWdo1xamvwZx4hqHmIiIiIhIZGcS0sp6LDm2lxvHRkehaRB6RnXxmnAcyJLjJRt346t5mzyCTBeO6Rmw+2uImYSBHF6yPHMvMm3DQsYN6uBRPlkfutqGl7Cn5W7b4J9I8O0C9/sjOSEWR/atn36P4eyovm3MycLA29u2noH+MnddlscwmWFdW+DQ7u6gny9G9HSXR9MfvpQcU2wCMOYJYMrXQFILz3/rMrHc1Tft8vw7rqElNaMgoZTDAOEh3dJdHz78QBQRERGRyGyA37Jpkuv8inocXmIvN+Z21HfAhVqmJiEp3l3qm7EzMH3eCopK8ORX7jJj+tsJA5EYXz/97CJZa+9MwgANL7FnEfKVOm5gO9S37rYgIa2NoJJjDixZttn9vfPoAe2QWJa4IlW7dkJ/j8EcH/6+Diu3+O87fElpKV6cutx1nq0xrxjXt8Z9/jqkJXtk/nJSeI10nQhcsBjodJzzPCcytxxU7mrlJxt7HgySqilIKBWyN0GdtWq7niURERGRBpBNyOEl9VXCaC83ru/JxhYGJjunNw54JuFb01djiy0L7fghHTCki1dGjNQKA0nNUxIClkl4sLDY9JH8bqF7qvGgLmlo5ZXBWB+6eQ3wiKThJd8tcD+/NGGISo19xcGjf/UoOwYe/XyROTjhD9/M3+QxCIRl4J29srB9waDiYbZsQn7eZNvKgn2S0gY47XvgqizguJcqvIp9W+Njoz0OhEn1FCSUCh3ao5XraC77EkZavwsRERERcepjCxLuyS3w6JkXSPb7CVaQkOxfdjN25vhl3cvb4LTRt2eswTUv/YoPf1/v+re0xgm4/Ng+db4PqbjkeJsfMwlZUXXVS7/ia1uZOJ0w1LdhDf6WnBjnkYkVKcNL2EPvh8XuTM1ebVPRzStrUqo2qndrHDPAnd2auTsPb/yyqs5PGycZv/nLao8y8PNH175Ngr0vIf/SzlnrnhZeI41aVDqJ2Z5J2D7Ncyq4VE/57VKhxklxGNi5ORZm7DZH47hg6qrR8yIiIiIRp7fX4IyVW7LrJfPCnkmYHoTJxhUFCdnwnoHS2gQt2cebQSVW4fAge2UZbddNGICURPdEUqk7Bs6sFkn+GFxSVFKKt6evxgez1nlMM6bTR3bFEX3cgx6DkU1ovbbW7oiMTMKZK7eb955lwtAOQd2ecHXV+H5YkLHL/A2jT2ZnmKGk/TvWfsDO/35dg/22fXPuUT2QmuzO3K0pbkujhFgzidzqS+jvCdb2TEL1I6w5ZRJKpQ63TTmepSnHIiIiIhGJfc44OMSywo+9rCpzoLAYeWVfEkMpk7CmJcfMgJq9egee/HIxzn7yR/z9jd/NF/PKAoSnjeyKkba2PuL/TMKd+w6iuKS01rfF/f/XV2bivZmeAcIWTRLx0LkjcNmxfWrci82f7Bl2nBBrD66Fq29tpcbsETqmX9ugbk84J/r87QR32TFfvo9/uciUzNfGlt15+OLPDa7zbZs3wonDO9dpG+NiojGsq3P+Ac1bt8sE5f2FJdZ8X1g02bjmlEkolWK/gOe+W2Z+/n3Vdpx3VA89WyIiIiIRJj42Bt1aN8Wqrc6hJSszs+t1sjG1CGKQsEtLz7JGVtAMKxviV1n5Had8MmOQpXIHq+n7xUyWUb1a4f/buw8oN8qrb+D/7b33ZnvtXXvXuAM2Ns1gAphqWigh2HQIIRADeXEoCS+hfQkJJBA6dvKGEkioAUMAm2Ljjg0uu7bXbXvx9l71nftoRxpptV1atf/vnDnSjEbSrGYlje7c596TclOQnRJlt+0mM/0Q3B6DQQ1l1y8baibou5sOYdXavX2CFjKM82dnH+MSGaBZyZb/Qwcq6jFrgvvWt5RA1PeHq03zp01LQwgb+ozYvOwknDkzHf/93jh8u7SmBa+u2av+f4frpc/z0KWLlEuZBAnyjdYJkxPxTV6Zut7S0YWdR2owZ6J9/oeLq5tUcFTDTMLhY5CQ+iXDTCanRGFfWT0KyhtUpzD9WToiIiIi8gy56dGmIOH+snoVJLHHj8GhDDXWsrScWfRfGl9oQ/RsZRJKUHPDvnI1ukYCGhJQGqzOo9QImz85iZksTuhwLJmcwwkSyu+cP7z/PXYW1lgsjwwJwC/OmY6Tp6bAVVjX6pPmJe4cJFy93bLeozTFoNG55cyp+O7QUdPJmPe3HMaJOcmYOSFuyI8hw5Y37Kswzc+aEKc+z+zh+KxE1SFZ+xiVIcf2ChLqhxoLBgmHj0FCR+puA9rqgNCR1wBwNjm4kSChkNoqF83LdPYmEREREZEDOxxLgFCabkjzAEc52ug6mYTakGMtSHio0ljnrbCqUQUFZdICqP2R4dozM+Nx4pQkNRrHmcOnvVFydEifoN9QG8xIxtXzn+5RGU16c7MTcee5011uX8r2xIQFoba3K6w7dziWzxot400rfcBsW/s0uFl+3gz8+vXNpmUy7Pj5m05R9QAHIydBXvjvHtO8BPRuPnOq3YbZy4mZ3PQY7C4ylrbYtL9SBTbt8fj6IKFstwyRpuFhkNCBgko+h89btwLppwCTLgQmXQBEjW4M/1iToRGSci9kSAWDhERERESeJzfNqnlJca1jg4RWw41jI0ZeCN8eJiRE4LuDR9X1I1VNuP6vX6ruoAOR2mmSESN1vI/PSlA/zMl5gTPJfNWGCQ+leUltUzue+minqimpFxzgpwIii2dnOLX24GDZhFsPGLvCFrhxh+ONeytQ39Jhml/MLEK7kZIJkpX58XfGTE2p0/fyF3kqM3Ywn+4oUmUXNGfNyrB7E1MZFq0FCSXzVzoSj0uwrA87EvrOxikxYaqcBrlR45Kvv/4a559/PlJTU9UH8HvvvTfg+mVlZbjqqqswefJk+Pr64s4774QrCy76BD49XUDhGmDtHcDLmcDfZwLrHwTKt8qpK7g6Sc9Njw1T13cV1lh8iBMRERGRZ0iKDkF0WKBpPq/EsXUJ9ZmEklXi7B9y+uYlEmjqL0Aor5EEj3535fF4664f4b5L5mDhtFQGCJ3M18dH/Q9rygbJJPw2vxw3v/B1nwDhMRkxeP7mU1RwxVUDhNZDjouONqtmDe7oY91Q46AAP5w2jQ1L7ElqCCbpOtV/tK0Q2w4ag8v9aW7rNCUJidBAfyxdOAX2Ni870WJ+4/5KuzyuPpOQTUvcMEjY3NyMmTNn4tlnnx3S+u3t7UhISMD999+v7ufquiIyYYiaaLmw6gdg48PAa8cDL2YAn98KlHwLVyVfjloHNqkZIPUCiIiIiMizyDFfji6bMN/RQUJdJqGzhxqLgbJkpLaddCX+47L5eP3OM3DneTNUBqGzA5tkSV+DsL9MQgmASO3Bh97eZpH8IFmI1y/Kwe+vmT/shifObl4ijVr0WV/uQvaRlr0rTp2agrAgZuPakwwtXn7BDItl0old3gf9eWNdgcV748qTsxATbv9M7/EJ4RZlAmTI8Wh19/SoRjga1iN0w+HGixcvVtNQTZgwAU8//bS6/uqrrw45sCiTpqHBWLOhp6dHTY4ij900fTlCTnsUvrX5wMEP4HPgA/iUm+sCoKkE+P55GIITYEg5Aa5q/uREvL3hoLq+Pq8cZ0xPg7eR/Sk1Sxz5P0Njh/vTs3B/ehbuT/u/nkRDJc02tMwqGf5V19yO6LAghzcuiXNi0xJ9Ztbk1CjsKzUO3ZS6aFJ2Z8GUZPVj1pWzyshI32DRVibhjsNH8eQHP6CyvtVieWZiBH61ZJbdh1OObfOSeou6ou7gkx1FFvNsWOIY0tTmguPH44MtR9S8dP5+8bM8/PJ8y+Ch9rn/3ubDpnkJ4i2Z65hyafKZKkOOpamK2FNUg4aWDkSGmjPah0u2X9+NmUHCkfH4moSPPfYYHnrooT7Lq6qq0NZmWQvF3gfl9fX1KrDk65sATLheTb6tFQgq/i+Cij9FUPk6+PS0ozr2JHRVmiPnfvX7EbnlXrSnn4X29LPRHe7cDk8xAQZEhfijvrVLpScXlZSpdHBvYrk/nZqAS3bA/elZuD89C/enfTU2ul92CTlPTrplkEGyCaUJh6OHG7tCJqEMV31y6Xz1N0uwKVE3RI/cgz4DsLG1E01tnQgPDlBDcWX45DubDlmsL2HfyxZMwk9PzXa7rFD5W2UYqNZspcDNmpdIxpfUvdPXBHW3IKc7uf70HGwpqFJBNC1AK92OpTmP3suf55nqeoobzsh16Htj3uREU5BQYntSZ/P0USQkWXc25nDjkfH4IOGKFSuwfPlyi0zCjIwMNWw5MjLSoT9yJDouz2MZVEoExkux0Ltg6GiCoWgNYicuklC6eZUjq+Bb8S2CKr4Ftv0GhrhpqumJQRqfJB0L+Ix9kOrEnBR8vL0Ind0GFDb64MQcyw8UT9f//iR3xP3pWbg/PQv3p30FBzs/+ELuY3JKtOoGqSVi5BXXOiRI2NXdo5pGuFKQUMiP4Rnj45y9GTRCKbpMQm04q/wr/7/3dvQJHkiQ7e4LZmLauFi3fL0lqD0xOVLVjHfHDscytFTrJi7OmeO6TWI8QXCgP+66YCbu+dsG9Z4QT330A164+VSEBRmDgD8cqca6/HLTfaaPi8VJOckO3S55DmkA1dphrKkpmeyjCRLqm5aIjHhjbwUaHo8PEgYFBanJmgR6HB3skQ+6AZ8nOBLIXtJ3eeV2y8ep3gVU74LP5keB8FRg4vlA1oVAxmmA/9gcVC3ISVZBQrFhXyVOnup9RWUH3Z/kVrg/PQv3p2fh/rQffmfRcOtXjU+IMNU3c1RdQgkO6Nv3xbvAcGNyf8lWtQRXrt2L7YeOols3/FAb1nrTj3IREujeP4WzdEHCQ5UNKjvPz01+p6zu7bgrAv19cfr0dKdujzeQgNxF8zJNGbXVje147tPduPuCGejpMeDFz/JN60q4Vjp8OzpwKydmjp2YYApOSiahnETy9xvZ/7H+ZICcfGKNy5Fxj08Rb3Pem8Cy3cBJjwLWtQqbSoEfXgDeOQf45Nox26RZmfHqwFE78yNvXiIiIiLyLLnp5uYle0vr+gRY7F2P0JUyCclzahJqAQf9/29seBAevuJ43HHudLcPEFrXJezo6lFdjt2B1ISUfaM5OTcFESFsWDIWlp02Bemx5uy6L3aWYMPeCqzbX40DFeZs1B/NTFd1WceCDDnWNLd3YVeRMfA9EuxsbB8MEroiidjHTQXmrQCu2gDcUgb86EVg4nmAny4rMtOq6UtXG7DtT0DdAbtvknT8mptlfANLfY+dvWetiIiIiMhz6OuCyRCwwqpGh3Y2FnEMEpIdSEJDVD9ND07OTcbzN5/SpwabO5uUFNWneYk7+O+OIlNJA7F4jnPr73sT6Stw94UzVVkJzV9W78K/tpSY5oMD/FQwcaxIjEGfrzjSLsfSO6BYFyhn0xI3DRI2NTVhx44dahKHDh1S1wsLC031BK+55hqL+2jry32l+Yhc37NnDzxaWDIw40bgog+B26qBC94FjlkGZJ5juV7hF8CXy4FXsoBV04Bvfg2UbQIM9sn6kw5vmm/3musVEBEREZHndDjWy3PAkGN90xLB4cbkiOYlIizIH/+zZBbuu2ROvwFEdyVdtyWRQ1OgywRzVZLZqe9qnBEXhmkZ5uxlGpts8UvnTzLN1zZ3qAalmitOyhrTEzfRYUEWJ6c27RtZkFCGT2uNfASblrhpkHDr1q2YPXu2moQ0GJHrDz74oJovKyszBQw12vrbtm3D66+/rq6fc45VsMyTBYQZ6xievRIIjbe8reB98/Xq3cDmx4DXTwBeSAP+exNw8CNjtuEIHZeVYPoi+nZvhYrWExEREZHnSI8PV4EVTX5JrUOHG0vWiv75iEbjjBnmpgezM+NV9qA0QvDEphhSt00ChRp3aF6y7UAVqnSZxJJF6In7xtVJR29bmXbS1f3ieZljvj3zdA2ySmqaUVxt2YBkKKybEzGTcOSc+o28cOHCAQNNq1at6rOMgakBHHc3ED0JOPABULpBXi3j8uZyYOdLxkmCjNNvBE7707D3lxT+nJUZp9qnyzCR/WX1mJzKVvVEREREnkK6pk5Ji8Z3B4+q+bxi+2cS6oMEUo+QQQKyl/OOHY+JSZHqf0qyk+T/2ZNlJUehoDc4KMON5beyK7+fPtY1LJHkkzNmsGGJM0jDkHsunIk7Xv0WPbp4zPWLctSQ5LE2LzsRq9buNc1v3FeJS+f3DWIOpOioZWkMBglHjjUJPUnsZGDu/wBXrjfWMTzzZWDSBYB/iHmdzmbA10Zh2IYjQ3qKBVPMbdAlm5CIiIiIPIt+6JdkZ0g9akdlEnKoMdmTBMiOyYjF1PQYjw8QWjcvaWrrQkV9K1yVvO/19eZOzEn2uCHg7kSSfa46Ocs0P2N8LE6dmuKUbclMjFBZjJpN+ytGlUkYHuyP6DD+b40Ug4SeKiwJmH49sOR94GdHgQvfB6ZdB4QkAFkXWq5bfxh4aQKwcirw9b3GLMR+6hieMNlcWJR1CYmIiIg8T26aZY0w6XLsqJqEbFpCZJ8goasPOf5UNSwxZ60tnp3h1O0h4OpTsnHPBTNw+dw0/PbHxzotC1WeV7IJNbsKa9HY2jmqzsaunFHr6hgk9AYBoUDWBcBZrxgzDFPnW95+4EPjZU0esOUJ4I0FwPMpwKc3GG/rNJ+Rig0PVsVOxZGqJpRUmzsIEREREZFnZRKKfDsOOZbhkEethhsT0cioodW6+QIX7XAswUEJEmpSY0MxY0KcU7eJjME5qdl5zsxkhAQ6tzasPkgo/y9Sv3I4itjZ2G4YJPQ2vn6Aj9VuD08B0k6SjwnzspZKYNcrwHsXAH+NA95bAuS9pm5akMMux0RERESeKjI0EGmxYQ5pXiLZIZ3d5hErHG5MNHIS2EmLC3P5TMLth46ivM6ceHL2rHFeMRychm7mhDjVyEqzcRhDjuV7pba53TTPzsajwyAhAZMvBa74Bri1AjjrVSBriWUdw65W4MD75iAh6xISEREReU02YV5Jnd2aB+qHGgtmEhKNvnmJqwcJV+salvj5+uDMmWxYQn2bqcyZGG+al2ap3T22S6BZK2TTErtikJDMQhOAadcCF74L/KwaWPIBMP0GILQ3c1CaoADqzPKEhAj4oQs3VV2H1s9+Cez7F1CdD/R08RUlIiIicnO56dEWWRqlNS12eVz9UGMRF8nhxkT2qksoQfg6XUaVK6htardoeDl/chJiwoOcuk3kmvRDjqVh1p6ioWWxF+nqEYpx8RF23zZv4tyB5+S6AkKASecbJ2liUrYJiM423bxgShJ2V69Hrt9e4AeZnjLe4BcIxOYC8dOAuGlA/DHG65ETpOiB8/4eIiIiIhqyHKvmJXkltRbDGkeKmYREjm9ecuykBJd5mT/7oRjdPbqGJXPGOXV7yHXN1QUJxcb9lZg+Pm5YTUsC/X0tOiXT8DGTkAYnNQyl2UmoOf13QU4yMn0Podtg9S/U3QFUfW8cmrxuhbGm4auTgR6r7kT1h4CmUqlezT1ARERENj3yyCNYsGABQkNDER1t2UyDHCszMQJB/ubjvPwS+zQvqdYNN5aaZNFhzCgistdwY1HgQkOOpUzBJ9vNDUuSokIshpQS6UmT1Mmp5v/nTfsqhp1JmB4Xroa008gxk5BGJCs5Eg+FXo41DQsxy3837jnJH4F1eUD1bqBmL2DoNq8cM9mYYaj3zQpg7z+B4BhzxqG6lAzEYywCkkREROSdOjo6cNlll2H+/Pl45ZVXnL05XsXfzxfZqdHYVVij5vOKa+0+3Dg2Iog/5ohGKSo0UDUA0t5bB1yow/EPR2pQUtNsmj97dgYbltCATshOwr5S4/9wUXWz+v/RN9IaLJNwXHw4X+FRYpCQRtwuXRqYvL+lDV93LcCJsbOx8ORU441d7UDtPuDoLqB6FxBk48y/3CbaaoGSb4yTntRBlIDhjJuAKT/mXiIiIvJCDz30kLpctWqVszfFK+WmmYOEBysa0dbZbdF9crTDjRMiWI+QyB6ykiJ1QULXyST8WNewRDKHz5yZ4dTtIfeoS/j3r/aZ5jftr8TF8zL7Xb+9sxsVus7Z7Gw8egwS0ohJXcL3txxW19/+9oBqZjIhMQLwDwISphun/kgTlIgMY7Cwqbjv7S0VQGEFMOlCy+USVPz4amMAUcs6lBqIUkORiIiIvF57e7uaNA0Nxh/MPT09apLhb3JJg5uiG/bVYzBgX2ktpmXE2i1IKJmEI9kX3I/uj/vQviYmRar6bUIyr5rbOhAS6O/U/djQ0oF1+WWm+blZCYgND+Tnr4txtfdiZmI44iKCUN1o/B7fuK8CS44f3+/6hVWN0Bcwy4gLdZm/xVn7cbR/P4OENGLTx8ciOiwQdc0dqvbFrS9+jTNmpOOahZOREDlI0O7kR83X2+qMw5RlkqChmnYCrUeNgUC9o7uBQx8bJ33NxOhJ5uHKpuBhDuA7urPdRERE5F4ee+wxUwaiXlVVFVpaWlBfX68OpH19WZp7MPFBXRbzW/cWI9Fq2XBV1Zu7JIf6GVBZaQxsDIf8AOJ+dG/ch/YVH2IOk8i17/YWIjsp3Kn78ZOdFejqNm/X/ImRI3q/k/e9F2ekRWBtvjFIuOtINQ4XlyE00Pbv+l0HjdnumjCfDq/8P+vR7cfmZvMQ/5FgkJBGzM/XF/dfMgePv7dDpbdL06r/fl+ML3eXYsncTFx+4iSEBwcM/kDB0UDaicZJr6USCLTs1oWaPX3vL92Xa/cbp4J3zYHD25sAX12wUm6HDxCVyeAhERGRk9x777144oknBlwnLy8POTk5I3r8FStWYPny5RaZhBkZGUhISEB4eLgqmSLXXeXHkCuTPpMJkftR1TuMsbi+C4mJlt0nh0OGhTW3m+tWj0uKGdHjyY8h7kf3xn1oX3OCIoDPDpjma9r9RvVeHe1+lEDFuoJ803x8RDAWzcliDVIX5IrvxYUzgLX5R9V1iTMXNgCnTLX9/1yfZ26qJf1KpmWlI9Dfz6v3Y1OTuUbjSDBISKMiLclf/dlCNez4n+sL0NTWhY6uHrz17QFVg+LKk7JwwfHjR/ZGDbXxQTD9RmDCYmOtQ8kqVJdS+3AP0GWuRYDorL5DkL/9LZD/OuAfAsRN7c041GUeRqRLscURvApEREQ0VHfddReWLVs24DoTJ04c8QsaFBSkJmvy40cmOYjWrtPgctJiUNVgHDK4t6R+VK9bTbPuWE0CB5EhI3487kf3x31oP8nRoSo5o6mt01RDdKw+42ztR6llWnTUnM101qwMBHhh4MZduNp7cfbEBAT5+6K9yzhsdnNBFRZOS7O5bnG1+f8sJSYMwYFDSFLyUD522o8MEtKoBQX44ccLJqluVf9cfwDvbz6Mzu4e9SX10ud5KoC4dOFknD49bfTdrOT+kRnGKXOxeXlPN1B/qDdguNsYCLQmAUUhwcSKbcZJT7IWVbOUm4FjrhnddhIREZFNcpZbJnIPuenR+CavzFRPsKqhdfCyMkPobCykIysR2Sc4kJUciR2Hq9V8gZM7HK/ebm5YIr/+5Hci0XDiC7Mz4011NrcUVKK7x2AzE1Xf2ZhNS+yDQUKym8iQQNx4Ri4uOG686kj0xQ8lqiZGZX0rfv/+9/j3xkO47vQpOG5SgvoisyupPRiTZZyyl9heZ/JlQNQkY7CwtqC3YodORwNQ+i2QfYnl8vYG+Ly/BBHB44H4CUBIHBAc23cKiuIwZiIiIjsqLCxETU2Nuuzu7saOHTvU8qysLDV0mBwvJy3aYj6/uA4JU0cWJKzWNS3RhiASkX1M0gUJj1Q1oau7B/5+Y58Z1tjaia/3mBuWHJeVgMQoNpmk4Zk3OckUJGxo7UR+SS2OsWqc1d3TgxJdJuG4eB4X2AODhGR3SdGhuOfCWbjkhIl4dU0+thRUqeUHKxpw/xtbMGtCHG44IxfZKeaOeWPihPvN1ztbgJp8c+ah1jClsbBvs5Tq3fApWoswuS5lDfvlA/zsKBCi+/AqeB848pmNoGKM5XW/QHv/tURERG7vwQcfxN/+9jfT/OzZs9Xl2rVrsXDhQidumffISo6Cv68PuqT4tNSLLKnFyVNTRt3ZWMQxSEhk1/eqRkZ1SaBQAodjbc2uElV+SnPO7HFjvg3k/uZmWZYe27ivsk+QsKy2xfTdJBgktA8GCclhJiZF4ndXzsWOw0fxyuf52FdmTHuXM1w/f3kdFh6TqoYhp8aq8NvYCggFkuYYJ732esDP6qx27b6hP65kE+qVrAN2PDv4/cb/CLj0v5bLNj1mzG60lbWoTdZ1F4mIiDzIqlWr1ETOHfY1MTkS+0qNx3H5JeYi8aMZbhwREqAem4jswzogeKCifsyDhNKwZPV35qHGseFBmJvt+AYq5HmkHIUMoS8ob1Dzm/ZX4PpFOf0ONRYcbmwfDBKSw82aEI+nrz9RpZ2vWrtXRfyFdEFel1eGc48dj6tOzkJ0WN8i42POOsgnjlmKngmLUXtwM2JCDPDtqAPaaoDWGuOlNvV09B1uLMuHwj+077JdrwJ1BYPcLxhY+Cdg5i2656wFNj1qO6gYogUXw9mkhYiIiIYkNy3GFCTcX1avspQCRjCMUZ9JyKHGRPaVHhdu0ezhgARXZo7tq7y3tA6HKhtN82fOTHfKkGfyDCdMTjIFCSUztry2Bckx5t/NRX2ChE5IPvJADBLSmJCGJZI5eGJOsup6/NrX+1Hf0qHSg6WxyWffF+OyBRNx8bxMBAe64L9lSDw6E44DEhOlPeLQ73fi74CZt5oDia3VloFFbYo/pu99hxJg7Grrm/nYVAps/cPA9/P1NwYLf7IFiNQNAShZDxz53Nj4RQKQMvnpL4OMl3LfhOmWj9nZDPj4AX5BDEASEen1dAFtvSeY2muNl4FRQNoCvk7kNnUJ399ivC7DCA9VNGByqmWtwuHWJGTTEiL7kqYOmUmRpmxfLbgyllZ/V2QxfzaHGtMozMtOxD++Ntf7kmzCC+dm2swklBNPYUHe29nYnlwwGkOeTM46X3j8BJwxIw3/2nBQNTNp7+xGS0cX/vblPny49QiuPiVbdcDyc5EW7KMSnmKcRuKyNUCbLqhonbmoTeGpww8uyg/WlkogMMJyefFXwIbfDn7/5OOBn2y2XPavs4DS9cbrEiiUSR9kVNeDgOk3AjNuNN+vqx1Y83Pjetp9+gQne6eM04yNYzTtDUBzueV91HMHOj5QaegBujsBQ5fxUl7THqvLsGQgKNJye6Wrtq11rS+nX29Zq7JwLVD4xeD3i54EnPSI5bZ+/zxQf9j4vxjWO8n/jVzK0HsiGtp7vqOx97NXAn21QHsdMNmq2dUPLwJ7/9m7Tu+6UjrCWuZi4OKP+cqTW8hNj7GYzyupG1GQUD/cmPUIiexPhhdrQcKD5Q3oMRhUssZYaG7vxNrdpab5ORPjkaLL+iIarqyUKDVkvaapXc1LIxN9kLDoqLlpCYca2w+DhOQUEuVfunAKzj9uvDo7IGed5EtMPgD+/PEuvLNJOiHnYMGUJPt3QnYXiSMcH5AwA7j8q/6Diqaptu/warnPUFhnL4puXTHy7nbjZOuH8cTzLee7WoCdLw/tea/cYBkkPLQa+OgK2+vqA4wh8cDSnRY3h+Y9D5+vvjIG+foLuEnwb8plwMI/Wj72s7HG128w574B5Oi2ryYPePv0of2tcj99kLDkG2CTVfCvvwCudZBQAhZFX9pePzCyN2iYAky9Bph2reXtNfuMwU4JKHvre5E8h8EAdLXqAn29l/K+iUgzr1f5PfD1r3qz/nrXk4CgBAqt3dlu+V6tPwQUrhl8W4ZajoLIBSRHhyAqNFCNAhH5xbXqpO9wdPcYj/M0CWxaQuTQ5iWShCFlntLGqP77l7tKVfKHZjGzCGmUJMAt2YSrtxszVH84XI2W9i6EBvmr+pf64cZsWmI/DBKSU8WGB+MX50zHRXMzsXLtXqzPL1fLi6ub8b9vb0NuejRuWJSLaeMsOxnRACTwl37KyF6imTcD4043Bvy6eicV8NPPtwGR4/veV35kSyBJv56t+0vQTk+WDZX1feWx+6M9n6zSYz5g0QTU7IJPcT+BM73WozYWDjHLVQKNFncbxkdun/sOMX1eApvWmsr6X18CuTLV7gXGLbK6rRFYOcV4PSDMnIWoz0TUshNT5vXNTCVylO4OczafNoQ3NNH4OaT3waVAc1lv1l/vunJfa+f9E5jyY93jtwFHrJpJ9UceOyzJPC/lGPRlHYJidB3tY8zLJOuXyE3ICdvctGiVxaFlEg5XXXO7OiGsiYu0ccKRiOzbvKS8YcyChFogR8hJhflTdN+NRCM0LzvJ9L8lpcq2HazCybkpqG5sV4FwDTMJ7YdBQnIJ8qZ+8LJjsae4Fi9/nofdRcYsrbziOtz1tw2qaOl1p0/B+AQGIRwqJts4jcQZzw2+jvpxYP6BoGhZfn2CizYCleHplveNmgBM/al5nf7ur/1o1/HpsQowqnqKAYCPv+Wl1A2zJoEICaxJEEBNAbYvI62yLMJSgbn39r++/lKay1hnFibPHfw5pZ6ktSXvA43FQHOpMWDYrJukhqVcSk1JCfb1F1yU26WRTn/NdK753pjFqjn8GfDDCzYCir3XJSOUmYnUKVl91b31WquBzhYg8xzL12XrH4EDHxiDfJLtLJfy/2gt50rg3Nctl0kJBClJMBjrzGDTZ4YPEBzdG+jrDfLJdWkCpS2zPnkx62fGWrQSWOf/OHmQnPQYU5BQspMk6DecpnNVuqHGgo1LiOwvMzFCZV9pAfmC8nqcMnWEpY+GoaCsXjU10jcsGUlzIyJrsyfGI9DfV9XDFZv2V6ogoXVnY2YS2g+DhORSpqbH4Mml89Wb/5Uv8k1v/o37KrB5fwXOnJWBn54ymcWu3ZX6wWw1ZFUCcfHTRvZ4kjE5wqzJuvlPI/GCN+ArzVgkuDacH/OXrB7Rc6og2cmPjey+UZnGaSRipxingUjWoARK9eR1ybnKMqBoawi5sA4wVn0P7P93/88nQU25T9KxwIXvWN5Wsd38eoUk9O0aTq5HsnVVVl9vwC9uqjG4pinfAmz5vfl2LSgoQ3+tO73fbu6KqNTtN9ZLHYytEgASxJMgoQTd9dl8+qCfTNYZiFETgdtqjJnZPsP8kSPBQSIPJJmEelL3TE7ijqRpiWCQkMj+Av39VLDkcFWjucPxGFi9w7phScaYPC95vuAAP8yaEIfNBVVqfvP+SlW+ouio5fEiOxvbD4OE5JJDWuSg8/isBNX1+O9f7VPpxD0G4JPtRVi7swRL5mXi8gWTEBbMDkY0QhKMkKYdntAgxx5sDRWOngic+5rlMsn0UgFDLRuxNztRXytSyPKBSN3HxsK+jXfE2juMNRiFBC5lKKca5pxqHt4s90tdMPIAMw1Qs6/FGMSTOntSj9K0z7qBr+62zPzTLqVzrz5L+NLPgPFnmOdlmO++twd/1eW5rUsQaFl9sj2m4bu6AJ+2LC637+Ndsd4YtJOTEcMhgWl5bCIykUYlPrp3+nCDhEetgoQcbkzkuCHHYxkkbOvsxtpd5uO+GeNjkR5nNSKGaBTmTU4yBQmlNu6+0jqLTMLwYH/EDCOznQbGICG5LOlufPbscVg4LQ3vbz6EN9cfUIVK27t68M/1B7D6u0JceXI2zjt2nDprRkRjQAKrUkttsHpqJz0GzP6FLguxzPZw53BdswhbAUZDtzGDUSZss3qORy2DhNLB+f9mGodc2+ySHWK+ftpTlp3Hpeu0DJG2vp/+PnIpjV6sg1FSP1KCma42tFSCepIhqs/oE/v+bfx7bQX75FKr9TnrNmDRM+b7SUbd988NXAtUI4+jpx/yL1mkElQOjrN9aV2SYN6vgXn3GffFcF9j67+diEZMCsVPSIzAoUpj8CGvZAgNvPrpbCxDxyJ4opfIIbKSI/HFzhJ1vba5XWXxOrKb+OaDtWjtYMMScpy5WYkW8zLKUB8klNJlXtvs1AEYJCS3SDG+/MQs1SHrjXUF+HDrEXR296ChtRMv/HcP3tt8CMsWTsHCaamqBgcRuQAZxi01I2UaiK6Ivcm0640dYvVBxZaKvp1lrbMQZehqZ5NxGswpT1jOl6wH1q0Y/H5Ss/O6fZbL3r8IOPiROcAowSyL4GTvlH0pMOtWy/t+fa+5lqT1+vrAZsJMy2zN2gJjYw1bQT7tUrrxShD25mLL55SMPul4PdxAn3y+yjaogG0vGY5rK9hnHUSOnwHccMh4uwz9Heizukf2s24ICYfvErmMnLRoU5BwX0m9GvLl5+sz7OHG8ZHB/EFH5CCTdB2OtWxCRwYJv8w3N/mLCAnASbm6UQhEdpAYFYJJSZE4UGHMjJXSZHXN5kZ0rEdoXwwSktuIDA3EzWdOxYVzJ+DvX+7Dmp0lKt+koq4VT7y3A//eeBDXLcrBsRMTnL2pRDRUtoJF81bYzoprqbTMTEw7yeqxfIG4Y3qb1rSaG9eoundWwciRdtmWwJ01dV9D73PKc/WTXZMwq2+AdItVsLI/F68GMs82z1dsBb64bfD7ScDQmgTxrPkF9Q30pZ7Qd70lHxoDl2qd2KF37NaCxkTk1nLTY0xdJqWrZNHRJpVdONzhxqxHSDR2HY6lecncbMtMLHuRkwYHKs3NxM6Ykc4RXuQQ87ITTUFC7WSVhp2N7YtBQnI7ydGh+NWSWbjkhEzV3GTbQePZq4LyBvz6tc2YMzEe152eg+wUG11picg9SY04GR6sHyJsTZqzLNvVd7kE42RIsATwtK7X0hBFL/ti4/214KJFd23dZaiN+ltyv476/u8n9RdHE5i0dV9bgT7TbTGWwb7uTsuafJLNmH2ReR2ZpEbnUDKxk+YMfZuJyONMSbUcwi9DjoccJNQNN3ZkVhORtwsPDkBydAjK61odXpdQ6sXrLWbDEnJgXcLX1xXYvI2ZhPbFICG5dSr9oz+Zh+8OHsUrX+SpIKGQ+e8OrsNp01LVMOTkmFBnbyoROZMEvyRIpgJllmfXTWKyjNNI6Ov22SJZkKqOn3Vn70Dg8m90gcVWG0HG3qBlpFUWXsJ04KyVfbP/JEA4WDdoqePIhi9ENALjEsJVbUKpES3yi+tUOZjBGAwGZhISjfHvJFOQsDf7yt7aO7uxZpex9qE4JiMG4xOGdtKAaLgmp0ap5iRSZ9PauHj+39kTg4Tk9iRzcFbmSfhqdylWrd1r+kKULlvf7CnDeceNx1UnZyMqNNDZm0pE3kiCdr6htpenWw2ZHirpOjxt2ag3jYhoOKT2s2QTbj90dFjNS5raulRAQV+TkIgc27xkfX65ul5W24Lmtk6E2blZ0Dd5Zeq9rRnKCQOi0Xz/zM1OwKc7LGttSyMsqVlI9uNrx8cicuqHxmnT0vDSrafiljOnIjLE+CXY1WPAe5sPY9kza1XTkzbdASoRERERDU9umnnIcWFVkwo+DKdpieBwY6KxrUtor2xCySKWxIzH3tmOZ1abS7yEBfnj5KkDlIQhsoN52X3L/qTHhQ+5gRYNDTMJyaME+vvhonmZOHNmOt7ecBDvbDyI9q4e9YUmWYYfbj2Mn546Wd3u58sYOREREdFw5KSbg4TSEmpvab0a1THUpiUigZmERA6VZdXhWMoyzRg/QD3jAdS3dGDjvgqVmShlnTq7e/qsc/q0VAQHDFLuhGiU5LsmwM/X4n+Q9Qjtj0FC8kiSTr/stCk4/7jx+L+v9uHTHUXoMciZ7HY89Z+deGfjIVx7+hTMn5wEn6EU6yciIiIi5KTFWLwK+SW1gwcJG4ylYDTMJCRyrNjwIESHBaKuuUPNHyivH9b9K+tbsWFvOdbvrcDOI9Xqd1R/IoL9VUNJIkcLCfTHjAlx2HagyrSMnY3tj0FC8mhyEHrneTNw8bxMvLpmLzbsq1DLC4824aG3tqkCu9cvysExGbHO3lQiIiIilyc1nlNjQ1Fa06Lm80rqBr3P0UZzoXkZFSYBDCJyHEmCkOYlWjBlKB2Oi442qWzB9XvLsa904KBiUIAfjpuUgAVTEjEp2hdJ0WwUSWPjhOxEiyAhMwntj0FC8grjEiLw28uPw+6iGrz8eT72FBsLbe8uqsXyVRuwYEoSrj09hx8yRERERIPITYsxBQnzi2tV9+KBRmboaxLGhAex5AvRGMhKijQFU45UNaGjq1uVZtLI+1aGIavAYH65SqIYSHhwAE6YnIgTc5IxZ2KCGl7c09ODyspKh/8tRBr5/1u5dq8qJyb/gzPGM9nH3hgkJK8iGYN/XDYfG/ZW4NU1+SiqblbLv91bgY37KnH27AxcfUo2h8EQERER9SMnLRpf7CxR1xtaO1Fa24K02LAhDTfmUGOisW9e0mMw4HBlo8ou3FNUg3X55er3jwwrHohk/UpQZsGUZBWM8fdjTXdyLvkOefSquSqwvSAnGdFhzEy3NwYJyevImW75QJk3OVG1UJeahTVN7erL8+PvCtVBrwxPvmzBRIQFGbskExEREZFRbrpVXcLi2oGDhLrhxvERwXwZiZzQvOSvn+xWAX1pRDIQKSdw4pRkFRyckhYNX9ZvJxf8DrL+HiL7YZCQvJZ0Nz5nzjjVjeudTYfw9rcH0dLRhfbObryxrkAFDK86OQvnHjsefuxtQkRERKRkJkYg0N8XHV09prqEi2akD2m4cTw7GxONiZTYUIQG+qvfN4PVD52YFKmCgidOScKExAg2diTyYgwSktcLDvTHVSdnq4ChBAf/s/UIunoM6izbc5/uwXubD+OaU7ORE2+u4UFERETkrWTIYXZKlKrtLPIHCD5IHTR95hIzCYnGhmQAZiZFmN6nepL/MDUjRg0jluBgSgwbjxCREYOERL2knsGtZx2DJXMzsWrtXny5u1QtL6ttwRPvfY/xcSGYPakGseHBiAkPRExYkKrTIfeLDgtkEW4iIiLyGjLUSws+HKxoQFtntyoib61aN9RYsCYh0dhZND3N9D718/XBrMx4lS04f0qS+k1DRGSNQUIiK3ImbcXFs3Hp/Il4+Ys87DhUrZYfqW7FkerDNl8vORsXFWYMHErXPn0AUS61ZXIZERLA2h5ERETk9s1LNN09BhSU1WPauL5dJo/qhhoLDjcmGjtSNiktLgwtbV2YMSFOdSgmIhoIg4RE/ZBhNI//ZB6+O3hUBQsPVjT2+1oZANQ1d6jpUGX/62ln8STzUB9IlOChRVCxd1lYkD9rghAREZHLyU2zLBqfV1JrO0io62wsONyYaGzNmhDPl5yIhoxBQqJBOiEfOykBsybEYteBYvgEhaG+uRM1ze2oa2pXl7VN7ajVLpvaVT3DgcjZdhl6Yz38xpYAP19TJqJlALFv1qLUViQiIiIaC5IRKNPRBmOmYH6x7bqEfTIJ2d2YiIjIZTGqQDTEYGFSZBASE2Ph6+vb73oGgwFNbV2obWrrDSR22AwkynXJOuwxDBxQ7OzuQUV9q5oGExLo1ycT0frSmK0YiEB/NmEhIiKi0clNi8Y3DeXqen6p7SCh/qRoeLA/T2oSERG5MKcGCb/++mv8/ve/x7Zt21BWVoZ3330XS5YsGfA+X375JZYvX47du3cjIyMD999/P5YtWzZm20w0WDBRag7KNC4hYsB1JUDY0NKhgobWmYkSQKzRBRT1XQH709rRjdaOFtVoZTBykG47gGi8TI0NQ3J0CJuxEBERUb9y0mLwTZ4xSCgZhVUNrUiIDOl3uDGblhAREbk2pwYJm5ubMXPmTFx33XW4+OKLB13/0KFDOPfcc3HLLbfgtddewxdffIEbbrgBKSkpOOuss8Zkm4nsxddHahMaA3OZg6zb1d2jAoVa0FAfQLS+lEzGwcg6MhVVNw841FmauKTHhSE9Lrz30ng9KjRwBH8xEREReZLcdHPzEpFfUtc3SKgbbsyhxkRERK7NqUHCxYsXq2monn/+eWRmZuLJJ59U87m5uVi3bh3+9Kc/9RskbG9vV5OmoaFBXfb09KjJUeSxZeipI5+Dxo6z96evDxCjmp1IcG7gDMWOrm6ViWgKHlpcl2BiR+9w53aVfTjQUOfCo01qAiosbpNMyfTYMNUtTS61AKIEFd1hKLOz9yfZF/enZ+H+tP/rSeQoWclRqiGb1FvWgoQn56b0O9yYnY2JiIhcm1vVJNywYQPOOOMMi2USHLzzzjv7vc9jjz2Ghx56qM/yqqoqtLVZFlK290F5fX29CkQMVMOO3IM77s+YACAm2geIDgIgU19tnd2ob+1CQ2sn6ls61fWa5g6U17ehvK4NFQ3t6OzuWzexsbUTeSV1atLz8QHiwwORHBWMlOhgdZkcFYSUqGDEhAW4TKdmd9yf1D/uT8/C/WlfjY2Ndn5EIrOgAD9MSorEvrJ6NZ9XXNuntEq1LpOQw42JiIhcm1sFCcvLy5GUlGSxTOYlO7C1tRUhIZbDG8SKFStUDUONrCu1DBMSEhAZGenQHzkSEJHnYRDC/Xnr/pTMAKkvVFzdjJKaZnWpTdbdCoX0Yalq7FDTzmJj1q4mOMAPaTayD2VZaNDYfhR56/70VNyfnoX7076Cg4Pt/IhElnLSo01Bwv1l9apEir+f8btVRi1oWYaCw42JiIhcm1sFCUciKChITdYkMODo4IAEIcbieWhseOP+lD81NTZcTdbaOrpU4LDIFDhsQknv9ZaOLptZiwcqGtRkTZqlaPUOMyRw2Hvdkc1TvHF/ejLuT8/C/Wk//IwjR8tNi8EHW46o6x1dPThU2YjslKg+Q40FhxsTERG5NrcKEiYnJ6OiwrI2msxLRqCtLEIicpzgQH9MSo5Sk54M4ZXGKlrgsNiUgdiE8tpWNfTImqwv0w9HaiyW+/v6qE7LWsZhRrzWQIXNU4iIiFxBTppl8xIZcqwFCWU0gh4zCYmIiFybWwUJ58+fj48//thi2WeffaaWE5HrZABJzSGZZk6I69MMpUyChrrAoTZ8Wbo3W+vqMeiap8Bm8xTrzsupse7RPIWIiMgTSNOyqNBA0/e4NC+54Hjjbfp6hII1CYmIiFybU4OETU1NKCgoMM0fOnQIO3bsQGxsLMaNG6fqCZaUlODvf/+7uv2WW27BM888g1/96le47rrrsGbNGrz11lv46KOPnPhXENFQBfj5YlxChJqsNbR2mIYrF/UGD2VehjRLcHGozVOkE3RiVEif4KFcSgaDqzRPISIi8gTyvSrZhJv2V6r5vBJz85KjDW0WxwASTCQiIiLX5dQg4datW3HaaaeZ5rUGI0uXLsWqVatQVlaGwsJC0+2ZmZkqIPjLX/4STz/9NNLT0/Hyyy+rDsdE5N4iQwIRmR6I3PSYvs1T6lvNgUNVB9F4Xf/jQyP10cvrWtW09UBVn+YpEiyUIcwxwcCk1A4kx4QhKToECZHBDqt/SERE5Mn0QcLSmhaVVSgBQX2Ts7iIIJ6oIyIicnFODRIuXLhQ1S/rjwQKbd1n+/btDt4yInIVfr4+SI4JVdPxWX2bp6jhyn2GLzehtaPbZvOUgvIGNSnflZlu85Vux1HBSIoKQVJ0KJKjQ9V1aZ4i8zJESraFiIiILFmf4MsvqcW87CSrICE7bRMREbk6t6pJSERk3TwlKyVKTQM2T9E1USmvbVHZhtakoUpFXauaYNVARUiAUIYxG4OIchlqCiDKfGw4g4hEROSdJqdGQU6jaV+v+cV1KkhYrcv4T4hkk0EiIiJXxyAhEXld85SS6ibsOVSKdgSisr4NFXUtqKg3DlFuauu0+Zgy7LmstkVNtkgn5sTe4KExiChZiMbrchkTHqSyFYmIiDxNWFAAxidE4HBVo5rX6gVbDzcmIiIi18YgIRF5X/OU+HAE90QjMTERvlZ1CJvbOlWwsKK+xZhZKMHDWi2I2IKW9i6bjyudmKUOk0z9Pa9kIpqyD3szErVAYkwYazUREZH7ykmPNgUJ95bUqZNu+tIf8cwkJCIicnkMEhIR6YQFB2BSskyRNl8X+dEjmYcqkKjLQDQua7FZC9GUwVhjbLxiS6C/FkTUMhG1LETjMikAz87MRETkys1LPtlepK63dHRh+8GjFrfHsyYhERGRy2OQkIhoGMKDAxCeHIVJyZZ1ELVaiI0qiKgFDXUZiarjcotqnmJLR1dPb+1E20HEIH9fU/1DramKfj4yJIBBRCIicprcNMvmJd/kmZuDCQ43JiIicn0MEhIR2Ylk+kWGBKop26qZihZEbGg1ZiKqoKEpgKgFFFvR3k8Qsb2rB4VHm9RkS3CAnzEDUXVmNmciSqF4uS0owE9lK2qXMvyZmYljS/a/ZJRKtql05laXncbrctlquuxd1nu7tlyWdRuAiOAARIYG9P6vBSBCLkON12VZRGgAQgP9uX+JaExlxIerzx7JIhSbCyotbmfjEiIiItfHICER0RiRoJwMG5Zpcmq0zSBSfUuHxVBm05Dm3rqIknFoiwSRjlQ1qWkofH1kiLMueOjvh0AVTNRd9/e1XEctk9uM6/Rdbnkf/WNJYxd3CUpKp+sOLVhnEbzrQnuHVTDPOsgnt3f2H+Sz1VnbEeT1luBhhAQO9QHE3nn5H1TXewONkb3zflY1OomIhsrP1weT06Kw41C1mteX35BP/9hwNi4hIiJydQwSEhG5CAmiRYcFqUlqO9kKItY1d6ghzOVWGYhadqJkqg2FBKtU4KqfzEV70wclLQOT+qCkMbDYN3jpq7Ih1XKr+8hyf1+gvq4NDT316Ogy6IJy8vf1Zuyp4F1v4M4iyGcVzOu97u6kkU5tc7uahiMsyN8UMLQMIFpmKmq3SbBR9pW7BICJyPFDjrUgoZ58r/n78SQEERGRq2OQkIjITUggJiY8SE05VrWftAy42qZ2i6BhdVObGsLc3tmDji657FZDlyVTrr2rW2UmGm83Xh9qkHG4xjoo6SpZNSGBfggO9FfBzBDTpQRB/Xtv80NIgL+6NK8nwU/z7XJdraeGEAONrZ1obO1AQ0snGuRS5ls6TNflslF3W39D2G1pbu9SU1nt0P9OGbquDX82BRf1gUYbt0ltT3l9iMiz2DrBJeIjg8d8W4iIiGj4GCQkIvIQvj4+iIsIVtPU9L5BxKHo7jGgU4KJFsFDq8BiZ48K9vVZbhVw1N9X/1jm2xwXlBwOyVjUgnhBWjBPBe96A3cqUGcZ5NMCelrwznhdFwwM9FPBM0eQQBsQNuT15bVWAcOW3uBibyCxoaVDBRwtA43G+abWTgx1ZLTsw+rGdjUNlU9vJ3FTbUXdkOjwYH/4drdjXK0BcZHBiAkzBsYlk5SI3DNIKN9LRERE5PoYJCQiIhPJ7vJTQa6xeVEkKKllOOoDi8YgZE/fgKONTEgZPtzc0oroiDAEB/mbMvO0jLy+QT4tc894u6dntEngMyHA2MRmOPuluc0cQJSAohZoVJmKumWmQGNL55CDvhKAbGrrVFMpWvpZq9BiToKHMmRR6pppgUN1vXdeux4VGuTx+9RbHD58GA8//DDWrFmD8vJypKam4uqrr8Z9992HwMAx+pCiYZH3aEpMKMpqLd/X8RGsR0hEROQOGCQkIiInD8mVobUj/zrq6elBZWUlEhMT4cvGG3bbLyq7L3TogRipmWnMWtQCiFpAUTc02mK5cT0Z3jwUTW1daiqubh5wPYkPSqDQGDwM7BNEVMFFFWQMVoFH1lN0Xfn5+er9/cILLyArKwu7du3CjTfeiObmZvzhD39w9uZRP3LTovsGCYdxkoKIiIich0FCIiIiGjUJthmHZ/sjMWo4WYs9uszETtQ3t6OoohrdvoGoa+lQdTZrmoxNWOS6vmNqf/Uvh9q0RYaEx1hlJkaHBVplKwarS8lCpbF19tlnq0kzceJE7N27F88999yAQcL29nY1aRoaGtSlBBxlkoC2XJJjTEmNxppdpRbLYsMD7fqacz+6P+5Dz8D96P64Dz1vP/aM8vuWQUIiIiJyGj9fX1NXbyEHNpNifPrNDJUu1BIslCCgCh5qU28Qsab3Uibp8jwQGR5dWd+qpsGEBvqbMhEtMhN7sxVVMFH9HYHs4upA9fX1iI2NHXCdxx57DA899FCf5VVVVWhpaVGPIQfSzDx2jKTQvj9O/LvbVMa3vcjnBPeje+M+9Azcj+6P+9Dz9mNz88CjbgbDICERERG5DTU8PdYfqbEDN2+Rg6TGtk5TwFCfjWgKKPZer2/uGLRRS0tHF1pqulBSM/iBV1RoYL91E011FcODVMdnaThEQ1NQUIC//OUvgw41XrFiBZYvX26RSZiRkYGEhASEh4errFe5ziChY8TE9SDQf5+qGavJykhGYny4XX8McT+6N+5Dz8D96P64Dz1vPzY1NY3qsRgkJCIiIo8jB0qqc3JIIMYnRAw65LmuuaNvhmLv9TrdsqHUUKxv6VDT4arGQWs/LpqehrsumAlvcu+99+KJJ54YcJ28vDzk5OSY5ktKStTQ48suu0zVJRxIUFCQmqxJUFAm+d/QrpP9Bfn6IjslCruLak3LEqJC7f56cz+6P+5Dz8D96P64Dz2Dj52ObxgkJCIiInj7kOe4iGA1DUaas5gyEnXDm/WBRRVcbGwftNuzdJEO8Pe+QNVdd92FZcuWDbiO1B/UlJaW4rTTTsOCBQvw4osvjsEW0mhNTY8xBQllCH5oEH9yEBERuQN+YxMRERENUVCAH5KjQ9U02HDnlvYu05Bmff1Efd3E9EGGTXsiGQoj01BIBqEECI899lisXLmS2X9u4pITJmJdfjmqG9twzamTnb05RERENEQMEhIRERE5YMhHWHCAmjLsWIvNm0iAcOHChRg/fryqQyiNRzTJyclO3TYamNTcXHnbQpV5Kx3PiYiIyD3wW5uIiIiIXM5nn32mmpXIlJ6e3idTk1w/UM4AIRERkXvxvkI4REREROTypG6hBANtTURERERkfwwSEhEREREREREReTkGCYmIiIiIiIiIiLwcg4RERERERERERERejkFCIiIiIiIiIiIiL8cgIRERERERERERkZdjkJCIiIiIiIiIiMjLMUhIRERERERERETk5fzhZQwGg7psaGhw6PP09PSgsbERwcHB8PVlLNbdcX96Fu5Pz8L96Vm4P+1LO97Rjn+8kf7Yj/9fnoH70f1xH3oG7kf3x33oefuxqalpVMd+XhcklBdOZGRkOHtTiIiIiMbs+CcqKsorX20e+xEREZG3aRzhsZ+PwctOLUuEtbS0FBEREfDx8XHY88jZaglEFhUVITIy0mHPQ2OD+9OzcH96Fu5Pz8L9aV9ymCcHiampqV47skF/7CevBY/P3B8/J9wf96Fn4H50f9yHnrcfI3qPd0Z67Od1mYTyIqWnp4/Z80mAkEFCz8H96Vm4Pz0L96dn4f60H2/NILR17KedIOb/l2fgfnR/3IeegfvR/XEfetZ+jBrFsZ93nlImIiIiIiIiIiIiEwYJiYiIiIiIiIiIvByDhA4SFBSE3/zmN+qS3B/3p2fh/vQs3J+ehfuT+P9F/JzwfPys9wzcj+6P+9AzBNkx/uR1jUuIiIiIiIiIiIjIEjMJiYiIiIiIiIiIvByDhERERERERERERF6OQUIiIiIiIiIiIiIvxyAhERERERERERGRl2OQ0EGeffZZTJgwAcHBwZg3bx42b97sqKciB3rsscdw/PHHIyIiAomJiViyZAn27t3L19xDPP744/Dx8cGdd97p7E2hESopKcHVV1+NuLg4hISEYPr06di6dStfTzfU3d2NBx54AJmZmWpfTpo0CQ8//DDYX43sicdn7u23v/2t+t7WTzk5Oc7eLBrA119/jfPPPx+pqalqf7333nsWt8tn/IMPPoiUlBT12X/GGWdg//79fE3daB8uW7asz/vy7LPPdtr20sh+07a1teG2225Tx9Th4eG45JJLUFFRwZfTzfbjwoUL+7wfb7nllmE9D4OEDvDPf/4Ty5cvVy2ov/vuO8ycORNnnXUWKisrHfF05EBfffWV+rDcuHEjPvvsM3R2duLMM89Ec3MzX3c3t2XLFrzwwguYMWOGszeFRqi2thYnnngiAgICsHr1auzZswdPPvkkYmJi+Jq6oSeeeALPPfccnnnmGeTl5an5//f//h/+8pe/OHvTyEPw+MwzHHPMMSgrKzNN69atc/Ym0QDkmFl+C0mA3hb5nP/zn/+M559/Hps2bUJYWJj63SQBC3INg+1DIUFB/fvyjTfeGNNtpNH/pv3lL3+JDz/8EG+//bZav7S0FBdffDFfWjeMTdx4440W70f5nB0OHwNP0dudZA5KhFd+6Iienh5kZGTg9ttvx7333mv/J6QxU1VVpaL28gY95ZRT+Mq7qaamJsyZMwd//etf8bvf/Q6zZs3CU0895ezNomGSz9P169fjm2++4WvnAc477zwkJSXhlVdeMS2Ts9iSWfKPf/zDqdtGnoHHZ56RSShZTDt27HD2ptAISEbLu+++q7JfhPwMley0u+66C3fffbdaVl9fr74LVq1ahSuuuIKvs4vvQy2TsK6urk+GIbnPb1p53yUkJOD111/HpZdeqtbJz89Hbm4uNmzYgBNOOMHZm0xDjE1IJuFof9syk9DOOjo6sG3bNpUqb3qRfX3VvLzByL3JB6iIjY119qbQKMgZmHPPPdfifUru54MPPsBxxx2Hyy67TH1Bzp49Gy+99JKzN4tGaMGCBfjiiy+wb98+Nf/999+rDKHFixfzNaVR4/GZ55ChqBJYmjhxIn7yk5+gsLDQ2ZtEI3To0CGUl5dbHI9FRUWpgD5/N7mXL7/8Uh2LTZkyBbfeeiuqq6udvUk0jN+0Er+QrDT9e1FKOYwbN47vRTeMTbz22muIj4/HtGnTsGLFCrS0tAzrcf3tupWEo0ePqrpKcgZMT+YlGk/uSzJCpXadDG+UNxy5pzfffFOVAZDhxuTeDh48qIanSnmHX//612qf/uIXv0BgYCCWLl3q7M2jEWSGNjQ0qINSPz8/9V36yCOPqCAA0Wjx+MwzSPBIMswkECFDqB566CGcfPLJ2LVrl6rRRO5FAoTC1u8m7TZyfTLUWIalSk3hAwcOqGMyOcEngV75PifX/00r7zc5fo6OjrZYl+9F94tNXHXVVRg/frw6mfbDDz/gf/7nf1TdwnfeeWfIj80gIdEwss/kIJS1b9xXUVER7rjjDlXDQZoKkft/OUom4aOPPqrmJZNQ3qNS14hBQvfz1ltvqTOfMtRFao7JcEI5+JGDHO5PIhL6zGKpKSxBQ/kxJJ8f119/PV8kIifQDwuXBnLy3pTmY5JduGjRIu4TF8PftJ69H2+66SaL96M0hZL3oQTw5X05FBxubGeS1ilnTKw7Acl8cnKyvZ+OxsjPf/5z/Oc//8HatWuRnp7O191NSSq9NBCSeoT+/v5qkhoOUjBbrkvmErkP+dKbOnWqxTKpncKhZ+7pnnvuUdmE8mNDDmp++tOfqiLa0smNaLR4fOaZJOtl8uTJKCgocPam0Ahov434u8mzSCkA+czl+9J9ftPKe1HKckhtST3GMNw/NiEn08Rw3o8MEtqZpOkee+yxqq6SPttF5ufPn2/vpyMHk4LK8iaUAr1r1qxRafTkvuQsys6dO1WGkjZJJpoMZ5TrHBLhXiS9XtLn9aSenWSVkPuReilSw1dP3pPyHUo0Wjw+89xGZJIdISeNyP3IcbUEJ/S/m6TshHQ55u8m91VcXKxqEvJ96T6/aSV+ERAQYPFelGNsOfHO96J7xya0Rl/DeT9yuLEDSH0sGRolwYe5c+eqzjLSlvraa691xNORg9N4Zejb+++/r2rdaPVRpKiydNwk9yL70LqeZFhYGOLi4lhn0g1Jlpk0u5Dhxj/+8Y+xefNmvPjii2oi93P++eerGoRSJFuGG2/fvh1//OMfcd111zl708hD8PjM/UkHXPmskJNBpaWl+M1vfqNOJlx55ZXO3jQaIJCrz2CRZiXyo1UK7cvnvZSV+N3vfofs7Gz1g/eBBx5QZSb03XPJdfehTFIb9JJLLlEBXwna/+pXv0JWVhbOOussp243Df03rVxKyQb5npR9GhkZidtvv10FCNnZ2H32o7z/5PZzzjlH/b6VmoTye0k6H0sZgCEzkEP85S9/MYwbN84QGBhomDt3rmHjxo18pd2QvEVsTStXrnT2ppGdnHrqqYY77riDr6eb+vDDDw3Tpk0zBAUFGXJycgwvvviiszeJRqihoUG9F+W7Mzg42DBx4kTDfffdZ2hvb+drSnbD4zP3dvnllxtSUlLU8XVaWpqaLygocPZm0QDWrl1r81h66dKl6vaenh7DAw88YEhKSlLf5YsWLTLs3buXr6mb7MOWlhbDmWeeaUhISDAEBAQYxo8fb7jxxhsN5eXlzt5sGuZv2tbWVsPPfvYzQ0xMjCE0NNRw0UUXGcrKyvg6utF+LCwsNJxyyimG2NhY9XmalZVluOeeewz19fXDeh6f3icjIiIiIiIiIiIiL8WahERERERERERERF6OQUIiIiIiIiIiIiIvxyAhERERERERERGRl2OQkIiIiIiIiIiIyMsxSEhEREREREREROTlGCQkIiIiIiIiIiLycgwSEhEREREREREReTkGCYmIiIiIiIiIiLwcg4RERC7Mx8cH7733nrM3g4iIiMgrHD58WB1/7dixw9mbQkQ05hgkJCLqx7Jly9RBovV09tln8zUjIiIictLx2ZIlS/os//LLL9VxWl1d3agePyMjA2VlZZg2bRpcEYOYRORI/g59dCIiNycBwZUrV1osCwoKctr2EBEREZFjdHR0IDAwEMnJyXyJicgrMZOQiGgAEhCUA0X9FBMTo26Ts9XPPfccFi9ejJCQEEycOBH/+te/LO6/c+dOnH766er2uLg43HTTTWhqarJY59VXX8UxxxyjnislJQU///nPLW4/evQoLrroIoSGhiI7OxsffPAB9xkRERHRIP7973+bjrEmTJiAJ5980uJ2Wfbwww/jmmuuQWRkpDpOs87U629kiWQuitraWnV/OT6UYzU5Lty/f7/pOVatWoXo6Gh8+umnyM3NRXh4uDoJLdmK/ZHH/MlPfoKEhAR1DCnHf9pJ68zMTHU5e/ZstR0LFy403e/ll19WzxEcHIycnBz89a9/Nd2m/V1vvvkmFixYoNaRbMmvvvqK/0dEZMIgIRHRKDzwwAO45JJL8P3336uDuSuuuAJ5eXnqtubmZpx11lnqoHHLli14++238fnnn1sEASXIeNttt6mDUgkoSgAwKyvL4jkeeugh/PjHP8YPP/yAc845Rz1PTU0N9xsRERFRP7Zt26aOn+TYTI6xfvvb36rjNgna6f3hD3/AzJkzsX37dnW7taeffloF9LTpjjvuQGJiogrCaUHErVu3qmO4DRs2wGAwqOO1zs5O02O0tLSo5wbwsMwAAAXfSURBVPm///s/fP311ygsLMTdd9894PHlnj17sHr1anVcKceL8fHx6rbNmzerSzmmlO1555131Pxrr72GBx98EI888oi6z6OPPqoe529/+5vFY99zzz2466671N87f/58nH/++aiurub/EREZGYiIyKalS5ca/Pz8DGFhYRbTI488om6Xj9BbbrnF4j7z5s0z3Hrrrer6iy++aIiJiTE0NTWZbv/oo48Mvr6+hvLycjWfmppquO+++/rdA/Ic999/v2leHkuWrV69mnuNiIiIvE5/x2fBwcHqGKm2tlatd9VVVxl+9KMfWdz3nnvuMUydOtU0P378eMOSJUss1jl06JB6nO3bt/d57n//+9/qedatW6fm9+3bp9Zdv369aZ2jR48aQkJCDG+99ZaaX7lypVqnoKDAtM6zzz5rSEpK6vdvPP/88w3XXnutzdv6275JkyYZXn/9dYtlDz/8sGH+/PkW93v88cdNt3d2dhrS09MNTzzxRL/bQkTehTUJiYgGcNppp6mzt3qxsbGm63IGVk/mteEpchZXzkyHhYWZbj/xxBPR09ODvXv3qiEfpaWlWLRo0YD7YMaMGabr8lgyHKayspL7jYiIiLySreOzTZs24eqrrzbNy3HYhRdeaLGOHIc99dRT6O7uhp+fn1p23HHHDek5JfPupz/9KZ555hn1ONpz+Pv7Y968eab1pLzMlClTTCNLhAxDnjRpkmleyssMdCx36623qpEq3333Hc4880zVqEWGCPdHRq8cOHAA119/PW688UbT8q6uLkRFRVmsqz92lW2Xv1+/rUTk3RgkJCIagATlrIf/2ovUmBmKgIAAi3kJLkqgkYiIiMgb2To+Ky4uHvFjDaa8vBwXXHABbrjhBhWIGy5bx3LGASO2SV3DI0eO4OOPP8Znn32mTihLeRoZsmyLVu/6pZdesghYCi0YSkQ0FKxJSEQ0Chs3buwzLwWjhVxKrUI5u6tZv349fH191RnmiIgIVTD7iy++4D4gIiIisiM5DpPjLj2Znzx58rACZ21tbSojUWoQ/vGPf+zzHJKtJ1mMGqnvJyNGpk6dOqrtl6YlS5cuxT/+8Q+V/fjiiy+q5dJ9WUg2pCYpKQmpqak4ePCgCp7qJ63Ria1jV9l2qd2oHbsSETGTkIhoAO3t7erssZ4MzdCKR0szEhmmcdJJJ6mC0VJM+pVXXlG3SYOR3/zmN+oAT4plV1VV4fbbb1dDVeRgTsjyW265RRXAlrPGjY2N6gBW1iMiIiKikZHmHMcff7zqXnz55ZerpiIyVFjf8Xcobr75ZhQVFamTunIspy8/I12HJYAoQ3xfeOEFdQL43nvvRVpaWp+hzsMhDUiOPfZY1ZlZjkX/85//mAJ5cswoo1E++eQTpKenqy7FMqRYGt394he/UNele7LcTxqqSKfk5cuXmx772WefVdstj/enP/1J3X7dddeNeFuJyLMwk5CIaAByACZ1Y/STBAQ1ckD25ptvqrqBf//73/HGG2+YzhxL/ZlPP/1UdSKWg9RLL71UDReRA1SNBBDl7LAcsMqB4HnnnYf9+/dznxARERGNwpw5c/DWW2+p47Rp06apwNv//u//qm7Ew/HVV1+pLsJyfKc/Hvz222/V7StXrlQBPTmGk3p/MoxYhglbDzEeDskWXLFihTq+POWUU1Tmo/wd2snqP//5zyooKdmDWjBShkK//PLLanumT5+OU089VXVyts4kfPzxx9UkdbPXrVunujJrJ7+JiHykewlfBiKi4ZN6Mu+++64qJk1ERERE5KoOHz6sAobSgGXWrFnO3hwiclHMJCQiIiIiIiIiIvJyDBISERERERERERF5OQ43JiIiIiIiIiIi8nLMJCQiIiIiIiIiIvJyDBISERERERERERF5OQYJiYiIiIiIiIiIvByDhERERERERERERF6OQUIiIiIiIiIiIiIvxyAhERERERERERGRl2OQkIiIiIiIiIiIyMsxSEhERERERERERATv9v8BxphZzwAZyz0AAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1300x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(13, 4))\n",
    "\n",
    "# Learning curves\n",
    "ax = axes[0]\n",
    "ax.plot(history.history['loss'],     color='steelblue',  lw=2, label='Train MSE')\n",
    "ax.plot(history.history['val_loss'], color='darkorange', lw=2, linestyle='--', label='Val MSE')\n",
    "ax.set_title('Spec-Built Model: Training Curves', fontsize=12)\n",
    "ax.set_xlabel('Epoch'); ax.set_ylabel('MSE')\n",
    "ax.legend(); ax.grid(True, alpha=0.3)\n",
    "\n",
    "# Forecast vs actual (sample 0, output dim 0)\n",
    "ax = axes[1]\n",
    "steps = np.arange(1, H + 1)\n",
    "ax.plot(steps, y_train[0, :, 0],  color='steelblue',  lw=2.5, label='Actual')\n",
    "ax.plot(steps, y_pred[0, :, 0],   color='darkorange', lw=2,   linestyle='--', label='Predicted')\n",
    "ax.set_title('Sample Forecast (Output Dim 0)', fontsize=12)\n",
    "ax.set_xlabel('Horizon step'); ax.set_ylabel('Value')\n",
    "ax.legend(); ax.grid(True, alpha=0.3)\n",
    "\n",
    "plt.suptitle('Section 9 -- Training a Spec-Built Model', fontsize=13)\n",
    "plt.tight_layout(); plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11e0855b",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 10 — Spec vs Keyword: Side-by-Side Comparison\n",
    "\n",
    "Both `BaseAttentive` (keyword) and `BaseAttentiveV2` (spec) produce\n",
    "equivalent models. The spec approach adds reproducibility and\n",
    "configurability via `dataclasses.replace`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7c25bb7a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:54.436277Z",
     "iopub.status.busy": "2026-04-21T02:46:54.436277Z",
     "iopub.status.idle": "2026-04-21T02:46:57.940703Z",
     "shell.execute_reply": "2026-04-21T02:46:57.939704Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Keyword model params : 346,323\n",
      "Spec    model params : 346,323\n",
      "Match: True\n"
     ]
    }
   ],
   "source": [
    "# Build keyword model\n",
    "kw_params = dict(\n",
    "    static_input_dim=4, dynamic_input_dim=8,\n",
    "    future_input_dim=6, output_dim=2,\n",
    "    forecast_horizon=24, embed_dim=32, num_heads=4,\n",
    "    dropout_rate=0.1,\n",
    ")\n",
    "model_kw2 = BaseAttentive(**kw_params, name='keyword_v')\n",
    "\n",
    "# Build spec model\n",
    "cmp_spec = replace(BASE_SPEC, embed_dim=32, attention_heads=4)\n",
    "model_sp2 = BaseAttentive(\n",
    "    static_input_dim=cmp_spec.static_input_dim,\n",
    "    dynamic_input_dim=cmp_spec.dynamic_input_dim,\n",
    "    future_input_dim=cmp_spec.future_input_dim,\n",
    "    output_dim=cmp_spec.output_dim,\n",
    "    forecast_horizon=cmp_spec.forecast_horizon,\n",
    "    embed_dim=cmp_spec.embed_dim,\n",
    "    num_heads=cmp_spec.attention_heads,\n",
    "    dropout_rate=cmp_spec.dropout_rate,\n",
    "    name='spec_v',\n",
    ")\n",
    "\n",
    "# Compare parameter counts\n",
    "x_s2 = rng.standard_normal((4, 4)).astype('float32')\n",
    "x_d2 = rng.standard_normal((4, 24, 8)).astype('float32')\n",
    "x_f2 = rng.standard_normal((4, 24, 6)).astype('float32')\n",
    "_ = model_kw2([x_s2, x_d2, x_f2])\n",
    "_ = model_sp2([x_s2, x_d2, x_f2])\n",
    "\n",
    "kw_params_n  = model_kw2.count_params()\n",
    "sp_params_n  = model_sp2.count_params()\n",
    "print(f'Keyword model params : {kw_params_n:,}')\n",
    "print(f'Spec    model params : {sp_params_n:,}')\n",
    "print(f'Match: {kw_params_n == sp_params_n}')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "bd0f48c1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:57.943780Z",
     "iopub.status.busy": "2026-04-21T02:46:57.943780Z",
     "iopub.status.idle": "2026-04-21T02:46:58.066095Z",
     "shell.execute_reply": "2026-04-21T02:46:58.064929Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Advantages of Spec approach:\n",
      "  * Immutable: prevents accidental mutation\n",
      "  * JSON-serialisable: saved/reloaded exactly\n",
      "  * Composable: derive variants with replace()\n",
      "  * Documented: typed fields with defaults\n"
     ]
    }
   ],
   "source": [
    "# Visual comparison: keyword dict vs spec fields\n",
    "fig, axes = plt.subplots(1, 2, figsize=(12, 5))\n",
    "\n",
    "# Left: keyword dict display\n",
    "ax = axes[0]\n",
    "ax.axis('off')\n",
    "kw_text = '\\n'.join(\n",
    "    [f'{k} = {v}' for k, v in kw_params.items()]\n",
    ")\n",
    "ax.text(0.05, 0.95, 'BaseAttentive(\\n  ' + kw_text.replace('\\n','\\n  ') + '\\n)',\n",
    "        transform=ax.transAxes, fontsize=9, verticalalignment='top',\n",
    "        fontfamily='monospace',\n",
    "        bbox=dict(boxstyle='round', facecolor='#e8f4f8', alpha=0.8))\n",
    "ax.set_title('Keyword Approach', fontsize=12, pad=10)\n",
    "\n",
    "# Right: spec fields display\n",
    "ax = axes[1]\n",
    "ax.axis('off')\n",
    "sp_text = (\n",
    "    f'embed_dim        = {cmp_spec.embed_dim}\\n'\n",
    "    f'attention_heads  = {cmp_spec.attention_heads}\\n'\n",
    "    f'dropout_rate     = {cmp_spec.dropout_rate}\\n'\n",
    "    f'forecast_horizon = {cmp_spec.forecast_horizon}\\n'\n",
    "    f'static_input_dim = {cmp_spec.static_input_dim}\\n'\n",
    "    f'dynamic_input_dim= {cmp_spec.dynamic_input_dim}\\n'\n",
    "    f'future_input_dim = {cmp_spec.future_input_dim}\\n'\n",
    "    f'output_dim       = {cmp_spec.output_dim}'\n",
    ")\n",
    "ax.text(0.05, 0.95, 'BaseAttentiveSpec(\\n  ' + sp_text.replace('\\n','\\n  ') + '\\n)',\n",
    "        transform=ax.transAxes, fontsize=9, verticalalignment='top',\n",
    "        fontfamily='monospace',\n",
    "        bbox=dict(boxstyle='round', facecolor='#f0f8e8', alpha=0.8))\n",
    "ax.set_title('Spec Approach', fontsize=12, pad=10)\n",
    "\n",
    "plt.suptitle('Section 10 -- Keyword vs Spec: Equivalent Models', fontsize=13)\n",
    "plt.tight_layout(); plt.show()\n",
    "print('\\nAdvantages of Spec approach:')\n",
    "print('  * Immutable: prevents accidental mutation')\n",
    "print('  * JSON-serialisable: saved/reloaded exactly')\n",
    "print('  * Composable: derive variants with replace()')\n",
    "print('  * Documented: typed fields with defaults')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "093cbec8",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## 11 — Registry Inspection & Visualization\n",
    "\n",
    "`ComponentRegistry` stores all registered builder functions by string key.\n",
    "Inspecting the registry helps discover available components."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4f9937d9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:58.069438Z",
     "iopub.status.busy": "2026-04-21T02:46:58.068351Z",
     "iopub.status.idle": "2026-04-21T02:46:58.081903Z",
     "shell.execute_reply": "2026-04-21T02:46:58.080691Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total registered components: 28\n",
      "\n",
      "  decoder.cross_attention\n",
      "  decoder.hierarchical_attention\n",
      "  decoder.memory_attention\n",
      "  embedding.positional\n",
      "  encoder.dynamic_window\n",
      "  encoder.hybrid_multiscale\n",
      "  encoder.residual_bilstm\n",
      "  encoder.temporal_self_attention\n",
      "  feature.dynamic_processor\n",
      "  feature.future_processor\n",
      "  feature.static_processor\n",
      "  fusion.concat\n",
      "  fusion.multi_resolution_attention\n",
      "  head.multi_horizon\n",
      "  head.point_forecast\n",
      "  head.quantile_distribution\n",
      "  head.quantile_forecast\n",
      "  pool.demo_last\n",
      "  pool.final_flatten\n",
      "  pool.final_last\n",
      "  pool.final_mean\n",
      "  pool.last\n",
      "  pool.mean\n",
      "  projection.dense\n",
      "  projection.dynamic\n",
      "  projection.future\n",
      "  projection.hidden\n",
      "  projection.static\n"
     ]
    }
   ],
   "source": [
    "# List all registered keys using list_keys()\n",
    "all_keys = sorted(component_registry.list_keys())\n",
    "print(f'Total registered components: {len(all_keys)}')\n",
    "print()\n",
    "for key in all_keys:\n",
    "    print(f'  {key}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "358af750",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-04-21T02:46:58.084675Z",
     "iopub.status.busy": "2026-04-21T02:46:58.084675Z",
     "iopub.status.idle": "2026-04-21T02:46:58.222361Z",
     "shell.execute_reply": "2026-04-21T02:46:58.220228Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABKUAAANMCAYAAACNUR6PAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAA/qpJREFUeJzs3QeUFGXa9vF7EkPOSBDJAiogAkuOEkRAicoSdV0Qc0RdxBXFrKCrguEzsUhSFBEUTEhGXEEFlJcclCgZhjSpvnM9bvX25B4GmvT/nTMv3V3pqaoez8713s9dEZ7neQYAAAAAAACEUWQ4DwYAAAAAAAAIoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAA4Y914440WERFh5yOdt84/WIUKFaxly5Z2NtF4Ne5z2WOPPebu16ZNm073UAAAOKsQSgEATpsjR47Yv/71L2vWrJkVLVrUYmJirGTJktahQwcbM2aMJSYmcndO8R/SU6dOTXeZQgT9ke3/5MqVy8qXL29///vf7bfffjvj74vCAZ3fzz//bGea//u//wtc1/nz55+y4wTfP/3ExsZalSpV7J577rE9e/bYmU73TvfwXAt6FNLlz58/zefx8fF2/fXXu3t1xx13mOd5p2V8AACEE6EUAOC0WLdunV1xxRV27733Wu7cuW3IkCH2//7f/7P77rvPEhIS7G9/+5s9/PDD3J1T6PHHH88wlJKyZcva+++/735eeeUVa9q0qb333nvWsGHDsIUab731lh09ejTb2ynI0PmdiaHUO++8YwUKFLALLrjA3n333VN6rNq1awfu4YgRI+ySSy6xl19+2Vq0aOFCkHD46quvbPXq1dneTvdO9/BcC6UyCuivvfZamzx5sj3yyCM2atSo87ZCEABwfok+3QMAAJx/FDJ06tTJNmzYYB9//LF169YtxfKHHnrIfvjhB/eD06dQoULWt2/fwPtbbrnFVbK99NJLrpLt/vvvP+VjUPWcfk41VaUcPnw43QqWk0mBqwKi6667zl1fBbEK/BRSnQoXXnhhint45513WteuXV0YOX36dOvevbudaqqyO9OE636HYv/+/daxY0f77rvv3O+WKtkAADhfUCkFAAi7t99+21VOKNRIHUj5/vKXv9htt92W4jP9Id2kSRPLly+f+2NSrz/99NM02/p9d5YtW2Zt2rRx66oqRcfTlMBjx47Z4MGD3R/sqtJq3ry5m1IVTKGLKhW++eYbN4VIU9c0/alWrVo2adKkdMec3fGtWrXK/TGqQEIBRY8ePWzHjh1p1j9w4IAL6jT1SmMoUaKE9erVy4V66Y3522+/dVUxlStXdutXrVrV/v3vfwfWU+WJX4Whz4OneGWldevW7t+1a9em+SP/9ddft7p161revHnd+bdq1cpmz56dblWIKuJKly5tefLkcZVXs2bNSrd/VHqf/f7773bTTTcF7onubePGjQPnqOugY4sq7vxz83sxzZkzx73XeqNHj7ZLL73UfQ90zTp37uzGf/DgwTTjVkiq7YYPH+7er1+/3t3D7FAQ9Mcff9gNN9zgzk3ByAcffGDhpN+J9O7h8ePH7emnn7bLLrvMXY/ChQvbNddcYz/99FOafahSTvegWLFi7l5feeWVbr30+kel99mvv/7qgjn9DuoelipVyt2zzz//3C3X75zunehz/x4G99gKdbyZ3W+f7oEqAfW7qPvfoEED++ijj9Kcd3Jysj3zzDNWsWJFt48aNWrY+PHj7UTt3LnTXZ/vv//eVc2lF0jpv0Ht2rVz56dj6r9Bb7zxRop1Lr/8citXrpwbX2qqvtL5jx07NnAOmjat/eh8CxYsaNWqVXNTcxWaAgAQTlRKAQDCzv9j7+abbw55m9dee81uv/12q169uj366KPuM/2R2aVLF3vzzTfT7GvLli3Wtm1b69mzpwt7NIXoxRdftOjoaPcHsaq1/vGPf9ju3bvdH6faj4KpyMiU//8ahUEKDvyATNPXFAgp2Ar+Azm749u6dav7Y1RVKy+88IIL0LSewhCNNTiQUuCiPk4KAfQH+Pbt293x9IfzkiVLXDgTTNMedX6DBg1yf/ArLNJYFWopKFOopWqdfv36uX5e2bkPCmJEPcCCaV8TJ05011phggID/bGuezBlyhQ3NcmnMGLGjBnu2igg2bhxo7sO+kM/KwoVtU9dP90TBW66RsuXL3f9mRT2KGTUNVBgoXPTOYqqvILpD3OFKwMHDnShyEUXXWT16tWzadOmuXPR9Us97U7fD90HP6DbvHlztnr/aB86T41JQYGmsCqMGDBggIVLevdQYUT79u1t0aJF7l6qp5Guq6ZP6jszb948d21E91b3TdPr9L2qX7++u/76LPX3Ij265gqx/Oo7fX/1e6jvssIZBbUKq/U9VyWZ7qWmHYqC1uyON7P7LZou99RTT7n9PfHEE+4ef/LJJ+57qml0+r32KUzV9Ed9xzT1WAGjlleqVCnb90HfHX2X9but4Ei/A6np/HWNFNwOHTrUBd5ff/213Xrrre4+6r8donNSFZyWXXXVVWm+cwq9dT6ic9V/oxTgad9RUVHud1Dfe93bcFQmAgAQ4AEAEGZFixb1ChYsGPL6e/fu9fLly+dVrlzZO3DgQOBzva5UqZKXP39+b9++fYHPy5cvr5TA+/DDD1Psp06dOl5ERIR37bXXesnJyYHPX375Zbf+F198Efjsvffec5+VK1fO279/f+BzvdZnRYoU8Y4cOZKj8X3wwQcpxnfbbbe5z1etWhX47K677vJy587t/fzzzynW3bRpk1egQAHvhhtuSDPm2rVre8ePHw98vmXLFi9XrlzeX//61xT70LrB2wfTGKtXr+7t2rXL/WzcuNF7//333XlHR0d7y5YtC6w7ZcoUt68333wzxT4SEhK8unXrehUqVAhc788//9ytO2DAgBTr+p+n/p8mGl/wZzqu3j/33HNeZmbPnu3W0zXJaJnOZefOnSmWJSYmehdddJH3l7/8JcXnhw8fdt/Zq6++OsU1ys7/lNq6dasXFRXlDRs2LPDZv/71L7ePlStXplk/vfujY7Zo0SKk42n7du3aBe7h2rVrvVGjRrnvgr6Twef+4osvpvkd8L/Duh7Bxxw9erRb98knn0yxrv+5xhhM2wZ/9umnn6b7/U/N/z7rfqWWnfFmdr+XLl3qlg0ZMiTNMTp37ux+xw4ePOje6/dS//248sor3fckeB/6XPvR70lWNLaYmBivbNmy7j5888036a63bds2LzY21uvVq1eaZfrvQmRkpLd+/Xr3Xv99yZMnj3fdddelWO+3335z6916662Bz6644grvkksuyXKcAACEA9P3AABhp2qg7PTQ0f/3X9VKd911l5tq4tNrfRYXF+emuATTtCC/MsCn6Tn6W10VBcFTwvxKmtTTmUQVCaoy8Om1qgv27dvnpgWd6PjKlCnjnrQVzK8e8cehsaraSFUZOh9Vk/g/qphQ9URwVZVPFUTBfXy0rSqK0ju/zGhqmqqq9KPqHlWkaLqWpiRq6o9v3Lhx7n6q8il4jOqVo2oMTRf0j63pa37FSTA9cdGvhsmMfy80LVBVKjnRv39/N/UvmKpGVAmlqXorVqxIUd2n762mOPl0XtmpklLlnKZO6bi+Pn36uMqUU9XwXN8P/x5efPHFrqJIU870fQw+d91DVflp+mXwPVQzdFXzLFiwINBwXvdQ1+nuu+9OcSxVewX/rmTEX2fmzJnpTpMMRXbGm9n91u+X/lugCrvg/ehH1X2HDh1yvZ5E33vdb313df6+OnXquGNmR1JSkvv+6vfGr9hKTd85VS7pO5d6bPq90nfJ/++KpvbpvycaY/BDCFTZqfWCv7e6/qo01DUCAOB0Y/oeACDsFNboj71QaWqJaOpaav5nqfsrpTcVrEiRIuku8z9P74ly6QUl6kkTfMwTGV96030U+ASPY9euXe61HyykJ/V0w8z2relC2aE+QJoOJep1pWmAmqalKZDBNO1R9zP19LjUvXMUjOlaacyaSpia+tqk7u2VmqZ6aRqT+vqoJ5WeLqdpdAog1YcsOzSe9OgP+CeffNJNe9KUL9FrBRrB0xCzQ2GGgieFeQoJ9PRJn6abaTqlzin1tc0pTfHUuej4miamKaya2pq6+biuu0KcjL5nojBEAYruoULV1E3CtU/9bimwzYye/KeASCGdQiHdN03901Rb/3crK9kZb2b3W/vRtVHAldl3N/h3OL11Ne70AuKMqJfahAkT3PdW03gVsur7n3pswT3AMhubaKqq+qrpu6TeVDovhVL6HVF459O0VgXICuN1H3V8TZnU1NszsSk9AODcRigFAAg7VWqo54v+yDuRXiyhCK5kCHVZdqpeTuX4/HH4/+qPUvW2yum+s3t+qsYK/oNYf7SqOkvhwcqVK10o5O9X4YD+yM7sngfLyePuFbKomklNsdVHSo3z1VvnwQcftOeeey7k/aihdXoUZKi/kKpxnn/+eRfm6fuq5vgn2m9n7ty5gV5OqlhKz2effebCgpOpePHiKe6h+hbVrFnTPXVPvdUUjvj3UJ8rtMpIZgFQdik8eeCBB1y1lO7hyJEjXa8jhYCq5srKiYw3vfut/ei7qHFk9HuTXth8MijgVDWUfq/8YCo48PJ/X9Wg3P9dSy34v5/qPaffMwWoCqX08ABV86kvVrBGjRq57+KXX37pjqkf/e7q90rVU6H0BQMA4GQhlAIAhJ3+INYf+QoT9P+1z4r/h5f+iPaf/uZTOBK8zsmmagU9kS2zY56q8emPak3L0RSnzKolwkVP/tIj6zXNcNiwYa4Jsx+yrFmzxgVWqatn0qu+UqWQpvOlrkLTExlDpeupaZj6UdN5NXdWgKQnLKqiKSehl191otBLT1T0n+YWPAUqu1QlpabzChjSq25TU3WFCSc7lEpNgYMf6uleqom4fw9Vmad7m974Ut9DTRvTtNTg+63m46qi0nc2FApQ9KNwSlM9VdWlhw+ocXhWT4PMzngzo/188cUX7sl1WU0f9X+HNa3Vb7ie+vc8uzQN7+OPP3b/TdRTBvXkTH8cfniZOljMjBqea1rlf/7zH/d90u+spoimpvumY+on+EEN2kb3AwCAcKGnFAAg7NR7RlNV9NQ79UBJz9KlS90fSqJ+LaraefXVV1NM+9NrfaY/sLLb0yVUmrKmp3r59FqPY9cf3pqGdCrHpz+29Qel/sBM7/H0kpO+ShrX3r17s7WN/nBWjytNvfKnLWoqloKmIUOGZDnFSH+EiwKRYHoaX1ZT9/zrn/qx9frD2/9D3p865ocl2T0/n6YzaWqTnoioqh5NsUs9bUvVJgooQhmz7l+7du1c3x9VxqT+UdWMqnX0xLlTTb3BFLDo98/v6aR7qCmaGVUepb6H6omkp9AF01TP4N+VjOie6PsSTL9Pmvp35MgRFzJmdQ+zM96sroUonNM5ZbYf3SMFZTpm8Lo//vhjmp5x2dGpUyf3hEp9d/X75Qdc+q4oyFQAnLo/luhaq+dU6vPR74MqB/UEQYVOqUNCTWtMTX2xcvL7AgDAiaJSCgAQdppGo6lK+sNflSH6Y12hjfoeqfpB00k0tUTTsUR/VKkKRv+ffFVT6DH0omBEvXkUHITSYPlEqEpBx/zb3/7m3qtHi3rzqMrLnw50KsenKU0LFy50f6DqR9VI6vuiKWUKctQrRsc5EdqX/pjWlDdViugP7r/+9a9ZbvfPf/7T3S+/75JCFV0fTRPSH+j6I1vXTb2L1CRa18Dvx6OG5qpqUoChP45VAaJwS1VX6reknlWZ0XdDVUz6Y1vBpoILBZi6H7r2fl8e9fhRE2kFm7pPukeqoPKbyWfFb3iuc5T0KvpUFaf7kNW0yIkTJ7pQwa9KSY+W6T4qAFO10KmkvlUKEFVVo2BJ91PVNWrYryoZVevoOqn3m77rmgamoEPX3g+V9Z1+5JFH3L2tX7++u28ffvih6xWWmJiY6fFVLaZQUlMJtb6mRGp6o37n9R33pxSq15SCWf0OKLBR8KvgSvc5O+PNjI7x2GOPuR/1XlKPJ4WRCgf1vdLvmJqni0JJ/Y7re67j6Z4pFNb7yy+/PFBRdyL030IFU926dQtUTGnaoEJxXW+Frgqc1FNN/41UE35V8SnAUuVacH88/T5q6ql/r1LTvvS7r+von6t+//TflVB+/wEAOKnC8ow/AADScfjwYfdo9yZNmniFCxf2oqOjvQsuuMDr0KGDN3bs2BSPXZcpU6Z4jRo18vLmzet+9PqTTz5Js189fj74kfC+YcOGpfvYdr3X51qe+nH0X3/9tffoo4+6x8znypXLq1Gjhjd+/Ph072dOx+c/ul7HTn2dhg8f7o6dO3du9xj56tWrewMGDPAWL16cZszaT2o6no4bbM2aNV7btm3dY++1XfD/LNC6l112mZeRhg0buvu1bt26wGe6Z02bNnX706PstY+uXbt6kyZNSrFtXFycd/fdd7t7rfOpX7++N2vWLK979+7usfbBbrjhhhTj2rBhgzdo0CB3/jqOrrNe//Of//T279+fYtvPP//cu+KKK9xYtA//mmd0nVPbtGmTFxkZ6Y6jMaem8wvlf0rVq1fPXau9e/dmuM6xY8fccapWrRr4TPvW+Yfy3UmPtu/YsWO6y+Lj471y5cq53zv/uiUkJHgvv/yyG6//Ha5SpYrXu3dv78svv0yx/R9//OHGVqRIEbdeq1atvJ9++smrW7eud8kll2T63dN6/fv39ypXruy21XnXqlXLGzFihLsOwcaMGeP2FxMTk+Z6hDreUO73Z5995rVr186dj37Py5Yt67Vv3957/fXXU6yXlJTkPfnkk+7aaT39jowbNy7D/7akR9cjX7586S6bMWOG+76WKFHCW7FihftswYIFXpcuXdxnug6lS5f2WrZs6a7X0aNH0+xj3rx5biy6FsnJyWmWP/PMM16zZs3c/vxz7dGjh7d06dIsxw4AwMkWof9zcmMuAADOfqpaUfWPqi3UhBinnhpXa2peKFPiwkEVJGp6rl5SqgxC5jSlza8sVJ8mnB6a7qt7oOq+jKbUAgBwpqCnFAAACKv0+uOoqfgvv/xyynqDnQhNnVLQoumCyPoeqteaGpafSffwfKTphJoS6U85BgDgTEZPKQAAEFbDhw93/XfUO0e9tn7++Wf3dDr1FHvooYdO+92YNGmS602kZtHqf6W+XUhJ/ajUkLxx48auGbd6h02YMMH1iCLEC7/Dhw/b9OnT3RNA1U9K96BUqVJ8bQEAZzxCKQAAEFbNmjVzzdsV+ugJYkWLFnVNo5944gkrW7bsab8bvXr1co2yNU41ckdaejjB6NGj3T2Li4uzkiVLuqbaeq8G8wgvNT/X91aN/9XoXA9eAADgbEBPKQAAAAAAAIQdPaUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2EWfyp0PmrbFftx+zL2OjDDLHR1hxfNGW62Sua1njUJWvURut2zptiN2y/Rt6e7j4mK5bEKPcoH3xxOT7aOVB+zr9XG2aV+8JXpmJfNFW90yeax/7SJWtmCMWy8p2bOPVx6wmWsP2cZ98Raf5NkF+aOtUdl8dsMVha1U/j/Xk8dm77TP1xwKjDM2KsKK5o2yS4rntu6XFrR6F+YNrLvtUIJ1nrA53bHmzxVps/9WKc0+/f0WiI20y0rkttvqF7NqxWNzdG0BAAAAAADOZqc0lPLFRJoLYf44nGi/H0iw3w4k2Mx1h+wfTUtYl0sKpVj3woLRViR3VOB9uUK5Aq8PHk+y2z7bZqt3H3fv88VEuBBqR1yiffJ/B61mydzufWKyZ/fO3G6Ltxxx6xXOHWnF8ka7Y/8ZaB2y16650KoWSxsMXVoi1vYeTbLthxJt68E4+2ZDnA2oU8QG/aVYmnUvyBdlF+T73yXMqxNNR40LYu1Igmcb9sXbot+P2P/tOm7T+pS33NEUqgEAAAAAgPNTWEIpVUe91/Ui93rlrmP2j6922Pa4RHt2wS6rXSpPinX/XqeoXVOtYLr7eX7BrkAg1e/ywq7iKFolSGb247aj5mc8H/yyPxBI9bi0kD3QtLhFRkTY0m1H7c7Pt9qB48n22Lc7bXyPiywi4s/tff44N++Pt4e/2WFr9sTb2z/us9ql81iDsv+rmJLO1QvazfXShlWp+ft844c99s6P+2zfsSRXvXXJfyvFAAAAAAAAzjdhL9W5tERuu79Jcfc6Kdns01UHQ9ou7niSq1ryp/Td2eB/gZTUKZPHav034Jrx32lzeWMi7M6GxVwgJZri165KAfd67d5495OR8oVz2bBWJQPvP/m/A3ayREWalcgbljwQAAAAAADgjHRakpHg6qiN++Otafn/VSANn/OH+/ENrFvEVSNtPpDgQiy5olSeNBVOwbSuaCpf6il1mrL3uR0KVEOlN4UveF1NETzspt79uc9gby3d5358HasWsMeCgizf3z75PTB9L0+0grLiVjxo2h8AAAAAAMD55rQkI56X8bLUPaWCezb5MsmjUq2XdsWg4qqQZDLUND2l/Cbrqf3yx59TDqV0gRirf2HKKYsAAAAAAADnm9MSSv2042jgdcXC/2tknllPqfKFYty0N1VL/bzjmHmel2G1lNZVL6jfD8TbkYTkFNVSfk8qt16qY6emdVXhJJWKpA2cQu0p9Z+bK9v/7T5ud83Y5qql1KtqXPe0/awAAAAAAADOF2HvKaVG5y8t2u1eR0WYXVM9/abmqeWPjbI2lfIHwqLR/9njnrLn+37LEVv237CrQ9U/+0YpUHp18R5L/m9plhqdf7X+v32piuZyPxn57UC8DZ+zM/C+S/WUTwnMDoVP6qU1sG5R916B2awNh094fwAAAAAAAGe7sFRK7T6S6Poq/XE4yXYdTnRT4lT19I+mJaxSkVy29GhiYN13ftxrU1b+r6m4qpxGd7rQvX6gSQn31DqFOv/+eb99tPKglckfbTsPJ9rB48n2aMsL7PJSeaznZYVt8e9HbPGWo/bRygP2zYZDVjA2yn4/kOCOXTA20h67smS6lUoa596jSbYjLtH8zGtAnSLW8KKUT94TNWn/7vc/n/Ln01hT97EKrqx6Z+mfT98b8/M+a1P5z5ANAAAAAADgfBOWUCoh2ezXP45bnpgIu6hQjNUsmdv+WqOQVS+RO826Ww8muh9f/lz/C3gK5Y6yd7uUtcm/HrCv18fZpv3xrql5iXxR1qpifrui9J+9mqKjIuxfV5exj1cesBlrD7kga9uhBCtdINoaXZTXbqxdxEoVSL//k8YZGx1hpfJH2yUlYu26ywpZ3TJpAylRyKafYElB1Vup5Y6OtJ41C9kbP+x11V4LfztsTcrlC+EKAgAAAAAAnFsiPDVnAgAAAAAAAM7lnlIAAAAAAAAAoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpU5A7969bfr06efNcXFqPf300/bKK69wmQEAAAAA5xVCKZw3rrnmGpszZ066yxT2KfQL93Hl4YcftrvuuuuUHBsAAAAAgDMVoRQAAAAAAADCLjr8hzz7bNiwwYYPH+7+bd26tSUlJaVYPnfuXHvzzTdt27ZtVqVKFXvkkUesQoUKgeULFixwy3/77TcrXLiw3XnnndamTRu3bMaMGfb222/b3r177dJLL3VVM2XLlj0px7355putbt26tnr1aluyZIkVKFDA3nnnHStVqlSW55zZmLWvwYMH2z/+8Q8bNWqUHThwwK6//nq3js7jmWeesR9//NHy5cvnPu/bt29gvzpf7Xffvn1WpEgRu+222+yqq65yy7Sfxx57zH7++WeLiIiwqlWr2muvvWaRkVlnp+PGjbOPP/7Ydu3aZYUKFXJVT3369HHLHnroIVu0aJEdO3bMhg4d6vanazxx4kRbtWqVDRw40F3bxMREa9asmdvmn//8p7Vr1y7L66xrrHGuWLHCNm7caI0aNbKnnnrKoqOjMz2uzJw5003di4+Pd9fp/vvvT3FOmX03MjsuAAAAAABnBQ+ZSk5O9nr27Om98cYbXkJCgvfRRx95devW9aZNm+aW//rrr16TJk28H374wUtMTPQmT57sde/e3b2WX375xWvWrJm3cOFC99nWrVvda1m9erXbdtmyZV58fLw3cuRIr1+/fifluDJw4ECvXbt23qJFi9zn69at8/bv35/lHc9szKJjNmjQwBs6dKjb37Fjx9w2MnjwYG/IkCHe0aNHvQ0bNnht2rTx5s+f75bps/r167vxyPbt272VK1cG9jtq1CjvlltucesdP37cW7BggbsOoZgyZYo7v6SkJG/FihVew4YNveXLl6dYp1OnTt7s2bPT3V7XtVevXmk+z+o66xp37drV++OPP9xP69at0xwjs+PKsGHDvBEjRqT4LLPvRqjHBQAAAADgTMb0vSxs3brV1q1b56p9VIXSrVs3V4njmzp1qqv0qVevnkVFRVmPHj1s9+7dbpvg5Y0bN3bLy5Qp417L7NmzrWHDhlarVi2LiYlxFTsrV6607du35/i4vubNm7sqGq1TuXLlFPvISGZj9qmq6O6773b7i42Ntcsuu8x9Nm/ePLvxxhstd+7cVrFiRWvfvr3NmjXLbZOcnOz2t2XLFjt8+LCr2LrkkksC+1R1lD7XuefKlcuaNGniPgtF165d3fmpGqlGjRp28cUXuwqxnArlOquiqkSJEu5HFU2bNm3K8XEz+26cyuMCAAAAABAuzPXJgqaZKWDRVDRRSFK0aNHA8h07dtjSpUsDwYskJCS4aWTVqlWznTt3Wp06ddLd9549e6x48eKB95pep4BHn3uel6Pj+sqVK2fZldmYfXny5HFhSLD9+/e7aXDB51SsWDHbvHmze503b14bMWKETZo0yUaPHu3G9uCDD7oQSfr372/Hjx930960r6uvvtruvffekKbvffnllzZ27Fh3XTSGo0ePuuuRU6Fc54IFCwaWKUDSOeRUZt+N0qVLn7LjAgAAAAAQLoRSWVAQpJ5AquBRQKSwSD1+fCVLlnT9i26//fZ0t9dyVf6kR4HN2rVrA+8PHTrkggUdU8fJyXF9qu7JrszGnNl+1XtKn6uSyA/QFKIEh2mquNKPgp0XX3zRnn/+eRcmSf78+V0IpZ/169fbgAEDrEGDBta0adMsg6NHH33UXn31VVfRpBBL/aR0zYJlFm5lVJEV6nXOTCihWna+GwAAAAAAnAuYvpeFCy+80DWUViNtTU+bMmWKa8jt69y5s02bNs2WLVvmpqcpPFADa79KR8u/+OILW7hwoavgURXS4sWL3bIWLVrYd999Z8uXL3frq6m1qm9UCZPT44ZCFUtdunRJ83lmY86MphkqQBozZowL1NSAW/tp2bKlW67qp2+//dZVMSkEUlijIMqn42kKmsIkVf7ovIKXZ+TIkSNuGz+wUdVU6mmMosqjNWvWpLsPVX1papyuY+prkdPrnNlxM5LZdwMAAAAAgHMBoVQI9FQzBQRXXnml6+uj3kU+TT3TU9FeeOEFa9Wqles5pCfX+ZU3Wv7kk0+6p8hpez01LS4uzi2rXr26DRkyxIYNG2Zt27Z1PZD0NDZ/25wcNxQKudTfKbXMxpwVnY+eJtexY0f3ND5VLPmhlIKjDz74wE3L05P8FBxp+p5P1VnqU6U+WLfeeqvrTVW7du0sj1mpUiW76aabbNCgQW6/P/30k9WsWTPNejoPBUrqc6UqrGCqsFIFV/fu3a1Dhw6up9PJus4ZHbdXr17uaX9aNnnyZPdavaNC+W4AAAAAAHC2i1C389M9CAAAAAAAAJxfqJQCAAAAAABA2NHoHGe8d9991957770Ml+vJeLly5QrrmAAAAAAAQM4wfQ8AAAAAAABhx/Q9AAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKVOQO/evW369Onn7HFvvvlmmzBhQqbr7Nixw5o1a2ZxcXF2pluyZIm1bNkyy/Wefvppe+WVV07acUePHu2uUf369bO8nueSs+m7AQAAAAA4fQilcEJKlSpl8+fPt/z5858zV/Dhhx+2u+6666Tt7/bbb3fXqHbt2nYqKKBUUJmRa665xubMmRP2456L3w0AAAAAwMlHKAUAAAAAAICwI5QKwYYNG+zGG2+05s2b2+OPP25JSUkpls+dO9dVjmiK2IABA2zTpk0pli9YsMD69etnLVq0sM6dO9s333wTWDZjxgzr1q2b2/a2226zLVu2nLTjahrem2++affdd5/bR8eOHd3UqlBs27bNbrjhBrfdQw89ZImJiYFl119/vTVt2tTq1atnhw4dSrHdwYMH3Vivuuoq69Chg7311lvmeV6aqXRffPGFderUyU3zevXVV92ycePGWdeuXd2+Ndbx48en2Hdm22Z1nWXMmDHWunVrt+/vv/8+8PnMmTPdvho0aGAjR45Mcy10jk899ZQ7Hx3/nnvusfj4+JDGfKJ0n/R90Hg1tjvuuMO2bt3qlq1atcp99swzz9i6devca/189dVXbrnul95rH0OHDnWve/XqFdI98q9xetcqq+Nm9d3Yu3evPfDAA26/1157rbt2vsyOCwAAAAA4R3nIVHJystezZ0/vjTfe8BISEryPPvrIq1u3rjdt2jS3/Ndff/WaNGni/fDDD15iYqI3efJkr3v37u61/PLLL16zZs28hQsXus+2bt3qXsvq1avdtsuWLfPi4+O9kSNHev369Tspx5WBAwd67dq18xYtWuQ+X7dunbd///4s77i269q1q/fHH3+4n9atW3uzZ89OsY7OQ+M5ePBgis/vuece78EHH/Ti4uK8Xbt2uXOYOXNmYLnG26BBA2/o0KFuLMeOHXPXSKZMmeLGmJSU5K1YscJr2LCht3z58pC2zew6a7u//OUv3tixY92y0aNHe7169Upz3sOGDfNGjBiR5nOd0+DBg719+/Z5x48f97766ivv6NGjIY3Zv57jx4/3smPTpk3e9OnTvUOHDrnzHDJkiHfHHXekWEffhfTOw9epU6c09y2rexTKtcrquBl9N3QNdR66dhs2bPDatGnjzZ8/P+TjAgAAAADOLVRKZUHVKaoK6du3r0VHR7uqpkKFCgWWT5061VWcqDIkKirKevToYbt373bbBC9v3LixW16mTBn3WmbPnm0NGza0WrVqWUxMjA0cONBWrlxp27dvz/Fxfap0atSokVuncuXKKfaRmXbt2lmJEiXcz6WXXpqmCis9Or56CakyK1++fFa8eHHX1+jrr79OsZ6qru6++243ltjYWLvsssvc56o40hgjIyOtRo0advHFF9vq1atD2jaz6+xTtZCW6Zps3rw5pOvgn5OqjwoXLmy5cuWytm3bWu7cuUMe84koX768qwZTXyadpyqaTsZ+Q71HJ3KtMqP7Nm/ePFf5p2tXsWJFa9++vc2aNeuUHhcAAAAAcOaKPt0DONPt27fP/RGtP+AlIiLCihYtGliuKVJLly5N8cd1QkKC7dq1y6pVq2Y7d+60OnXqpLvvPXv2uFDAV6BAARdA6HNNp8rJcX3lypU7ofMuWLBg4LUCs+PHj2e5jT81sGfPnoHPkpOTrUqVKinWy5Mnjwu7Uvvyyy9t7Nixbj+aqnj06FF3TqFsm9l1lrx587pwLzvn45+Tgqjg+5TdMZ+I/fv3u6mEusdHjhxx+/bHnxOh3KMTvVZZnY/OIfg6FitWLEXwdCqOCwAAAAA4cxFKZUFB0LFjx+zw4cMuIFJYpN44vpIlS7q+TnrSWnq03O8FlJr+KF+7dm3gvXrw6A9xHVPHyclxfao6CRc9dU3HU88nv5Io1DEpLHn00UddjyhVf6nyqE+fPin6UWW0bVbXOafnpP5RCvtSh2GhjlkBi4Kf7Bg1apTFxcXZxIkTXVWY+ocNGzYsxToKKjOj8ZzoPcpMVsdNj6rMdFxVavnhqsLX4KAVAAAAAHB+YfpeFi688EKrWrWqa8qsKUhTpkyxAwcOBJarofa0adNs2bJlLnhQsKTG2X61jJYrAFi4cKGrFFFFz+LFi90yNeT+7rvvbPny5W79t99+21U5lS5dOsfHDcWkSZOsS5cudrKoCkbTEV966SUXqOh816xZ45pYZ0XVQApz/JBCFUippyJmJrPrnNNzatKkiT333HOu2kfX99tvv3WBYahjrlSpkqt4yk4wpeunqXsKJBXepNdAXSGZpnqmbigePHZd/5N1j0I9bnpUAaUG6Gpkrmu3ceNGd7/U3BwAAAAAcH4ilAqBnrym8OjKK690PZ/UQ8inPkIPP/ywvfDCC9aqVSvX20lPgfOrSbT8ySeftNdee81tryfiKQyQ6tWr25AhQ1wFjPoUqWfQ008/Hdg2J8cNhUKu4Kf9hUJPWtMT1/SUNVGvI71ftGiRez98+HAXol133XVu3E888YSr9sqKgpubbrrJBg0aZG3atLGffvrJatasGfK4MrvOWVEfI52DQr3Jkye71+rv5dM5qVpJ62lsn376qatCCnXM/fv3d0+8U58k9QgLhfapqW26t7feeqvrC5aaqrPUN6t79+7uPqhHWTBdA52Tejfp6YzB53Mi9yir42b13dB3XVVnerLenXfe6arKCKUAAAAA4PwVoW7np3sQAAAAAAAAOL9QKQUAAAAAAICwo9E5cBpoWltGNMVP0/cAAAAAADiXMX0PAAAAAAAAYcf0PQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlDoBvXv3tunTp583x8Xps2PHDvekvri4OG4DAAAAAOCcQigFmLmwT6FfRq655hqbM2dO2I9bqlQpmz9/vuXPn5/7BAAAAAA4pxBKAQAAAAAAIOyiw3/Is8+GDRts+PDh7t/WrVtbUlJSiuVz5861N99807Zt22ZVqlSxRx55xCpUqBBYvmDBArf8t99+s8KFC9udd95pbdq0cctmzJhhb7/9tu3du9cuvfRSe/jhh61s2bIn5bg333yz1a1b11avXm1LliyxAgUK2DvvvOOqbzKr3Jk8ebIbT8WKFa1MmTL2zTff2P33328dOnTI9LihbJvZ+WY25pEjR7pjDRo0KLDu9ddfbzfddJO1b98+yylwuo7aZ3x8vF1++eU2ZMgQu/DCC23VqlU2cOBAd20TExPdVDn55z//ae3atbOHHnrIFi1aZMeOHbOhQ4daZGSkG+/EiRPdegcPHrSXXnrJrRMVFWVdu3a1AQMGWEREhBv/4MGD7cYbb7T333/fcufObY8++qg1aNAgy+P656drrGPPnj3bXQufrt8zzzxjP/74o+XLl8+t27dvX7css+MCAAAAAHDG8JCp5ORkr2fPnt4bb7zhJSQkeB999JFXt25db9q0aW75r7/+6jVp0sT74YcfvMTERG/y5Mle9+7d3Wv55ZdfvGbNmnkLFy50n23dutW9ltWrV7ttly1b5sXHx3sjR470+vXrd1KOKwMHDvTatWvnLVq0yH2+bt06b//+/Zmer/bfv39/7+jRo16HDh28iRMnuvH+/e9/z/K4WW2b2flmNebZs2d7Xbp0Cay3Zs0ar3nz5u5YWdm0aZM3ffp079ChQ96xY8e8IUOGeHfccUea8+7Vq1eG++jUqZMbQ2r33HOP9+CDD3pxcXHerl273D2bOXOmW6Zr9Je//MUbO3asO5fRo0enOUZWx9X3Rff94MGDKT4fPHiwOw+d/4YNG7w2bdp48+fPD/m4AAAAAACcbkzfy8LWrVtt3bp1rgolOjraunXrZoUKFQosnzp1ql111VVWr149VynTo0cP2717t9smeHnjxo3dclUP6bWo+qVhw4ZWq1Yti4mJcZUzK1eutO3bt+f4uL7mzZtbo0aN3DqVK1dOsY+MlCtXzlXXqKJKFU8XXXSR7dmzJ6TjZrZtZueb1ZibNm1qhw4dsl9++cWt8+WXX7rqMR0rK+XLl7dOnTq5vkyxsbGuaktVUzml81a/p/vuu89VKxUvXtz1nvr6669TrNerVy93LjqvzZs35/i4qqyaN2+eq4TS+es6q1ps1qxZp/S4AAAAAACcTEzfy8K+ffvcH/4KHUTTsooWLZpiatjSpUtTBAIJCQm2a9cuq1atmu3cudPq1KmT7r4V1ijI8Gl6lkITfe55Xo6O61NIlF0KMvx//R9/6mBmx81q28zOt3Tp0pmOWcGcwrCZM2dajRo17KuvvrJhw4aFdD779+930/807iNHjrjxaH85pWshPXv2DHyWnJzsphn68ubNGziWgrjjx4/n+Lg6H51D8LUsVqxYiuDpVBwXAAAAAICTiVAqCwqC1NPn8OHDLiBSWKR+Pr6SJUu6p6fdfvvt6W6v5ap6So+ChLVr1wbeqxJI4YGOqePk5Lg+PyTKKR0/q+Oqp1Rm22Z2vqGMuWPHjnbPPfe4flzaZ0ZhX2qjRo2yuLg41wdKVVfqiZU60FLolxn1kkpN1WAa6xdffBFSxVZ6sjpuetSXTMdVpZZ/7RTspb6OAAAAAACcyZi+lwU1w65ataqNGzfOTZuaMmWKHThwILC8c+fONm3aNFu2bJmrklHQomoeVQ/5yxVaLFy40FW3qHJq8eLFblmLFi3su+++s+XLl7v11QBcVU6qGsrpcUMxadIk69KlS7a+MDk5bmbnGwo1Rlcg8+yzz7rpaqEGOgqkNHVP4Z7Cm/Hjx6dZp0SJEm4aoc4nPapKWrNmTZrPNB1Rjc51DN1fraNG46HK6rjpUQWUpjOOGTPGBZcbN25037GWLVuGvA8AAAAAAE43QqkQPPXUUy5MufLKK10PJPU58mkqmZ4g98ILL1irVq1cjyU9bc8PTLT8ySeftNdee81tr6fLKcCQ6tWru6fAqWqnbdu2rs/R008/Hdg2J8cNhUKuLVu2hP5tyeFxszrfUKgf1Pr1691UvlDpiX2a2qbx3nrrra5fVWrqkaVeX927d3fHUP+rYLpvCt8Uhunpej491U+h4XXXXefu0xNPPOGq20KV0XE1PVFP5NNT9fzz1ns95U90HfUkQVWP6WmOffr0IZQCAAAAAJxVItTt/HQPAgiVgqGxY8e6qXgAAAAAAODsRaUUzhqqDNKUQ00hBAAAAAAAZzcaneOsoJ5Jmgap6W7dunVLsUzT2jLSt29fN30PAAAAAACcWZi+BwAAAAAAgLBj+h4AAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIu+hTufM97/W2+E3fp7usyF9ft4jcBW3vmD6Bzwr3fM3yXHpV4P3Br561wwvfCrwv9chKi4iJtUOzX7a4Oa/8b2cRkRaRu4DFlKxu+VvcYbGVGgcW/fFSc0vavzXdMRS/ZbrFlL7Ujvz0kR2Y+tB/9xVhFpXLIvMUtugSVSxPrWstz+XdLCIyMv19RkZZRK68FlWwtOWq0MDyNbzBootVDKwbvO+iN4632IoN7fjGxYHzjshbxC64e7ZF5i7g3u+bfJcd++Vziyp8oV1w77zAfrzkZDv688d2dNknlrBztXnxhy0yb1GLKnKRxVZpbvkb3WQRufJkfVMAAAAAAADO9VAqICqXC3+CReQprKQlxWdH/jM2EEp5CcfsyI+Ts9x1dKlLLSIyyhL+WO0CsL1bl9kFd35tUYXKpDxebH4XMqX4LJ0QJ7p4ZS2xxH2bLX7DQvdzdMVnVrT3/7OI6Fxp91m8kguoEv9Y436O/vSRFb7+VctdtVUIF8bMO7LPDi962wpceW/G6yQct70TBrqxOFG5LLpoBfOSjlvC1uWW8NsSy1Ors0XnKhvSMQEAAAAAAM6LUCqqQAkrPvDjNJ+rYuh/K8VY/MbFrgoopmQ1O7r8U/OO7nefW1JChvtWxVV0kbJ2ZNlUOzDlfrOEYxa/dbnlSRVKxZS+zIr9bUKWYy3Y8XFXzZR87JAd+OxRO7ZimsWvn29xc1+1Aq3vz3Cfx9fNd1VO3rGDtv+je+yCe+ZYZN4ioVweO/zde5a3QX+Lylcs3eWH5rwSCKRy1+hoha55KlBZpfDu2JpvLTJ3wZCOBQAAAAAAcCY4Y3pK5b7kzwqpI/953/17+Puxbiqd/3lmvOQkS9r7W+B9dJFyOR6PQp/CnZ+xyAIX/DmupR+Y53kZrh9bpZkVaHnXn+M5Hueqq0IRXfoyNxUvbu7odJfrmEd/+vDPMeUvYYU7PxcIpCQiJrfluayDReYhlAIAAAAAAGePsFRKaXrb9mGaFvc/pR9fn+J9npqdLH7jIlchFXtxS0vcucpyVW7mpsdlZte/WqR4n7/lXWmmCoqm9gWPIXXPpvQo8IkpU9OOr55lyYf3WPKRvRlWM0mu8n8JvE7ctS7TfQfG22SAHfhsmB1ZMtHyNf57muXuuIf3utcaiz/l8MDMJ+zI4jH/20/Lu6xAq7tDOiYAAAAAAMD5USmlnlJla6f4SW+dPHV6mhd/xPZrGp6Z5WvQP6SeUjEXXu4ahkvc/Nft+Lq0YZP6PwUfX9uFJFXfq0xXzca6gXHlLmT5m9xslhRvcbNfzmLl/92u6MJl3XkDAAAAAACcjU5rT6nU8v2lj3vanqa/RRUp5yqmEratyHQbv6eUeivterWtJR3YZofmjnJPpDuRnlLBtM+Ebb+415H5irqn3WUmfvOSwOvUTdUzk6/hjXb4+3/b0eVTLfqCqimWReYr5o6taqmE7b+Yl3jcIqJjLV+jv1neer1tx5MhhmsAAAAAAABnkDOmp5REFSptuau3da/z1u9rEZHZGV7E/14mHs/xWNTofP+nQyw5bpd7ryquiIigY6RyfP0C1wzdjSQ2v+Wp0THkY2lKXv4Wd5glJ1nijv9LuSwiwh3bjengDjvw+WMuLAMAAAAAADibhaen1KFdtvut7ik+U6VPZL7iadYt3G2keZ2GW0SIT5PbN+lWi4iKsaR9v7ueTxL732ArWML2X9OMoUDbhyy2Qv0Unx38fJgLuBL3bTZLjHefqbdVgZZ3pr/P/9fVVWclx+12n0XE5LHCPf7lqpuyI2/dnnZ40TuWtO9/DdsD42xxpyVsXWbxGxbZ0R8/tGMrpltU0fKWfOxAto4BAAAAAABwXoVS6peUsOXnFB8lH9qVbigVERPrfkKVuGPlny+ic1l0iYstd81Olr/ZbWnW05TA1GPwju5Puz81KI/KZZF5C1t08cqW5/IulufybulWbbl9bvvFInLldcfOVaGBC9uii1Ww7FKwlr/V3Xbgv/20UiyLibWi/ca4JwBqil/iztWWuHuDReYv5o4ZW62N5al5TbaPCQAAAAAAcLpEeJ7nnbajAwAAAAAA4Lx0RvWUAgAAAAAAwPmBUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilTrFrrrnG5syZc1L21bt3b5s+fXq6y3QMHSu1p59+2l555RU7V+zYscOaNWtmcXFxYd32RIwePdodr379+jZhwoSwHBMAAAAAgLMFodQ57uGHH7a77rrLzhWlSpWy+fPnW/78+dMsU2Cn4O5Etj0Vbr/9dne82rVrh+V4AAAAAACcTQilAAAAAAAAEHaEUiE4ePCgPf7443bVVVdZhw4d7K233jLP89wyTZm77777rG3btjZmzBj3/sYbb7Rjx44Ftl++fLn7XOu89tprgW1l7ty5rrqnZcuWNmDAANu0aVNg2YYNG9y+mjdv7o6flJQUWKZ9vP7669auXTvr1KmTO0awmTNnuqljDRo0sJEjR6ZYtmTJEnc8jbd169bWsWNH+/777wPLNQYdV9s/8sgj1rNnzwynDaam89R0Qf98Ncbg850xY4Z169bNHf+2226zLVu2BJYdOHDA7r33XmvVqpVdeeWVdsstt1hycnJg+fXXX29Nmza1evXq2aFDhwKfr1q1yo31mWeesXXr1rnX+vnqq6+y3Fb27t1rDzzwgLsW1157rY0bNy7ka6V1u3bt6vatZePHjw/pOgEAAAAAcL4jlArBsGHD7MiRIzZlyhQbO3aszZo1y7788svAcoUsN998s33wwQc2efJki4iISBESfffdd267f//73/bZZ5/Z7Nmz3ecrV660oUOHulBL+2zfvr0NHjzYhU8KcjT1rlGjRvbtt99ajRo1bP369YF96jMFRdrn+++/bwsXLkwx5quvvtpNHdO/6dH5REVFueBGYcrLL78cWKYgqm7duu4YCnGCjxuK4PPVGBW8yZo1a1yPq8cee8y+/vprq1KlijvH4IBHYZ4CtS+++ML69evnrqXvww8/dD+pVa9e3Z3rkCFD3D71Wj8K7LLaVhRmxcTE2Oeff+6ug8a9YMGCkK5Vvnz5bMSIETZv3jx77rnnbNSoUbZixYpsXS8AAAAAAM5HhFJZ2L17tws4FBwpgChevLirAlKo4qtQoYKVK1fOypQpY7lz57ayZcvanj17Asu7dOliRYoUcctVPeSHNFOnTnXVVwp+FHr06NHDHU/VPlu3bnX/9u3b16Kjo13wVahQocA+tQ+FLqVLl3b7VrVOdvXq1csdV5VYmzdvdp9t27bNVR797W9/c0FN586dUxw3FKnP12/0rjCuYcOGVqtWLbfvgQMHumBu+/btbrkCqMOHD7tzz5UrlzVp0iRFKHUqJCYmukBJlWG6dxUrVnThoELCrK6V6LpXrlzZIiMjXXB48cUX2+rVq0/pmAEAAAAAOBdEn+4BnOn0xDbRFDafppSpIsenQEKBhX5E/wZPtStWrFjgddGiRd20PH/fS5cuTRGAJCQk2K5duyw+Pt6FJArCROGMtvXt27cvxRiCjxGKvHnzurBLFBAdP37cvVaYFhsbG2gGnvq4ocjofLVvhXq+AgUKuGPpc4Vr/fv3d+N46KGHbP/+/a7KS9P5dH1PFR1H9yp4XBp/cPCU0bUSVcypKkz3Uvs5evSou4cAAAAAACBzhFJZ0BPbFDJpOplColAF91FS9VNw/yI/tClZsqTrJ6WntKWmaiFNZVPlkIIp7U/bBoc9we+DK7NyQmNT6BIXF+eCqdTHDUXq8/VDLe177dq1gWXq7aRj+ct1PIVQ+tGUQfXYUk8s9WsKxYlUVRUuXNjdX43ZH4euZShBnIKoRx991F599VVX7abwrE+fPinuvR9kBffGAgAAAAAATN/LkipoNOXspZdeckGNqmHUG0kNsEP16aefuoocTVPTtD81zhZNjZs2bZotW7bMhRYKadRPSZU2F154oVWtWtX1WdIUM/WzUiNwn5qBq8eR9qmqqU8++eSkfJ815a5atWqur5KOq7Gr0XtqkyZNctP0QjnfFi1auM/1r/pNqd+WzvHtt992x1KVlKgvlpqsK9Txgxy/YisUJUqUcMdM3cg8M6qAUuilRuYKATdu3OgCSP8eZUa9pjRWP8BS1ZSmXKZWqVIlVxFHMAUAAAAAwP9QKRWC4cOHuyfKXXfddS6IUP8oVfGESqGWmnYr9OjevXsgpFEPIjX6fuGFF+z33393lViquFEfJnnqqafcU/f0RDd9pt5FPu1DvZ9uuOEGNwVO/aWCnzanHkh6sp2mAaqCSP2r1BBcTw7Mio6r5u5q3K6n4Om4qauQFJAFPzkvo/NVzyUFaKLjqxm59q0gTe/V+Nzft6rDnn/+eVddVbBgQdfnqXbt2m6Zzu2JJ54IVCHpKYii5uKNGzd2r3Xt9FrXWGGTnqjnh3eZbasxPfvss66JeZ48eVy1UyihlMKmm266yQYNGuT2rXtQs2bNNOtpWuI//vEP149K/ceCn+4HAAAAAMD5KsJLPdcISEUhjqapKWzKiprA33///SGFOgAAAAAA4PzF0/eQxs8//+x6Omm6mabfqeJJVV0AAAAAAAAnC9P3kMYff/xhjzzyiJuiV7ZsWTfNLTu9nQAAAAAAALLC9D0AAAAAAACEHdP3AAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0IpAAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRygFAAAAAACAsIs+lTv/++rRtiRufbrLXqr8NysQlccGrHnNvS8clc8+rznU8kfldu8f3DDWvtz3s5XJVcRm1vxnYLtkL9mm7vmPfbZnia05ut2OJydYiZiCdnn+Cta/ZEu7JG/ZwLoz9i61j3ctttVHt9nR5HgrHl3A/lKgiv2t1JVWOU+pwHqvb/vC3tj+lXsdYREWGxFtRWLy28V5SluXYvWtdZFaKcZ++dL7Mjzn+Zc/ZQWj86TYp79fnVu1PGXs5tJtrUHBqidwRQEAAAAAAM4NpzSU8sVERFn1vBem+KxQVF5LNi/wfn/SYfv3ztl2e5mrM9xPfHKi3bX+Hfvu4Gr3PndEjJWLLW67Ew7ZjL0/utd+KPXYpg/skz3fu9f5I3PbRbHF7Ldju2363iX29b5l9lKVm6xxwWppjlE1TxkXYG07vse2x++zeQdW2rXF/mLDy//VIiIiUqxbJDqflY0tluKz6Ii0xWfV8ujcPVt3dIcL6X5Z95t9WuMfVipXkRCvIAAAAAAAwLklLKGUKpnGVb8nzec/HFqX4v24nfOsV4lmVjQmf7r7eXP7V4FAqmPRuvZIuR6WNyrWvV91ZKvtSTjoXn+zb3kgkGpVqIY9X6m/5YqMtrVHt9nANa/bvsTD9sjGCa4yK09krjQVXBfGFrVdCQfdOosPrbFpe36wK/JXtG7FG6ZYt1mhS+2JCr2yPH9/n5/s/t4e2/yBHfMS7JfDvxFKAQAAAACA89YZ01Oqep4L7UjycXtrx9fpLvc8z6bsXuxeaxreo+WvCwRSbvu8F1qTQpe415/tXRL4/MGLurhASi7OU8Z6lmjiXu9JPGSLD67JNEh7umIfyxXx57Yf7frOTiaqpAAAAAAAwPksLJVS2+L3penDtKzuiyne31iqlT3128c2edci63dBizT7UHXT3sQ49/rSfBdZ7lQVTsE2H9vl/lXPqjKxRVMsqxY0jXDTsT8yHXexmAJWPncJW3t0u60/tjPNclVQ6cdXL39le6fa7WnWu3f9e+7fdUe3W3RElN1YspXVyFcu02MDAAAAAACcy05bT6nUCkblsb+VbGWvbJthr2//Ms1yL6j/lJqGhyK9tSJD3NaX7P3vuKml7ilVKU/JdNdbfXRr4HXpmCLWotCl2RoDAAAAAADAuea09pRKrXfJ5jbhj/n2+Z6lViXo6XhSNDq/+1G11Mojv7un7sVGxqS7H1U3bTi20w4mHbVtx/emqJZS7ylfhdwXZDqe3QkH7ffju93ryrnTBk6h9pSaUeMRS/SS7PZ1/89+P77H7tswxqZdNiTF9EMAAAAAAIDzyRnTU0rUdPzm0u0syZJt9dFtKZbpyXd+o3E1IX/yt4/cU/J8ahw+/8BK97pT0XqBz5/fMtU9tU80DW/SrgXudbHoAtawYNUMx6JjDN04weK9P7dN3eQ8uxSUDS7bJbDvD3ctzNH+AAAAAAAAzmZhqZRSCNN31b9SfNb3ghauZ1Nq3Uo0tLE759iW+D1plg0q3c5+PfK7ewKfejl9tW+ZXRRbzPYkHHIVVLeUbueql9oUqWVditW3qXv+Y7P3/2JXLn/USsQUcr2mFHjFRkTbkxV7p3nynt//6VhyvG09vscSLdl9dm2xv1j3dEIphWCpz+upCn1cAJUeTdurkruUrTu2w97fOdd6XdAsw2ovAAAAAACAc1lYQqkEL8lWHP4tzdS49EIp9Z+6tcxVNnTThDTL9BS916oMdGHT9D1LbO3Rba5Zuaqe2he5wloWrhFY9/EKf7X6BS52T+xTT6ffju1yx9Nnfyt1ZZrpgb41R7dZrogoKxZT0KrmLWNdi9W31kVqpbuumq/rJ1hw9VZqqvbSsXVuuxMP2dTd/7GeF/z5NEAAAAAAAIDzSYTnZdLJGwAAAAAAADjXe0oBAAAAAADg/EAoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFJnkd69e9v06dPPm+OmZ8eOHdasWTOLi4s76447Z84cu+aaa07quAAAAAAAOFtFn+4BANlRqlQpmz9//nlzXAAAAAAAzlVUSgEAAAAAACDsCKVCcPDgQXv88cftqquusg4dOthbb71lnue5ZUuWLLGWLVvamDFjrHXr1taxY0f7/vvvU2y/YMEC69evn7Vo0cI6d+5s33zzTWDZjBkzrFu3bm4ft912m23ZsiWwbMOGDXbjjTda8+bN3fGTkpJS7Hfu3Lluap22HTBggG3atCnF8ptvvtnefPNNu++++9w+NDZNQ8tKZsd94IEH3D6DXX/99fbFF18EjjlixAi74YYb3PYPPfSQJSYmumU6ts5R10lT4e644w7bunWrW6bpgf3797dOnTrZnXfeac8884xbT9cn+DhNmza1evXq2aFDh9KMO7PrnBFdk19++SXN54MHD7bJkydnedzMzlffkddff93atWvnzmv58uUptt27d6+7njrPa6+91saNG+c+nzp1qt11111pxrRmzRq3LwAAAAAAzgWEUiEYNmyYHTlyxKZMmWJjx461WbNm2ZdffhlYrmVRUVH21VdfuZDj5ZdfDiz79ddf7eGHH7Zbb73Vvv32WxdS5M2bNxAyPP300/bYY4/Z119/bVWqVHHr+oGGXjdq1MhtV6NGDVu/fn1gvytXrrShQ4e6wEnjad++vQtSUgdXGvN1111ns2fPtldeecXy5MmT6blmdVydnx9Aydq1a23nzp0uGPMtWrTIBTUff/yxLV261IVFcvz4cRfqffrppy4wKliwoD377LOB7SIiIuyjjz5yoVjFihXtiSeecOP3ffjhh+4nPZld58zUqlXLbZve/i6//PIsj5vZ+WocCtv+/e9/2/vvv28LFy5MsZ2Ct5iYGPv888/dd0braVsdV/c3NYVnGi8AAAAAAOcCQqks7N692/USUviTL18+K168uGtWrRApWK9evVwwpWqZzZs3Bz5X1YsqrBo3buyWlylTxr0WBUUNGzZ0QYPCiYEDB7owYvv27a6CaN26dda3b1+Ljo521VSFChVKs19V72i/PXr0cGPVNsE0HgVMWqdy5cop9pGerI6riiFVC/nVRQrnVOmTO3fuwDqq5ilRooT7ufTSSwMVXOXLl3cVQ/nz57fY2FgXUK1evTqwXbly5dx+1L9JodRFF11ke/bsyeoWZXmdQw2levbs6QI3hWwKGhUShiKj81Ulm5aVLl3aihQpYl27dg1so2qqefPmuYo0nbPOV8GiAsYKFSq4cFBVcwo6FSqmDsoAAAAAADjb0eg8C/50NwUWvuTk5BSBhSpyFOCIwiVVBPkUcNSpUyfdfStwUcjlK1CggAtr9LlCCYUVCsL8KqKiRYumGJeqchRi+BISEmzXrl1WrVq1FEFPduzbty/T4+o8Ff7MnDnTVVEpNFElWTBVQPmCr8f+/ftt5MiRbtwKfVTV5V83UZjk/+v/pK78ykhm1zkzCnlU4aTrmStXLvvuu+/cmGvWrGmRkaFlthmdr65l8PekWLFigde6Fjq34Puv5Qo0dc11fAWUGo/uh8JCBYGalggAAAAAwLmAUCoLqtpROKIKmuBqoFCVLFky0DcpNYUQmv7mUwWSAg2FQAqljh07ZocPH3YBkd6rB1HwftVP6vbbb8/0+H7QEyodO7Pj+lP47rnnHmvTpo1bHmoYNGrUKIuLi7OJEye66itVEqUOtFLze3fl5DpnpmrVqvbHH3+46YSqSPrss89c+HMyKpJ0LYOvXXDVV+HChd29UXWbH/ppuf/ar+D67bffrE+fPq6qbtu2bXbJJZfkeFwAAAAAAJwJmL6XBVWyaIrdSy+95AIVVbeoF5QanIdClS0KtNRPSNuqomfx4sVumRpyqxJGDbBV5fT222+7KidN97rwwgtdYKLm15rqpd5KBw4cSLHfadOm2bJly1zllgItVS9pP6GaNGmSdenSJcVnWR1XNEVNoYr6QWnKmSp7QqHrp6l7CrsUwIwfP95Olsyuc2ZUqVW9enWbMGGCm+6nc9d+TkYo1apVK1dJpumYqpr65JNPUhxXUyHVIF8h4MaNG91x/d5cCqW0raZc6vunIE/fDVViAQAAAABwLiCUCsHw4cNdQKNKmiuvvNI14FYlUSg0xe3JJ5+01157zW2rp7UpnBGFIUOGDHHVQm3btnX9ldT43A95nnrqKRdaaTtN5VJAEbxfNfZ+4YUXXPihnlJqkh1qQCQKm4Kf9ufL7Lg+9YNSA3RN5QvVoEGD3PQ0jVcNydXrKlQKaPTEPj0Jzz++3qvJeFbXOSsKoBSyKYBUMKXqJu0vlONmRqGjnqqnJ/PpyYIKoYLp3sfHx7vKMz1xUBVRfiil42scGo/Gpn5VNDkHAAAAAJxLIrxQ50cBQVSVpScRqoIHAAAAAAAgu6iUQrapukdT/2i6DQAAAAAAThSNzpEt6nukaXL16tWzbt26nfFX791337X33nsvw+V6eqGeugcAAAAAAMKL6XsAAAAAAAAIO6bvAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0IpAAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRyh1Fundu7dNnz79lB/n5ptvtgkTJpzQttdcc43NmTMnR8d/+umn7ZVXXsl0HR1DxzrVRo8ebc2aNbP69euf8DXZsWOH20dcXJyFw8kYMwAAAAAApxqhFM44Dz/8sN111112Jrj99ttt/vz5Vrt27QzXUVCowDAjpUqVcvvInz+/nSljBgAAAADgdCOUAgAAAAAAQNgRSoXg4MGD9vjjj9tVV11lHTp0sLfeess8z3PLlixZYi1btrQxY8ZY69atrWPHjvb999+n2H7BggXWr18/a9GihXXu3Nm++eabwLIZM2ZYt27d3D5uu+0227JlS2DZhg0b7MYbb7TmzZu74yclJaXY79y5c12FjrYdMGCAbdq0Kc00vDfffNPuu+8+tw+NTVPJQrFt2za74YYb3HYPPfSQJSYmus8feOABt89g119/vX3xxReB98uXL3dT69q2bWuvvfZa4FoFXy+t36lTJzfN7NVXX3XLZs6c6d43aNDARo4cmeIY2sfrr79u7dq1c9vpGKHSWHQNNB7dJ73XdT127FjgOgVPc7v//vvTnGN6Vq1a5cb7zDPP2Lp169xr/Xz11Vcprk3Tpk2tXr16dujQoRTbHzhwwO69915r1aqVXXnllXbLLbdYcnJyYLnWf+qpp9x3Ttfsnnvusfj4eLds3Lhx1rVrV7dv3dfx48eflO8zAAAAAADhQigVgmHDhtmRI0dsypQpNnbsWJs1a5Z9+eWXgeVaFhUV5cIIBQQvv/xyYNmvv/7qpqPdeuut9u2337pgJW/evG7ZmjVrXP+kxx57zL7++murUqWKW1cUEuh1o0aN3HY1atSw9evXB/a7cuVKGzp0qAtbNJ727dvb4MGD0wRXGvN1111ns2fPdn2a8uTJE9IXY9GiRTZixAj7+OOPbenSpS5YE51fcAC1du1a27lzpwtNfN999527Tv/+97/ts88+c8cOpjBI+1OQooBOgYxcffXVbtqZ/k1N10DT5LTP999/3xYuXGjZoeBP4dMHH3xgkydPtoiIiGwFW+mpXr26G++QIUPcvdNr/Sg483344YfuJz0KlnQtFMbpmiq41Lh8jz76qO3fv98FZv53yw+t8uXL5+7PvHnz7LnnnrNRo0bZihUrTsr3GQAAAACAcCCUysLu3btd0KDwR0FA8eLFXaWNQqRgvXr1csGUKos2b94c+Hzq1KmuIqVx48ZueZkyZdxrUVjTsGFDq1WrlsXExNjAgQNd2LR9+3bbunWrq77p27evRUdHu1ClUKFCafarChztt0ePHm6s2iaYxqNgS+tUrlw5xT4yo2ClRIkS7ufSSy8NVGGpMkcVPL/88ot7rzBDFWK5c+cObNulSxcrUqSIO1dVJ6miK5iqru6++243ltjYWLvsssuyHI/2oTGVLl3a7VtVQtlRoUIFK1eunBuTxlq2bFnbs2ePnU4KoA4fPuzuda5cuaxJkyaBUMr/3qlKrXDhwm65rqV/nXX+up+RkZEusLz44ott9erVJ+37DAAAAADAqRZ9yo9wlvOnu/Xs2TPwmapVVBnjU+WTgiNRuHT8+PHAMlUR1alTJ919KxRRKOArUKCAC2n0uSqlFEAoOBCFFUWLFk0xLlUwqcrFl5CQYLt27bJq1aoFPlMQcyIKFiwYeB18TjpPhWGq7lEYogoeVd4EK1asWOC1xqxpiMFUraWwKzv27duX4poHHyMUCm8UzOlH9G/qqrJw69+/v7uuCp5UEaUKMU3n01h1fxVEBX8/gikMVJWT1tN5HD161N3/k/F9BgAAAAAgHAilsqAnpynA0PSq4GqgUJUsWdJVwqRHwYqmv/lUgaSQQkGOQilN7VIljYIpvd+7d2+K/aqflJ60lhk/hDmZNI1M/Y3atGnjxpU6dFM1jk9jTh0gnciYdE2Cz/9kVDn5fZQUBAYHVLrmqSmYC+73FCx4yl126Gl8CqH0o6mZ6gumflqqRtP3Tv2jFDKmDvAULGlqn3pxqVJOIVafPn3S9IVKb8w5/T4DAAAAAHCyMH0vC6pU0RS7l156yeLi4lx4oV5QatgdCjU2VwCgHkjaVpVTixcvdsvU+Fz9l9TbSFUub7/9tqty0hS1Cy+80KpWrer6Dmm6m/r/qDF28H6nTZtmy5Ytc8GDAi1VL4VSLeObNGmSm2qXXZrOpyllzz77rOtllTqU+fTTT13lj6YhalpYcL+pE6Vm4KrK0j5VNfXJJ5/YyVK+fPlAfyk1eE+vN1OlSpVcZVp6wZRCI40rdSPzrOg7oWmRCpP8AElBlf+903Q+9YvStdR9VV8tBZXqB6Vt/Mo5VU2lnraZ0Zhz+n0GAAAAAOBkIZQKwfDhw10wpIbhasr9xBNPpFtNkx5NcXvyySfdU+i0rZptKwzwG2WrSbamv6lfkHoCqfG5H/LoyWsKrbSdek2ph1DwftUI/YUXXnCBjXpKqXl4dqp2FHIFP+0vO/TUNlX3aCpfago91LRb09OuvfZaF76FQn259PQ6hWtqRq7X6rMl2of2pScCar+qJjpZVGWk6iM9KU/VR+rxlZqOqafWqUeX+nwFU7WS+oR1797dXRe/sbtCNJ2D9itapvdqIi+qoFNvLe1TjfD1RMDatWun+N6p75aui6rSFPapKkph00033WSDBg1yn//0009Ws2bNkMeck+8zAAAAAAAnS4THs+BxAhQcqafRxIkTuX4AAAAAACDbqJRCtqnXkab+aQohAAAAAADAiaDRObJF/bE0HVFT1rp163ZGXD1NicuIpq1pmhsAAAAAADizMH0PAAAAAAAAYcf0PQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilziK9e/e26dOn29lux44d1qxZM4uLiwvL+Y4ePdodr379+jZhwoRsbavjaxxZefrpp+2VV145accFAAAAAOBcRyiFsCtVqpTNnz/f8ufPH5bj3X777e54tWvXPmXHePjhh+2uu+466ce95pprbM6cOSdhhAAAAAAAnFkIpQAAAAAAABB2hFIhOHjwoD3++ON21VVXWYcOHeytt94yz/PcsiVLlljLli1tzJgx1rp1a+vYsaN9//33KbZfsGCB9evXz1q0aGGdO3e2b775JrBsxowZ1q1bN7eP2267zbZs2RJYtmHDBrvxxhutefPm7vhJSUkp9jt37lw3tUzbDhgwwDZt2pRi+c0332xvvvmm3XfffW4fGpumzoUis20zO+6BAwfs3nvvtVatWtmVV15pt9xyiyUnJweWX3/99da0aVOrV6+eHTp0KMUxMztf/zr7Vq9e7fbhGzdunHXt2tXtW2MdP368nSwah8ajcd1www22fv36wLKZM2e6KXoNGjSwkSNHZmu/mV2rhx56yO1X13zo0KHuda9evQJTCvv372+dOnWyO++805555hn33dN3CQAAAACAswWhVAiGDRtmR44csSlTptjYsWNt1qxZ9uWXXwaWa1lUVJR99dVXLhB5+eWXA8t+/fVXN7Xr1ltvtW+//dZef/11y5s3r1u2Zs0a14voscces6+//tqqVKni1hWFXnrdqFEjt12NGjVShCErV650YYVCI42nffv2Nnjw4DTBlcZ83XXX2ezZs13Pozx58oT85Uhv26yOq3Do2LFjLqz54osvXBgXERER2OeHH37oflLL6nyzki9fPhsxYoTNmzfPnnvuORs1apStWLHCTgaN4/LLL3fjaty4sT3yyCOBUPLqq692U/T0b3Zldq10Dtqvpjo+9dRT7vXEiRMD22q9jz76yAV5FStWtCeeeMLdLwAAAAAAzhaEUlnYvXu3CwQUwij4KF68uOvzoxApmKpYFEypmmbz5s2Bz6dOneoqrBRmaHmZMmXca1HY07BhQ6tVq5bFxMTYwIEDXeizfft227p1q61bt8769u1r0dHRrpqqUKFCafaraiHtt0ePHm6s2iaYxqOgR+tUrlw5xT6ykt62WR1XYcnhw4fd+HPlymVNmjRJEUplJKvzzYqqpDTGyMhIF2hdfPHFrprqZChcuLCrcNO4NL61a9fatm3bcrzfE71WUq5cOcudO7cLrRRKXXTRRbZnz54cjwkAAAAAgHCJDtuRzlL+lLWePXsGPtMUK1U1+VT5pMBCFC4dP348sGznzp1Wp06ddPetEEEhl69AgQIWGxvrPlcljkIHBWGisKJo0aIpxrV06VJXreRLSEiwXbt2WbVq1VKEFycqvW2zOq6mlen8Nf1s//79roJIU9QUFmVm3759mZ5vVlS5pio2jU9VW0ePHnXjOhk0Dj8s0vh0j/bu3WsXXnhhjvZ7otdKFAj6//o/qavkAAAAAAA4kxFKZUGVKPqDX9OrFJpkV8mSJV0lTHqKFSvmqm586rGkkEIhiEIpTe1SJY2CEL1XEBK8X/V10hPeQgkvTkR622Z1XD1RT8GKfjTtTT2n1G9JvZ4yo3PO7HwVBAWHLlrPpyDq0UcftVdffdVVcCnU6dOnT2CKnU+BYXB/q1BpHNqXX9mke6R7F6qMjhvKtQoloPKlPl8AAAAAAM5kTN/LgiqZNMXupZdesri4OBeMqBeUGm+HQtO+FGgtXLjQbavKqcWLF7tlanz+3Xff2fLly11Vz9tvv+2qjUqXLu2qcKpWrer6DiUmJrp+QWqMHbzfadOm2bJly1zgoUBLvYmyUx00adIk69KlS8jrh3Jcnacanysg8cMYhS9Zyep8y5Yta/Hx8YFpgurfFdzTS8fzK6tUNZV6GqNUqlTJVXllN5hSFdOnn37qxqXxaZyahhmqjI4byrXS90/fNwAAAAAAzjWEUiEYPny4CyTU9FtPSVNT6eBKncyov9GTTz5pr732mttWT7VTuCXVq1e3IUOGuEbqbdu2dT2Q1PjcnyqmBtcKrbSdek2pZ1LwftUY/IUXXnBPb1NvJz3lL9SeRKLQJ/hpf6GeT2bHVVXY3Xff7fpRqbm7nqZXu3btQJCkp8jpCXyiJxnq/aJFi7I83yJFirin02nfqijSVMfg0Oemm26yQYMGWZs2beynn36ymjVrpjtdTk9S1NjUGypUGodCOI1LY9U4g3uJ6RwUzE2ePNm9Vm+wUI6b2bXy6fuifauhvM4bAAAAAIBzRYTHnB8AAAAAAACEGZVSAAAAAAAACDsaneO8pul2GdFUO00JBAAAAAAAJx/T9wAAAAAAABB2TN8DAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0Kps0jv3r1t+vTpp3sYZ53Ro0e7p+zVr1/fJkyYcMYdd86cOdaxY0e3rrY51z399NP2yiuvnO5hAAAAAABOM0IppEvhl0Kwc8Htt99u8+fPt9q1a5+Rx33ppZfs1ltvdetqm5PlmmuucYFXuO9vZseVhx9+2O66665TcmwAAAAAwNmDUAo4zbZv324XX3zx6R4GAAAAAABhFR3ew52dDh486KpZFi1aZFFRUda1a1cbMGCARURE2JIlS2zw4MF244032vvvv2+5c+e2Rx991Bo0aBDYfsGCBfbmm2/ab7/9ZoULF7Y777zT2rRp45bNmDHD3n77bdu7d69deumlroqkbNmybtmGDRts+PDh7t/WrVtbUlJSinHNnTvX7Xfbtm1WpUoVe+SRR6xChQqB5TfffLPVrVvXVq9e7cZZoEABe+edd6xUqVIZnuuqVats4MCB7liJiYluSpn885//tHbt2mV6XFXfTJ482Z1LxYoVrUyZMvbNN9/Y/fffbx06dHDLx40b585z1qxZbp3HHnvM/Sva7plnnrEff/zR8uXLZ9dff7317ds3MDb/Wv/jH/+wUaNG2YEDB9w6up7a78cff2y7du2yQoUKuSqgPn36nJT7r3uk8923b58VKVLEbrvtNrvqqquy/G5kpX///rZx40ZLTk62m266ySIjI+2vf/1rSNVSmZ3vQw895MZz7NgxGzp0qNuvvlMTJ07M0f31v1NVq1a1FStWuLE3atTInnrqKYuOjs70uDJz5kw3dS8+Pt7dN30vUl/njH4XMjsuAAAAAOAs5SFL99xzj/fggw96cXFx3q5du7yePXt6M2fOdMt++OEH7y9/+Ys3duxYLzEx0Rs9erTXq1evwLa//PKL16xZM2/hwoVu+datW91rWb16tdekSRNv2bJlXnx8vDdy5EivX79+bllycrI7zhtvvOElJCR4H330kVe3bl1v2rRpbvmvv/7qttXxtd/Jkyd73bt3d699AwcO9Nq1a+ctWrTIfb5u3Tpv//79Id1xHSf4PHyZHVfb9O/f3zt69KjXoUMHb+LEie5c//73vwf2qXP45JNP3Dnp3Hr37h3Y9+DBg70hQ4a47Tds2OC1adPGmz9/fmC5jtmgQQNv6NCh7jyOHTvmrq9MmTLFnV9SUpK3YsUKr2HDht7y5ctTjF3XY/z48dn6xmss9evXd9dQtm/f7q1cuTKk70aox9U1WbVqVbbGFcr5durUyZs9e/ZJu7/+uXTt2tX7448/3E/r1q3THCOz48qwYcO8ESNGpPgss9+FUI8LAAAAADi7MH0vC7t373a9fu677z5XvVO8eHHXM+frr79OsV6vXr1cpUzz5s1t8+bNgc+nTp3qqmoaN27slqt6SK9l9uzZ1rBhQ6tVq5bFxMS4CpaVK1e66Vxbt261devWuUohVYN069bNVcSk3m+9evXcfnv06OHGqm2CaTyqKtE6lStXTrGPE5HVccuVK+eqxVSNpQqoiy66yPbs2RPYXpVinTt3duekc1MVlypyVLUzb948V3Gm7bVt+/btXUVVMK139913u/OIjY21yy67zH2uCiWdn6pzatSo4abDad85pSomneeWLVvs8OHD7rwuueSSbH03ToVTdb6hfK9UUVWiRAn3o4qmTZs25fi4mf0unMrjAgAAAABOH+a+ZGHHjh3u3549e6YIKjStyZc3b97ANCL9QX38+PHAsp07d1qdOnXS3bfCGgUZPk2vU9Cizz3Pc+GMwg7RdLCiRYumGNfSpUtThDYJCQluOle1atUCnykkOpkyO64oyPD/9X+Cpx3qHPypbTo3/3xz5crl1gu+HsWKFUsR8EmePHlcKJHal19+aWPHjnXj036OHj3qxpVTurcjRoywSZMmuSfj6Xo++OCDLggK5btxqpyq8w3le1WwYMHAstTf9xOV2e9C6dKlT9lxAQAAAACnD6FUFlQZo2Dliy++cCFRdpUsWdJVPaVHocvatWsD7w8dOuT+0FZwo1BKvXlUnaPwRu/Vayd4v+ojlFX/IT8kyq6MeiJldlz1jEqPxu7TOei99q9z889XFVQaq6py/PBNgURwEJfR+ShIUR+vV1991VX4qHpI/ZWCj+sHGQqNskuVbfpROPPiiy/a888/7wKhUL8bJ3rcjIR6vvr8ZN7fUGV23Ixk9rsAAAAAADg3MX0vC6re0LQiNbOOi4tzVSlr1qxxTbdDoalqCi0WLlzotlXl1OLFi92yFi1a2HfffWfLly93gYeaPKsaRZUhF154oWvsrIbWmrI2ZcoU19g7eL/Tpk2zZcuWucBDf8SrkXR2qmVU/dOlS5d0l6kaSVOntN/U55OT4+7fv98+/fRTd046N00705RGVZo1bdrUxowZ48I4NbPWdWvZsmWW+zxy5IgLZPwAQ1VEqacxSqVKlVwVUHYCIo3322+/dZVICnIUuOTPnz9b340TOe7JOF+NT+MJ5/3N6rgZyex3AQAAAABwbiKUCoGegKcQ5brrrrMrr7zSnnjiCVflEwpN83ryySfttddec9vqKWIKMKR69eo2ZMgQGzZsmLVt29b1BNLTyfwqFj1dTH+oazv111EPoeD96ulkL7zwgrVq1cr1/tFT/kJ56ptPIZd6JaVHFTiqDurevbt7cp56/pyM4+ocFHjonBTU6dr42+pa6MlsHTt2dE/UU/VPKKGUQh89vW7QoEHuqYY//fST1axZM92n3elpeeqzFfxUv8wo/Pnggw/s6quvdvtW+KPpe9n5bpzIcU/G+eq7pkBJvbn0RMBw3N/Mjqu+a3ran5bpKY16rd5RofwuAAAAAADOPRHqdn66B4Hzg6b3TZw40SZMmHC6hwIAAAAAAE4zKqUAAAAAAAAQdjQ6x3lNU8gyoql2miIXbu+++6699957GS7Xk/H0tEIAAAAAAM5mTN8DAAAAAABA2DF9DwAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0IpAAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIRd9Knc+WdPfG87/m9vusva3FvHcuWLthlP/ifd5UXLF7BuzzRN8dmauVts3psr3OuICLOer7S0/MXy2LaVezLcT7CeL7ewNXO32k9T1gXeFyiRN82+OzxS38pcWizd/UbHRlnBUnmtWsuL7LKrygc+X/rR2sB+U7usfXlr1P/SLMcHAAAAAABwvjiloZQvMjrCilUomOKz2Pwx5nle4H2BC/JY7oK5Au8LlcqXZj9r5m0NvNama+dttSu6VrFceaKtRJVCgWV7Nh205ETPYvJEWeEL8wc+j4o+8cIwjS82fy47sD3O9m4+ZN/9e6XlLZzLKjYonWbdYuULWGTM/45VoESeEz4uAAAAAADAuSgsoVTewrmt8/DGaT5XJZJP4VLVFmUz3MehP47YjlV/Vl0Vr1TIdm84EAililcslGL/k+6aY3G7j1qxCoWs0z8bnJRz8Md3cOcR+/Deue6z7f+3L91Qqs19dQIVWAAAAAAAADiLe0q5KinPLE/hWGs2oIb7TAGRH1SdDvmL5T5txwYAAAAAADibhaVSSlVLb/eemeKzAROuTvFe/Zz8nk5yRbcqVrfHxe61pvmtm//n1L3KjUu7qYBFyxWwvb8dcmFVqepFT/k5/PTJOlv59W9u+p5cdEUJu6RtuXTX/eDuPyupfH6PKgAAAAAAAJzmnlKppe4plS+oCmn7/+21Q7uOutcXN73Q/VulaRn7z4TVtvH7Hdb4hktdA/JQqEH6iaxz6I+j7kd0rPJ1S1pM7vQvX+qeUup5BQAAAAAAgDOkp1SwzHpKqXeU7/Mnv3f/Jif92SQ94WiibfxhRyCsykpweHU8LsEKlPjz9bG4hMDnMbFpL0vzQTVdEDXvzeW2eekftvCdX1zQVqLS/xqs++gpBQAAAAAAcJb3lEo4luiqoXzxRxLdT+LxpMBna+f+L7TKSrHy/6vYWvnVZkuMT7Ij+4/b+gV/7iMqJtIKlUn75D//iYFNBtSwqFyR7ul/P3609gTPCgAAAAAA4Px2xswrU8+m/5v1W+C9psZ1eLi+C6T8AKr7802tSNkCgXV+mbnJFr//f7Z95R6L23PU8hfLk+VxylxWzEpUKWS71h2wNXO3Bhqo+y5rXz7DaXmSt1CsVWtR1vWX+v3nXbZn08E0UxO/efHHFNP3Sl9S1Or3qh76xQAAAAAAADjHnTGVUurXpKDI/9m94YD73IVGZlaodL4UgZRUqF/S/auqpeApfpmJiIywq//xF6txdQXXxyoyMsJVPqkPVKMbLrW/9KyW5T5qdqpoEVF/Np76+dP1aZbv2Xwoxbkc2HY4pLEBAAAAAACcLyI8PdoOAAAAAAAAOB8rpQAAAAAAAHD+IJQCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0IpAAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCqWyYM2eOdezY0Zo1a2ajR4+280Xv3r1t+vTpdibYsWOHu/5xcXGneygAAAAAACAHCKWy4aWXXrJbb73V5s+fb7fffrudLNdcc40LvJC1UqVKueufP39+LhcAAAAAAGcxQqls2L59u1188cWn7m4AAAAAAACcJwilQtC/f383ZSw5OdluuummNNP35s6d66a4tWzZ0gYMGGCbNm0KLBs3bpx17drVmjZt6qb+jR8/PrDsoYcecvvSlLShQ4e617169Qosv/nmm23ChAmB9/fff7+9+eabgfdLlixxx/ziiy+sU6dObvtXX33VLTt48KA9/vjjdtVVV1mHDh3srbfeMs/zQvpSbNiwwW688UZr3ry520dSUlJg2QMPPJBiDHL99de7MfhjHjFihN1www1ue51jYmKiW6bzvO2226x169ZurHfccYdt3brVLdP0QF1nncedd95pzzzzjFtvxowZKY6j61ivXj07dOhQmnEvWLDA+vXrZy1atLDOnTvbN998E3Kl2iuvvOL+bdu2rb3++usprlVm13nv3r3ummis1157rbvfwTTOp556yt0D7eOee+6x+Pj4kO6Rzl3noeuof7/88svAsgMHDti9995rrVq1siuvvNJuueUW9/0M3rZbt27umLrmW7ZsSTEu3Sfdx/vuu8/tX99N3R8AAAAAAMIlOmxHOouNHTvW/asw5N1337Vq1aoFlq1cudIFSv/617/siiuusE8++cQGDx5sH3zwgUVFRVm+fPlcSFOxYkW37sCBA61WrVpWs2ZNe+6559w+FIYocFKAkF3Hjh1zYYzCrty5c9u6devc58OGDbNcuXLZlClT7OjRoy4Auuiii6x9+/aZ7k+hyMMPP+zCjrfffts+/fTTFP2kFF68/PLLNmjQIPd+7dq1tnPnzhRjX7RoUSC4Usim8Wn58ePHXfjy/PPPW0xMjAtknn322UDAExERYR999JF1797dmjRpYk888YS73tpGPvzwQ9u2bZsLf1L79ddf3bi1vwYNGrgxBYeDWfnuu+/cfda1UrB4ySWXpDinjK6zwjOdy+eff+4q6RT2VKhQwYVn8uijj1p0dLQLF/PmzesCTD88yuwe6Xi6PvpeNWrUyAVG+/btC4xH4ZfWmTlzpkVGRtoPP/zgrp+sWbPGnn76aXvttdfceej66tr432OfjvvYY4/ZCy+84K5Vnjx5Qr5eAAAAAADkFJVSOTR16lRX6aLASiFUjx49bPfu3YHQQlVSlStXdsFBjRo13PS/1atX28miKqS7777bChUqZLGxsXbZZZe546vvkqpgFIoVL17cBV9ff/11lvtT5ZLG3rdvXxemqNpG+/YpbFH1zy+//OLeq3pHVUIKanzt2rWzEiVKuJ9LL700EA6VL1/eVRqpH5TGqrAp+FqUK1fO7Ud9oxTiKaDZs2dPtu5D48aN3X0oU6aMex2qLl26WJEiRdx2qpZK3eMrveusz+bNm+eqyjRujVmB0qxZs9w2/n1QtVjhwoVdAKV9a92s7pGCK52HKpwOHz7srokCJp8CKH2u+6X9KsTzQ6nZs2dbw4YNXfipwExBqAJRhWbBVCGlwEvH0Xc0+D4DAAAAAHCqUSmVQ6pgWbp0aSCIkISEBNu1a5erqFJoowoVradpcKqI0fKTRdUtCn9Sj0l69uwZ+EwhR5UqVbLcn6pxFJooKBEFHUWLFg0sV1Cl8EcVOgrZvvrqK1fxE6xgwYKB1wpFVCEl+/fvt5EjR7rrdeTIEXc9tD+fwhH/X/8neOpgZlQZVadOHTtRxYoVC7zW+WoKY1bXWeej8SlQCt7P5s2bA/dBgVHw8lDvkaqqVGE3adIkN1VUgd2DDz7orrloqqOuqwIvjePqq6920/kUfirICz5mgQIFXJCmz0uXLh34XPsEAAAAAOB0IZTKoZIlS7p+Uuk9jU/Bg6ZvafqUKqkUGPTp0ydNbyd9nh4FCcGhjCpjUvODnGCqqtHn6oEUXMEUCgUymhamYymY0ljVNymYpvCpN1KbNm3c8lDDoFGjRllcXJxNnDjRVeVoKlvqQCu1UPtg6T74/alOhCqXfDrf4CAuo+us6id9rm399RX8+K91H9Q/SgFl6kArlHukSi/9KMR88cUX3bRHfwqeqs0UQuln/fr1bsqhpi2qkk3BmKZV+lTZpgArlHMCAAAAACBcmL6XQ2pAPW3aNFu2bJmrdFEAoCoiBQmqBlKo4ocBqpryp/UFU1WL+gClpuluy5cvd6/VS2nFihUhjUn70/Stl156yYVACra0fzXsDqYqHE1bC3bhhRda1apVXc8iTU9T3yE11Q6mKXkKZNS/SdPV/GljWdFYFKYo7FJ4E9z0/WTcBwU8CxcudOeryqnFixeHvL16Z6niSFPcNIVOzdKzoiovhUBjxoxxQd7GjRvdGPxeVLoPmlan3mHat74T3377rVs3q3uk9bWuKut0fRVc6tr5dJ6aFqnvl6rR9N3zl2vs6pGl746Oqd5gqtoLrpICAAAAAOB0I5TKIU2nUhNpNYtWc3D1lFJDbAUJlSpVck/rU1NwVRX99NNPrsF5amqOrSBLAY8qXnyqqlK1lZ46p2or9QgK1fDhw12odN1117mns6lpeOpKK4VNqZ/KJnpanEINbadeROo3lJr6QalCR1P5QqXroKltuk633nqr62cUKk0T1FPvdC384+u9mqr79+HJJ590zb01bl1ThT2hUkCkJ/f5TwDUGEMxZMgQVw2l6jE9NVD3LLhBuu6DqsLU8F3fAYVffmVcZvdIYZOa5WtanrZTmKnpez5VhanHlfpC6Vqqr1Xt2rXdsurVq7txqQpNPazUt0uNz0MNDwEAAAAACIcIL9T5UUAQhWiaSqapeGe7nDz9EAAAAAAAnBgqpZBtqgzS1D9NmQMAAAAAADgRNDpHtqhnkqbJqXF7t27dzvir9+6779p7772X4fLgpyYCAAAAAIDwYfoeAAAAAAAAwo7pewAAAAAAAAg7QikAAAAAAACEHaEUAAAAAAAAwo5QCgAAAAAAAGFHKAUAAAAAAICwI5TKhjlz5ljHjh2tWbNmNnr0aDvXnW/nCwAAAAAAwifC8zwvjMc7q3Xu3NkGDhxonTp1Oqn7veaaa+z++++3li1b2rlwvtOnT7eJEyfahAkTTtnYAAAAAADA2Y1KqWzYvn27XXzxxXa+ON/OFwAAAAAAhA+hVAj69+/vprAlJyfbTTfdlGY629y5c613796u0mnAgAG2adOmwLJx48ZZ165drWnTpm4q3Pjx4wPLHnroIbevHTt22NChQ93rXr16BZbffPPNKaqNVE315ptvBt4vWbLEHfOLL75w1Uza/tVXX3XLDh48aI8//rhdddVV1qFDB3vrrbcs1KK4zM7XP6Zv9erVVq9ePfd61apVbt1nnnnG1q1b517r56uvvspy2+Bz1jned9991rx5c3fNdH2yus5ZVW717NnTXQ/t84YbbrCNGzeGfNwZM2ZYt27d3HFvu+0227JlS4ptFyxYYP369bMWLVq46rJvvvkmsCyzMR84cMDuvfdea9WqlV155ZV2yy23uGvu03G1P41H/3755ZeBZXv37rUHHnjAWrdubddee637ngXL7LsBAAAAAMAZQdP3EJq6det6q1atSvHZr7/+6jVp0sT74YcfvMTERG/y5Mle9+7d3WuZMmWKt27dOi8pKclbsWKF17BhQ2/58uUp9tGpUydv9uzZaY43cOBAb/z48YH39913n/fGG28E3uuYDRo08IYOHert37/fO3bsmPfLL7+4Zffcc4/34IMPenFxcd6uXbu8nj17ejNnzszx+eqYLVq0CLzXcq0XbNq0aV6vXr3S7C+UbXXO7dq18xYtWuSuoa6dzi2r65wZjUfH+eSTT7yEhAR3DXv37h3ScVevXu2Ou2zZMi8+Pt4bOXKk169fv8B2ut7NmjXzFi5c6LbbunWrey1ZjXnUqFHeLbfc4h09etQ7fvy4t2DBAi85Odkt02f169d345Ht27d7K1euDBx38ODB3pAhQ9x6GzZs8Nq0aePNnz8/pO8GAAAAAABnAiqlcmjq1KmuGkkVP1FRUdajRw/bvXu3qxQSVUlVrlzZIiMjrUaNGm46nCqETpbExES7++67rVChQhYbG2uXXXaZO/78+fNd1U++fPmsePHirm/V119/bWcDVQY1atTIXU9dO51bVtc5K4ULF3bVRtHR0da3b193D7Zt25blcWfPnm0NGza0WrVqWUxMjOuxtXLlSje1UfxxNW7c2G1XpkwZ9zp4WUZjjoiIsMOHD9vWrVstV65c1qRJE/eZqGJK26gqS+uUKlXKLrnkksA9nzdvnt14442WO3duq1ixorVv395mzZqV5XcDAAAAAIAzRfTpHsDZTlO8li5dmiIQSEhIsF27dlm1atXclKuxY8e69ZKSkuzo0aNu+cmSJ08eK1GiRJoxiaas+RRyVKlSxc4G5cqVy/Z1zkrRokUDgY+COoU0e/bscSFSZsfVOgr1fAUKFAhsW7p0adu5c6fVqVMn3WNmNWZNkzx+/Libxrl//367+uqr3XQ+BZh58+a1ESNG2KRJk9zUSY3twQcfdMGm1tV3KXhcxYoVs82bN2f53QAAAAAA4ExBKJVDJUuWdD2Dbr/99nRDiUcffdT18lG1jMKGPn36pOntpM/To/BD4YNPFTOpqZomNVXV6HP1E1IlzckUypj88OdEts3onDK7zqFQDyZdd786SWGQgqqsjquwZ+3atYH3hw4dSrGtxqVKp/RkNeb8+fO7EEo/69evdz2nGjRo4PqPiSqu9KMg68UXX7Tnn3/eBZyq+tJYVXXlj0MhWSjnAwAAAADAmYLpezmkKWHTpk2zZcuWuWokhRYzZ850QcKRI0dcEOKHBaqaSm+6mSpe1qxZk+bz8uXL2/Lly91rTTVbsWJFSGPS/jTl7KWXXrK4uDgXBGn/an4dTFU4Xbp0ydb5li1b1uLj4wPn4TcxD6bqHE1v07XI7rYncp1DoeqiTz/91E1pU1NwTaMMrpLKiJqXf/fdd+4+6Fhvv/22q3JSlZQ/LoV/CxcudNdZlVOLFy8OaczaRo3P9R3R1ECto6DKH++3337rKusUpCm49JdpCqKCqzFjxtixY8dc03aNIbiJPAAAAAAAZzoqpXJI06kefvhhe+GFF+z33393lUmqimrbtq1VqlTJPb1u0KBBLnho166d1axZM80+9OS3Z5991qZMmeKCGwUfoqoqTdm6/vrrXY8j9TUK1fDhw+2VV16x6667zoVjmv6lSpxgevpb6ifJZaVIkSLuKXHqVaRg5oorrkizjs5fFT7du3d3AYqeEqcnzIWy7Ylc51Do+ikcUsVRhQoV7Mknn8ywoitY9erVbciQITZs2DDbt2+fe//0008HttW4tK/XXnvNjU9VTHfeeWdIY1aFlaqfVMVVsGBB1yOqdu3abpm+Lx988IG7j6IgTOPw6bW+M3pKoKbp6btCKAUAAAAAOJtEqNv56R4EcCpNnz7dJk6caBMmTOBCAwAAAABwhmD6HgAAAAAAAMKO6Xs46zVr1izDZX379g2pdxQAAAAAAAgvpu8BAAAAAAAg7Ji+BwAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAg7QqlsmDNnjnXs2NGaNWtmo0ePPnV35Tz39NNP2yuvvHK6hwEAAAAAAE6hCM/zvFN5gHNJ586dbeDAgdapU6eTut9rrrnG7r//fmvZsqWdLXIy5rPxfAEAAAAAwMlFpVQ2bN++3S6++OKTfAsAAAAAAADOP1RKhaB///62ceNGO3r0qMXGxlpkZKT99a9/tdtvv90tnzt3rr355pu2bds2q1Klij3yyCNWoUIFt2zcuHH28ccf265du6xQoULWu3dv69Onj1v20EMP2aJFi+zYsWOWK1cut9+yZcvaxIkT3fKbb77ZVRNpG1F1UdWqVW3QoEHu/ZIlS2zw4MH2j3/8w0aNGmUHDhyw66+/3u688047ePCgvfTSS27/UVFR1rVrVxswYIBFRERkeb7az2OPPWY///yzW1/HfO2119z4shpzTs535syZbupefHy8Ow+db7BDhw65aX0LFy60I0eOWO3ate355593+8rM9OnT3bguvfRSmzVrllWsWNGdn/716VrXrVvXVq9e7a5rgQIF7J133rFSpUrZjBkz7O2337a9e/e6fTz88MNu3L4FCxa4+//bb79Z4cKF3fVv06ZNlt+NzK6z6Ljadt++fVakSBG77bbb7KqrrnLLNJZnnnnGfvzxR8uXL5+7Xn379g2MKbPvBgAAAAAAZwRN30No6tat661atSrFZ7/++qvXpEkT74cffvASExO9yZMne927d3evZcqUKd66deu8pKQkb8WKFV7Dhg295cuXp9hHp06dvNmzZ6c53sCBA73x48cH3t93333eG2+8EXivYzZo0MAbOnSot3//fu/YsWPeL7/84pbdc8893oMPPujFxcV5u3bt8nr27OnNnDkzpPMcNWqUd8stt3hHjx71jh8/7i1YsMBLTk4Oacw5OV/fsGHDvBEjRqT5XOc0ePBgb9++fW5cX331lRtjVqZNm+bu3SeffOIlJCS4a9i7d+8017pdu3beokWL3L3TOeiarl692t3fZcuWefHx8d7IkSO9fv36BbbT9W7WrJm3cOFCt93WrVvd61C+G5ldZ31Wv359Nx7Zvn27t3LlysBxdR2GDBni1tuwYYPXpk0bb/78+SF9NwAAAAAAOBMwfS+Hpk6d6qpX6tWr5yqSevToYbt377Z169a55apQqly5sqt+qVGjhpv+p2qckyUxMdHuvvtuV5WkKq7LLrvMHX/+/Pl23333uSqa4sWLuz5OX3/9dUj7VNXO4cOHbevWra4KqUmTJiFVWJ3K8/XPSdVWqkbSuNq2bWu5c+cOaXtto55g0dHRrqJIY1L1UrDmzZtbo0aN3H3UOeiazp492xo2bGi1atWymJgY11Ns5cqVbipn8P1v3Lix265MmTLudSjfjcyuc3Jysttmy5Ytbh1VbF1yySWBez5v3jy78cYb3fmr4qt9+/auCiyr7wYAAAAAAGeK6NM9gLPdjh07bOnSpSkCgYSEBDd9rVq1avbll1/a2LFj3XpJSUluCqCWnyx58uSxEiVKpBmT9OzZM/CZQg5NHwt1uuLx48ddALR//367+uqr7d577w1MK8vMqTpf7U/BjQK2E1G0aNFA4KOgTiHNnj17XIjkK1euXJrttE7wMTWtz9+2dOnStnPnTqtTp84JfTcyu8558+a1ESNG2KRJk9yTHjW2Bx980AV9WlfXNnhcxYoVs82bN2f53QAAAAAA4ExBKJVDJUuWdH2T/P5SqUOJRx991F599VVXLaOwQf2VUj/wMKOwR+GHwgefKmZSUzVNaqqq0edffPFFyJVEwfLnz+/CEf2sX7/e9aJq0KCBNW3aNNMx5/R8M6NzUq8pBTonErSoB5PG4VcnKQxSUJXVtVTYs3bt2hR9rYK31f1XpVN2vxuhXGdVXOlHQdaLL77o+mcp8FPVl8aqqit/HArJQjkfAAAAAADOFEzfyyFNCZs2bZotW7bMVSMptFDDbgUJasatIMQPC1RF5E/dCqaKlzVr1qT5vHz58rZ8+XL3WlPNVqxYEdKYtD9NOVOj87i4OBdsaf9qfh1MVThdunRJs70aiW/atMmNXVPWdF4KULIac07PN6tz0vS25557zlUK6fp+++23rml6KLTNp59+6qa0qem5phUGV0llpEWLFvbdd9+5+6BjquG5qpxUJeXff4V/uma6zqqcWrx4cZbfjayus8ar81OlmYI0BXn+Mk1BVHA1ZswYd/5qwq8xqCk+AAAAAABnCyqlckjTqfQ0thdeeMF+//13V5mkKiH1O6pUqZLddNNN7ml5Ch7atWtnNWvWTLMPPfnt2WeftSlTprinuin4EFUZacqWnpqmHkfqaxSq4cOHuyfVXXfddS4s0vQvVeIE0xPZ1LMoNVX+qCpH1UUFCxZ0vYv0pLusxpzT8+3Vq5cbjyqiFMSoJ1P16tXtrbfeSnFOWs9/+l5w9VZmdP0UDqniSE+/e/LJJ0Pqk6XjDxkyxIYNG+aegqf3ekKgv63uv/alp+bpe+A/fS+r70ZW11nX74MPPnDnLArCNA6fXusaduzY0U3T03eFUAoAAAAAcDaJULfz0z0I4FSaPn26TZw40SZMmMCFBgAAAADgDMH0PQAAAAAAAIQd0/dw1mvWrFmGy/r27RtS7ygAAAAAABBeTN8DAAAAAABA2DF9DwAAAAAAAGFHKAUAAAAAAICwI5QCAAAAAABA2BFKAQAAAAAAIOwIpQAAAAAAABB2hFIAAAAAAAAIO0IpAAAAAAAAhB2hFAAAAAAAAMKOUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIRd9Knc+Z4hd1rCLz+nu6zww09ZRP4Ctu/hu9z7gncPsbxtOrjXiTu32+4B17vX+Xr9zQr0vsm9Ttq72+ImvGfxP/3gXkfkyWNRxYpbdIUqVuj2Bywid2633o5rmrl/c1/Z3grfOzRwzKRdO+3wR+Pt+I/fW9LuXRYRE2PR5SpY7pbtLO/VXSwiKiqwrr8PKfrcaMt1aS33+sjMqXbwtZHudfG3P7TokqVP8lUDAAAAAAA490WH5ygxFlP54hQfRRYoaJ7nhbwLLznZ9j3+oCVuWGsWFWXRF1UwL/64JW7dYombNljBAXcGQqn0JGxcZ3uH3m3eoYNmEREWVfpCS447ZAmrV7qf4z8ssiKPPmcRUWkvyaGx/8+KPTsqmycNAAAAAACA0xpKRRYtZsVGvJnm8+Mrfgp5H0lbf/8zkDKzIsNftNhaddxrBVPHl35vEbnzZLitwq8DLz75ZyAVk8uKPj7CctW8wgVdB994yY7OnGrxP/7Hjkz/yPJ1+Wua7RN+XWbHlyy22HoNQx4vAAAAAAAAzoWeUl5y4OXRWTMt/tdllnz0iEXkirXcjZpbRGxshpsmblrvfiRP89YukJKIyEgr8Ldb3TRAt9/ZX6XZNrrSnxVeh97/f9mq7AIAAAAAAMBprpRK/mNHih5NUmr6/GztI+qiCparVh2LX/6jHfv2C/djkVEWU/USy9uxm+Vp2TbDbRO3/JYmZPJF5slrUaUutMSN61Ks54u9or5F5M3nemMdmz8rW2MGAAAAAADA6ayUUk+papem+JGIEDaNiIgI/FvksRFWYMCdFnNJTbPoaLPkJEtY9YsdGDncjnwzI8TBpHPU/x4jIwX63+z+jRv/jnlJSSEeBwAAAAAAAGdkT6ngPlBe3MH/vVbvp3TW0dPy8nW+3v24XlI/L7EDIx437+hRO754fuDpfalFly0XeJ24YU2KZZoCmLR9S5r1guW6pKbF1m9ix/+z0I6GHH4BAAAAAADgjOwpFaUQKCZXoJ9T0r495h0/bkc+nxJYJ7piFfdv0q6dFvfBWEva/Yd7r15SsbXqWkS+An++z5c/w+NEV6jsftxx5s2y+P82WFej80Pvve5CLcnTql2G+8ivaqnISEtcnzLUAgAAAAAAwBlaKZUR9XPKd013Ozxlonuy3q7+Xf6cSvffhuIx1S5zfaTEO3bM4sa95X4iLyhlkQUKWtKObeYdjnPb5GnWOsPjaOpfoXuH2t6hd5sXd8j9G1WmrCUfOmjewQNunVx16lveTj0y3EdM+UqWu3kbOzYnbTN0AAAAAAAAnEWhlOS/4RaLLH6BmxbnGo0nJ1lU8QsstmEzy9/rb+4JeRJZrITl7/N31+g8cdsWS9y80SKio/9sdH7t9RZbr2Gmx4mpdLEVf/k9i/tonMX/+L0l7dzupgNq+9wt21neDl0tIioq87H2+bsdW/CtWWLiSb0GAAAAAAAA55sIz/tvWRIAAAAAAABwPvSUAgAAAAAAwPmJUAoAAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilsmHOnDnWsWNHa9asmY0ePfqk3YQdO3a4fcbFxZ20fQIAAAAAAJzJIjzP8073IM4WnTt3toEDB1qnTp1O91DOWdOnT7eJEyfahAkTTvdQAAAAAADAKUSlVDZs377dLr744lN3NwAAAAAAAM4ThFIh6N+/v5tel5ycbDfddFOK6XtLliyxli1bBtZdvXq11atXL/D+wIEDdu+991qrVq3syiuvtFtuucXtx3f99ddb06ZN3TaHDh1Kcdy9e/faAw88YK1bt7Zrr73Wxo0bF1jmH3fMmDFuuaYVfv/99yHfeB3rqaeesg4dOrj93HPPPRYfH39SjrtgwQLr16+ftWjRwlWXffPNN4FpirfddpvbTtfwjjvusK1bt7plq1atcp8988wztm7dOvdaP1999VXI5wQAAAAAAM4e0ad7AGeDsWPHun8VHL377rtWrVq1kLdVoHPs2DGbOXOmRUZG2g8//GARERGB5R9++KFt27bNhT+pKaCJiYmxzz//3FVp3XzzzVahQgUXYsmRI0csKirKBTdvvvmmvfzyyyFPe3v00UctOjrarZ83b16bO3duICzLyXF//fVXe/jhh+3ZZ5+1Bg0a2M6dO23Tpk1u2fHjx10I9vzzz7v9P/744269V1991apXr27z589n+h4AAAAAAOcJKqVOMQVQhw8fdhVBuXLlsiZNmqQIpTKSmJho8+bNsxtvvNFy585tFStWtPbt29usWbNSrNerVy8XEDVv3tw2b94c0ph2797tAqCHHnrIChcu7MbVtm1bd5ycHnfq1Kl21VVXWePGjd3yMmXKuNdSvnx5148rf/78Fhsb6wIqVZYBAAAAAIDzD5VSYZj6pwohBUD79++3q6++2k3nU9VUZrRuUlKSFS9ePPBZsWLFUgRAqnBStZOo8kjHCYWm0SmICt73yTquKqPq1KmT4TmNHDnSli5d6qqtdBx/PwAAAAAA4PxCpVQOqeJH4YpPVVHBVBWkEOqjjz5yU90+++wzW7RoUZb7VQWTKo1U1eTbs2ePFS1aNKdDtlKlSrn+Ubt27Trpxy1ZsmSgT1Rqo0aNsri4OPd0vTlz5tiTTz5pqR/+GEoVGQAAAAAAOPsRSuVQ2bJlXcCj5tySujH3woULXU8lhS+qKlLfJgVVWVEFkXo4qaG4elJt3LjRvvjiixRN1UPRpUsXmzRpUorPVAWlaYTPPfecq15KSEiwb7/91h0np8dVY3Otr/NWWKfKqcWLF7tlCqR07vny5XNB1/jx49NsX6JECdfHKnXTdwAAAAAAcG4hlMqhIkWKuCfq3X333TZgwAArUKBAiuWqGtIy9V669dZbXa+m2rVrBwIsPWFOT+AT9VjSe7+SasiQIS7w0hPu7rzzTuvTp0+2Q6ktW7a4JwCmNnz4cCtUqJDrDdWmTRv79NNPA1MKc3LcGjVquAqo1157zT1tUE3SFUbJoEGD3DRAPYlQ16JRo0ZptlczefWg6t69u7ses2fPztb5AgAAAACAs0OEl3r+FAAAAAAAAHCKUSkFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlMqGOXPmWMeOHa1Zs2Y2evTok3YTduzY4fYZFxdnZzqdt8Zav359mzBhwgnt43w736w8/fTT9sorr5ySfZ9ruFYAAAAAcO6I8DzPO92DOFt07tzZBg4caJ06dbLz3c0332wtW7a03r17p1k2ffp0mzhx4ikLcc60883KNddcY/fff7/b/myR2ZhP5f09G68VAAAAAODEUCmVDdu3b7eLL774BC81AAAAAAAAfIRSIejfv7+bwpWcnGw33XRTiul7S5YsSVHVsXr1aqtXr17g/YEDB+zee++1Vq1a2ZVXXmm33HKL24/v+uuvt6ZNm7ptDh06lOK4e/futQceeMBat25t1157rY0bNy6wzD/umDFj3HJNK/z+++9Drka57777rG3btm57vb/xxhvt2LFjgaqg4CoYVa68+eabWe531apV7to888wztm7dOvdaP1999VVI55sV/5y/+OILV62mfb/66qtu2cGDB+3xxx+3q666yjp06GBvvfWWBRcBzpgxw1W6NW/e3P375ZdfhnSds5LZtXrooYfcGDVdcejQoe51r169AuvOnDnTfdagQQMbOXJkmn1rzN26dXPnfNttt9mWLVtSHHfEiBF2ww03uHPSsRITE0Mas86va9eu7j7oezN+/PjAsszGHMr9nTt3rqsm05gHDBhgmzZtCmnMZ+q1AgAAAACcOtGncN/njLFjx7p/FaS8++67Vq1atZC3VQCgsEd/VEdGRtoPP/xgERERgeUffvihbdu2zYUhqemP/5iYGPv8889dlZb+uK5QoYILE+TIkSMWFRXlQgEFIS+//HLIU6r0B3yjRo3c+XzyySd266232vLly13vpBNVvXp1mz9/fqbTuzI731DoWi5YsMAFKblz53bhiAwbNsxy5cplU6ZMsaNHj9odd9xhF110kbVv395to8DqX//6lztnBR/79u0L+TqfqOeeey7TKWlXX321+3nsscfSbLtmzRrXP+m1116zSy65xIVvDz/8cOC7KIsWLQoEYApwdF1CmfaWL18+F9JUrFjRVq5c6aak1qpVy2rWrJnpmLO6v9qXAiVd5yuuuMJ9rwYPHmwffPCB+55mNuYz9VoBAAAAAE4dKqVOMQVQhw8ftq1bt7rQpEmTJilCqYyokmPevHmugknhiwIEBSyzZs1KsZ7+wNYf/KoA2bx5c8jjUuhSrlw5K1OmjNt/2bJlbc+ePXam03W5++67rVChQhYbG2uXXXaZ7d6924Ulqv5S4FK8eHEXbnz99dduG1Wm6Rqpekb3olSpUi68yM51DrfZs2dbw4YNXVikwEzBkUIfhWa+du3aWYkSJdzPpZdemqIqKTOqkqpcubILSWvUqOGmpKrCL6emTp3qKtUU3up69+jRw90bPzjMyZhP17UCAAAAAJw6VEqFYerf8ePH3ZSh/fv3u2oPTedTIJAZrZuUlOQCFl+xYsVSBE958+a16Og/b6H+GNdxQqXjKzjwK1j0r453psuTJ48LFoKp8kl69uwZ+ExBVJUqVQLXSZVBkyZNctMuFcY9+OCDLpAJ5TqfDgoIg8dUoEABF8Lp89KlS7vPChYsGFienfuvqYuqItJ107mrsiwhISHHY9b+li5dmiLQ03537doVqC480TGfrmsFAAAAADh1CKVySH/8Boc5qsQJlj9/fhdC6Wf9+vWuz4764mQ1Naxw4cIuKFKlSdGiRd1n+iPbf30q+D2Ysjon/w/74N5YwUKpBDtRfogWTJVP+ly9plTtlJ7GjRu7H4UkL774oj3//PMumAn1Omd0vqFcq6wCyPQoGFu7dm3gvfpvKUjJ6f1XcPToo4+6KW6qaNLY+vTpk6L/VlZjzuj+lixZ0vWTuv322094fGfStQIAAAAAnFpM38shTXuLj48PTFEKbvosCxcudFOF9Ee/H2woqMqKKqAUXKkRuXoibdy40YUu2e2D06VLF1chlB3ly5d3/aVE/Z9WrFiRZp1KlSq5qpj0ghpVMmnqVHYbmZ8oVclo+tZLL71kcXFxLiRSnyE1RhdVQ3377beuIkiBioIP/x6Eep0zOt9QrpXGp/FkR4sWLey7775z+1aQ9vbbb7tqI7/y50SpD5m+i35go6qp4Ol1oYw5o/urBvLTpk2zZcuWueuk5eqllp0qrDPpWgEAAAAATi1CqRwqUqSIe6Ke+hypCkpTh4Kpl5SWqeeTmomrd1Ht2rUDAZaeKKYn0omeGqf3asosQ4YMcYGXnpB25513uoqW7IZS6qOkJwBmh46jihqNSxU16tWT3rREPfFO59W3b98Uy1SBo6qk7t27u3NSz59Qzjcnhg8f7vpDXXfdde4ph0888USgakkhjJpta+pkmzZtXAij6Xu+UK5zRucbyrVS43SFM+pVpe9IcD8wnb+WTZ482b1WPyS/qbjGpQbuekqiej6pmXdOq9AUrukJkoMGDXLX4qeffnINzkMdc2b3V9Mh1WD8hRdecE+bVE8pNRTPzpjPpGsFAAAAADi1IrzU83YAAAAAAACAU4xKKQAAAAAAAIQdjc5xRtB0rIxoupymmyFr7777rr333nsZLteT8XLlysWlBAAAAACcdkzfAwAAAAAAQNgxfQ8AAAAAAABhRygFAAAAAACAsCOUAgAAAAAAQNgRSgEAAAAAACDsCKUAAAAAAAAQdoRSAAAAAAAACDtCKQAAAAAAAIQdoRQAAAAAAADCjlAKAAAAAAAAYUcoBQAAAAAAgLAjlAIAAAAAAEDYEUoBAAAAAAAg7AilAAAAAAAAEHaEUgAAAAAAAAi76FO58x5LP7fv9u+wsrnz2/dNegY+X7Rvu1334wz3+sVLmlnjIqWt4aIP091Hwehc9n8t+qX4LHh7+aRuR6tfuFSKdRos/MC2HItzr6MiIixfVIyVic1nDYuUsr9fdJlVylvopJ4rAAAAAAAAzpBQ6kSUis1rpWPzBd7nj45Js86H29emeZ86lApsHxVjVfIVsi1H42zV4X3u54Nta+2Nmq2sTfFyp+AMAAAAAAAAcNaFUr3LVLP7K9XJcPnhxAT7/I+N7vXlBYrbskO77bOdG+2Jqo0sT1Ta06lZoJh9VLejez13zxa79ZfZdiAx3u74ZY4tanK9FY3JfQrPBgAAAAAAAOdET6nP/thoR5ISLVdEpI2u0cpiIiLtUFKCzfhjU5bbtihW1u6reIV7rW2m7dgQhhEDAAAAAADgtFRKqbfThbPeCWndFzf+5H5815W+2P51afPA+8n/nbqnqXcV8xa0VsXK2le7f3NT+LqXrpLl/hsETfNbc3h/Ns8EAAAAAAAAZ00opaqmywoUC7yPS0ywtUf2h9RTqnyeAoHXvx09ZIv373Cvu5eu/Oe/paq4UErNz7cq/MqdP9OxJJuX4/MBAAAAAADAWRBKXRCb1z77y7UZPj0v1J5SqpLyI6V7fp1nERHzLdnzAmGTlt/z3+l5Gfl+/87A66r5Cp/A2QAAAAAAAOC86SnleZ59tH1d4L16Qh1MjLe4pITAZ5ODlqdn3p6t9q//Tg0sEBVj15asdApHDAAAAAAAgLPm6XsTtq222Xu2pPjsgyuudk/Z++3YIff+35e3dT2lfF/u2mw3Lf/GNh09aP/Zv8PqB/WNWnFoj3X84VPbeuyw7Yo/6j7LExlto2q0tKK5ePIeAAAAAADA6XDGhVI7jh9xP8ESvWTXyFzyR8VYs6IXpljeouiFljcq2j2VT+sFh1KqpFp+cI/li4p20/UaFS5tA8vVcE3SAQAAAAAAcHpEeJoXBwAAAAAAAITRWdNTCgAAAAAAAP+/vTsPs+lM9/9/11yUeehQEbOYRbrFTJlnYgiOMdrPGBFBEgffJmhEQospopOTRoxxIsTXPM8n3yY6dGtEDN2CbpShCjXv33XfOXtn71LUrmkX8X5dV121az17rfWsteSPfK77udcvB6EUAAAAAAAAfI5QCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFIAAAAAAADwOUIpAAAAAAAA+ByhFAAAAAAAAHyOUAoAAAAAAAA+RygFAAAAAAAAnyOUAgAAAAAAgM8RSgEAAAAAAMDnCKW8MGjQIFmxYkWWP4w9e/ZI+/bts/w8CxYskAYNGkjNmjXTfV1Xr161Y0RHR3u9z5EjR6RRo0aSlaZNmyZz587N0nMAAAAAAICMI5R6Cg0bNkz2798v1atXf+h3NmzYID179nzoeJEiRewYuXLlksfJuHHj5I033sjuaQAAAAAAgFQQSgEAAAAAAMDnCKW8dPnyZXn11VelYcOGMmbMGElISHCN7d2716qKdGnagAED5MKFC66xZcuWSadOnaR+/frStm1bWb58uWvM4XDIwoULpUWLFtKuXTs5fvy41w9Ol/mNGjVKmjdvLosXL7a/+/XrJzExMSkuORw9erQsWrQo1eOeOnXKluVNnz5dzp49a5/1Z9u2ba7vdOvWza6nRo0aEhUVJWml823atKndj2+++ca1PTIyUt5++20b69Chg927lJb/bdmyxe6XzmvevHk2tnnzZvu7Vq1aMmvWLI/9rl275roO/alXr57HMslHndd5zofNGQAAAAAApA+hlJcOHTokM2fOlC+//FKOHj0qBw4csO0nT56U8ePHW0C0c+dOadWqlbz11luSmJho42FhYbbfvn37ZMaMGTJ//nw5ceKEje3atcuWyS1ZskQ+//xzOXjwYJoeXufOnS18Wr16taxZs0b8/PzSFGylpEKFCrYsb+zYsVK2bFn7rD8anDl98cUX9pMe9+7dk4CAAAu5NOCZM2eOa0yDsKCgINm4caNt1/vivM9OGrrpNg33duzYIU2aNLHtrVu3tnnq7+QKFy7sug7t2/XCCy9YmOfteR81ZwAAAAAAkD6EUl7SUEbDDf2pVKmSqxpq3bp10rJlS6sa0uDilVdekevXr1uVkdIqqTJlyoi/v79UqVJFypUrJ6dPn3ZVWOlxixYtKvnz57fvpkXJkiWlePHiEh4eLqGhoVKsWDG5ceOGPO569Ohh90qrzi5evGjbtPJMgzut9tJrKVWqlAV8GvS50++NGDFC8ubNKyEhIVK5cuU0nVtDQWdfrbScN6U5AwAAAACA9AvMwL5PlTx58rg+a1VNbGys6y10WjnlHmLEx8fbkrHy5cvL1q1bZenSpfY9rZ66f/++jaubN29aNZJTwYIF0zQnDbo0KNEfpb+dFVqPq5w5c0pgYOAD9/HWrVs290KFCnncj+QBUI4cOSwYTA+trNIfrUpz3jNvzvuwOQMAAAAAgPQjlMqgZ555xvpJOStv3GkQNWHCBOt7pJVUGiL16tXLekmpAgUKWD8jp8yocnIeW6uI3AOqu3fvPvBdDViSkpJSPI4uBfSlfPnyWVCkVWZ6X5z3w/nZyRkmpdUPP/wg7733nlVK6bnSel4AAAAAAJC5WL6XQS+//LJ8/fXX8t1331nAo42/tem2VkNpLyINiZwBh1ZNOZf1qcaNG1ufoitXrljV1FdffZXiOTp27CirVq1K07xKlCjh6i+lTdqdfazclS5d2qq8UgqmtBpJ55WeRubpoZVI2jxdG4pr36jz589bQ3NtMp5R0dHR1sj8zTfftJ5ZvjovAAAAAAB4OEKpDNI+UePGjZMPPvjAQibtKaVNsrXSSEOf/v37y+DBg6VZs2Zy7NgxqVq1qmvfiIgIe9ubvtWvb9++Fo6k5NKlS3L79u00zUsrsrRSS9+Up5Va1apVe+A7es47d+5Yn6TevXt7jGllV926daVLly7Spk0b2b17t23XEE3fYKfHVTqmf2sj+IzS5upxcXHWTHz48OF2Dd6GQ9rzSeehgaA2fdfPAwcOdL1R8B//+Ic1mne+gc85/4yeFwAAAAAApI+fw7neCwAAAAAAAPARKqUAAAAAAADgczQ6R6bRZXEPo8sDdRkjAAAAAACAYvkeAAAAAAAAfI7lewAAAAAAAPA5QikAAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD5HKOWFQYMGyYoVK7L8YezZs0fat28vvnL16lV7Y150dLT4UkxMjIwcOVIiIiLs/Pfv3/fp+QEAAAAAQPYjlHqKFSlSRPbv3y+5cuVK874anmmIlh47d+6US5cuyfbt2+38OXLkkCfZhg0bpGfPntk9DQAAAAAAniiEUvC5K1euSPHixSU4OJi7DwAAAADAU8rP4XA4snsST8Lyveeff15OnDgh58+flzp16sjUqVMlMDDQxvfu3SuLFi2Sy5cvS9myZeX//J//IyVLlrSxZcuWyZdffinXrl2TvHnzWkVNr169bExv/ccffyxfffWVBTQtWrSw6iGtvEmNfkePXalSJas8KlWqlLz77rv2W0VGRsr06dPl22+/lbCwMOnWrZv07t3btb/+rfPVpXS7d++W3Llze3W9Y8aMkUOHDtl+Omd/f38pVqyYrFy5MtU5a2XV7373O4mPj7drd4ZS27Zts2qp1OZ85MgReeutt+Q///M/Zf78+XL79m37zvDhw+XOnTsye/Zsm1tAQIB06tRJBgwYIH5+frZvVFSUzJ07Vw4ePCj37t2T6tWry/vvv29zeNQzUps2bbLne/PmTcmfP7+89tpr0rJlSzl16pQMHDhQEhMTJSEhQUJCQuz7eo36LAEAAAAAwMP9lKogVRp2aDChevToIQcOHJBGjRrJyZMnZfz48fLhhx/Kiy++aAGTBierV6+2cETDlZkzZ1pYpN/VEKNatWpStWpV2bVrl4VLS5YskdDQUBkyZEiansQPP/xgc9Hz/9d//ZeFYcuXL7cxDXeCgoJk48aNVpmkQZMGZfXr17fxL774wkKpDh06pOl6Z8yY4Vq+N3r0aNvmLf2uLtfT4545c0ZmzZrlMZ7anJWGYToXvU69Z2fPnrXtEydOtIBp7dq11qPq9ddfl+eee05atWpl4xMmTLBQTXuD5cyZ04LEpKQkG3vUM9LzTZo0yZ6vhnPah0vDKVWhQgW7Hn2GGsr5ou8YAAAAAAC/FCzf85JWvhQuXNh+tDrpwoULtn3dunVWNVOjRg0LoV555RW5fv26KyzRip0yZcpYRVGVKlWkXLlycvr0aRvTYESPW7RoUavA0e+mRb58+eTll1+2sEUrivS4GjRp1c6+ffukX79+Ftxo2KLhjFZUZfR6s4q3c9bvjRgxwiqatDKpcuXKdr81HBo1apQFTIUKFbLQTKvOlHNcq7z0nml41bx5cztPas9Igyt9rtoD6+7du9aHq2LFill6LwAAAAAAeBpQKeWlPHnyuD5rNU9sbKx91sqZo0ePeoQnujxNl4KVL19etm7dKkuXLrXv6TIvreLRcaUVN7rcz6lgwYJpengFChRwLU/TMEZDmhs3bljooufScMb92BcvXszw9WaVW7dueTVnXeanQZk7vbeqe/furm0aJjnvrY7rPXE/trtHPSOtqtIqqlWrVsmCBQusF9Y777xj4RUAAAAAAEg/QqkMeuaZZ6wH0bBhwx4Y05BDl43NmzfPKqm0Ekd7FTnbeGmopH2UnDRQSgvdV4+lwZRW8WhwpMfUaiCt7tEKIf3beWzn58yg15KZvJ2zfic5rV7S7Vu2bHFVPyUfj4uLs6AwpUDrUc9I1a1b1340qPrDH/5gvag0xHJyBoMAAAAAAMB7LN/LIF0+9/XXX8t3331n1TnaUHvz5s0WYGhDbQ03nMGKVuQ4l/Wpxo0bW5Nv7Z+kVVPajyolHTt2tEqdlKqL1q9fb0vatFm3LjsLDw+35Xzah2nx4sXWE0mblWtgk5b+T6nRqiPtC5VZMjJnnUvt2rWt0Xl0dLRVO+nctDG6c7xevXrWD0vvmT4b7eel50ntGen39btaPaXhk4ZWuXLl8ji/Bl36DPXZAwAAAAAA7xBKZZAu4xo3bpx88MEHFjJpTyltxK0BRunSpaV///4yePBgadasmRw7dsyaZztFRERYo/FXX31V+vbt69HQ2532M9I3zSWnfZA0DGvSpIm9Ve73v/+9q2pn7NixVh3Utm1bezudVv84Ax4Nwho0aGBvrlNt2rSxv7W5ube0CbmGb9r3Sd9ylxkeNefUTJ482cK5rl272v2YMmWKVY+5j2sfKm3ars9CwzwNmFJ7RhpYadP61q1b27gGVrp8z51WWGklVZcuXexe6tsMAQAAAADAo/k53Ncp4YnBG98AAAAAAMCTjEopAAAAAAAA+ByNzpFpdAngw/Tu3duWyAEAAAAAACiW7wEAAAAAAMDnWL4HAAAAAAAAnyOUAgAAAAAAgM8RSgEAAAAAAMDnCKUAAAAAAADgc4RSAAAAAAAA8DlCKS8MGjRIVqxYkeUPY8+ePdK+fXv5JZg2bZrMnTvXY9uCBQukQYMGUrNmzSy/nz179pQNGzZ4bLt69aqdPzo6WnzpYefVZ63P3NfnBQAAAADgcUAohXRJLVAZN26cvPHGGx7bhg0bJvv375fq1atny10vUqSInT9Xrlxp3jcjAVJGzvsoGrpp+Obr8wIAAAAAkBkIpQAAAAAAAOBzhFJeunz5srz66qvSsGFDGTNmjCQkJLjG9u7daxUrjRo1kgEDBsiFCxdcY8uWLZNOnTpJ/fr1pW3btrJ8+XLXmMPhkIULF0qLFi2kXbt2cvz48TQ9vDVr1tgxW7ZsKX/84x+lRo0aD11yOHr0aFm0aJFrWddrr70mTZs2teVdr7/+uvz4448e+86cOTPF69XPuo8eY/z48fa5R48ern03b95s22rVqiWzZs1K0/XcuXNHJk2aZNfTpk0b+eSTT+weeePcuXPSr18/m68eIzEx0WO8W7du9gz0HkVFRXmM3b59W0aOHCmNGzeWJk2ayJAhQyQpKcmr63XeL723o0aNsvPrM9Hvp3Zepc9cq7CaN28uH330ket6jxw5Yv+enE6fPu16vqdOnbJ5TJ8+Xc6ePWuf9Wfbtm1eXW9kZKS8/fbb9vw7dOhg/0adnOddvHixjeu1fPPNN149AwAAAAAA0oJQykuHDh2yoObLL7+Uo0ePyoEDB2z7yZMnLazQQGLnzp3SqlUreeutt1yhSFhYmO23b98+mTFjhsyfP19OnDhhY7t27bIlWEuWLJHPP/9cDh486PWD02Bi3rx58oc//EG++uor+ctf/uL1vrGxsRb6rF+/Xnbs2CF58uSR9957z6vr1WvQJWG6NGzq1Kn2eeXKla79Wrdubdv0d1pNnDhR7t27J2vXrpWlS5fa/dy6dWuq+2mQo8sF69SpY/e0SpUq8sMPP3h854svvrCflGgoExMTY4Hali1bpE+fPuLn5+fV9TrpnLt27Sq7d++2Xlo5cuRI9bzq8OHDdq36b+D//t//a/unpkKFCjaPsWPHStmyZe2z/mi46c31apgVFBQkGzdulDlz5ti5nc9X6TMICAiwkEtDKf0OAAAAAACZjVDKS/o//IULF7afSpUquaqh1q1bZ5U9WpGi/yP/yiuvyPXr162CRWmVVJkyZcTf39/CknLlylnVi7PCSo9btGhRyZ8/v33XW7pv7dq1pXz58pIzZ07p3r271/uWKFHCKrO011BISIgFVM45pXa9WUXvmQYrGu5pkFeoUCGrINq+fXuq+2qVl97v3r17S2BgoHTu3Fny5s3r9bk1gLp7964dJzg4WOrVq+cKpbylFVIaium/AX3e3p6/Y8eO9uzDw8OtWkqfa1bSijcNSLWqLDQ0VEqVKmVBqgaA7rQaTK9Fr+vixYtZOicAAAAAwNMpMLsn8KTQaiInrTLRaiOly7S0ksj9f+rj4+Pl2rVrFhhppY9Wwuj3tHrq/v37Nq5u3rxplS5OBQsW9Ho+N27c8Ph+Wva9deuWLa3TeWtVjM5LwxxvrjerOJe7uYdruoTO/f48jN5HDVg0zFIaKBUoUMDrc/ft29euT5fq6b3RKi9dzqdBoreKFy8u6eH+3HTOugwxK+n16fPW0M99Du7Bk4aczn8Pvnj2AAAAAICnE6FUBj3zzDPWT0rfLJdS0DJhwgRbZqeVVBpy9OrVy9U3SEMI7e/jHjR5S4ME9yVqyffVCij3vkpaCeSkSwijo6NtGZpW9Gh1ji6dS4u0BDbJadDh7NnkpMvjtDJHl89pwJQWeh91+Z1eowZTen/d72tqtGJMQyj90XuqfcG0J5b2ZPL2enXu6a0Qc9I5O0OqRz0/p7RWc6l8+fLZXPW8zuBO/+2kJcQDAAAAACAzsHwvg15++WX5+uuv5bvvvrOgRZtKa28irYbSKiQNSJz/w69VU85lfUoba2vfnitXrli1j/aGetgSr1WrVnls02VV2o9Il93peVavXv3AEj1n43Rt0u7sY6U0kNIgRgMcDSTcm697Syttzpw5I+lRunRpq9JyD6b0eLoccfbs2TY/DWT0+Np4OzXPPvusPP/889YbSpenaX8nbV7uLe3lpcsT9Vk5AzO9P5l1vY+ifb20ekn/DehSRWdz82LFiklcXJzr34t7E3MnXVqp+6XUQP1htAJKwzZtZK5B3vnz5y0IdG+qDgAAAACALxBKZZD2idIm2x988IGFTNpTSptGaxWLhi/9+/eXwYMHS7NmzeTYsWNStWpV174RERH29jN9y50uIXOvzHF36dKlB0KWihUr2lvztLpHe1G9+OKLHuNakaWVWvoWNq3UqlatmmtM56PLtXS+Q4cOtV5IaaVvnNPwTfsRaWWRey8ifROcjunbAfXzwIEDPfbVa9U37Wmwpn2gnCZPnmyhkjYM17fgTZkyJcUKoZRoE3IN6XQ/bT6vfZ2cNNDReei9UNpDS//WZu5Ke0mNGDHC5qP3Q/stVa9e3avrfZTUzqs0iNPG6npP9N+C/ptQ2mdK3wKo89Lz5c6d+4Hja/Vd3bp1pUuXLnZsZ5P01M6rDdI18NIm5sOHD7d/K4RSAAAAAABf83M415LhiaYVUxoueFNZBAAAAAAAkN2olAIAAAAAAIDP0egcj73PPvtM/vSnPz10XN98GBwc7NM5AQAAAACAjGH5HgAAAAAAAHyO5XsAAAAAAADwOUIpAAAAAAAA+ByhFAAAAAAAAHyOUAoAAAAAAAA+RygFAAAAAAAAnyOUAgAAAAAAgM8RSnlh0KBBsmLFiix/GHv27JH27dvLL8G0adNk7ty5HtsWLFggDRo0kJo1a2bZ/dR72LZtWzuPng8AAAAAADyeCKWQLhqeaQD0MOPGjZM33njDY9uwYcNk//79Ur169Sy767Nnz5ahQ4faefR8v/T7DAAAAADAk4pQCr8oV65ckXLlymX3NAAAAAAAQCoIpbx0+fJlefXVV6Vhw4YyZswYSUhIcI3t3btXevbsKY0aNZIBAwbIhQsXXGPLli2TTp06Sf369W1Z2fLly11jDodDFi5cKC1atJB27drJ8ePHJS3WrFljx2zZsqX88Y9/lBo1ajx0yeHo0aNl0aJF9vnq1avy2muvSdOmTW2Z2+uvvy4//vijx74zZ85M8Xr1s+6jxxg/frx97tGjh2vfzZs327ZatWrJrFmz0nQ9d+7ckUmTJtn1tGnTRj755BO7R97o27evnTcpKUn69+//wPK9TZs2SefOne0Z6bVfunTJY3+9Zr0/o0aNsmvW+6rXmNrzVQcOHJA+ffpIRESEvPzyy7Jjxw6v7vPt27dl5MiR0rhxY2nSpIkMGTLE5u/NfQYAAAAA4InnQKoGDhzo6NSpk+Pf//63/TRt2tSxe/duG/vb3/7mqFevnuPPf/6zIyEhwbFmzRpHly5d7LNau3at4+zZs47ExETHiRMnHLVr13YcP37cxnbs2OFo3bq14/Lly47IyEhHt27dHO3atfPqifz97393NGjQwHHq1CnH3bt3HUOHDnX85je/8Zjz8uXLXX+PGjXK8fHHH9vnCxcuODZs2OCIiopyxMTEOMaOHet4/fXXvbpeJ51n8m3uJk6c6Jg5c+ZD76f73JzefPNNxzvvvOOIjo52XLt2zdG9e3fH5s2bHWmh90DvibvTp0/bM/ruu+8ccXFxjlmzZjn69OnzwJxatGjhOHTokD07fWa3bt1K9fn+9a9/tedw8OBB2/bjjz/aZ2/u8/z58x1Dhgxx3L9/3xEbG+s4cOCAIykpKU33GQAAAACAJxWVUl7SaqbChQvbT6VKlVzVMuvWrbPKHq1SCggIkFdeeUWuX78uZ8+etXGtkipTpoz4+/tLlSpVbGnZ6dOnXRU4etyiRYtK/vz57bve0n1r164t5cuXl5w5c0r37t293rdEiRJWmZUrVy4JCQmxqiTnnFK73qyi90z7QGmlUlhYmBQqVMj6KW3fvj3Dx969e7fdq2rVqklQUJAMHDhQTp48aUv93GmFVJ06dew56jPLmzdvqs/XOV63bl0bDw8Pt8/e3Gc/Pz+5e/euVU8FBwdLvXr1bBsAAAAAAE+DwOyewJMiT548rs8abMTGxtpnXV519OhR2blzp2s8Pj5erl27ZoHR1q1bZenSpfa9xMREuX//vo2rmzdvStmyZV37FSxY0Ov53Lhxw+P7adn31q1btrRO533v3j2bV2BgoFfXm1WcS+XcwzVdyuZ+f9JL75WGXE65c+e2kEi3ayDoVLx48RTn9ajn+69//Ut+/etfp+s+65JDva+6VE+/27p1a1vOpwEmAAAAAAC/dIRSGfTMM89Yv6GU3vSmgcaECRNk3rx5VmmjYUOvXr1cfZIKFCggkZGRru9rSOItDaF++OGHh+6roYuGIE5akeM0f/58iY6OlpUrV1o1kFZdTZw4MQ1XLRkKTjTkcvZOcipSpIhVGm3ZskVCQ0MlM+m9+v77711/R0VFWRik99+dnj8tz9c57t4nyl1q91krqDSE0h99ltqvSntxaf8xJwIqAAAAAMAvFSUZGaSNrb/++mv57rvvLGjRwEObfWs1jVbHaADlDD+0asq57Etpg+tt27bZMjKtmvrqq69SPEfHjh1l1apVDyw1O3z4sC0H0/OsXr3aY1yXjjkbp2uT9hMnTrjGNCjRQESXyWmY5d583VtaeXTmzBlJj9KlS1v1kHswpcfTJXazZ8+2+Wmgpsc/cuSIZJQ2INd7pfdDn8unn35qVU7uVVLpeb7OcQ3SDh48aHPWyqn/+Z//8eo+6z66LFL/jTiDOv1+Zt1nAAAAAAAeZ4RSGaR9osaNGycffPCBhUzac0jfxqa9gTR80TfBDR48WJo1aybHjh2TqlWreoQlHTp0sLfc6VIu9woZd/qmOH1Tm7uKFSva29y0ykZ7Ub344ose41qRpZVa3bp1s0ot7afkpPO5ePGizXfo0KHWRymt9G11Gs60atXKKnyc9A1x+qY4HdO3A+pn7eHkTq9V37SnwVrv3r1d2ydPnmxv+evatau9jW7KlCkeFV7pVaFCBRk7dqxVKTVv3tyCvGnTpnnVv+lRz9c5/vvf/14++ugjm7PeFw2jvLnPWmE1YsQIuw863q9fP6levbpX9xkAAAAAgCedn3Y7z+5JIOM0aNEgKjMqiwAAAAAAALIalVIAAAAAAADwORqd47H32WefyZ/+9KeHjuub8YKDg306JwAAAAAAkDEs3wMAAAAAAIDPsXwPAAAAAAAAPkcoBQAAAAAAAJ8jlAIAAAAAAIDPEUoBAAAAAADA5wilAAAAAAAA4HOEUgAAAAAAAPA5QikAAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFIAAAAAAADwucCsPPi9jR0k6erBnzf4+Ytfjl+Jf8EXJPiFNyXgmZq22ZGUKPGn/iQJZ7+QpFunRRLjxC8sXAKLNZGgam+Kf65nPY7rSLgv8X9dKAnnv5akO+dEHInil+s5CSzeSoJfGCF+IfkfmIOOh3X/S1ZeLgAAAAAAAB6HUMrFP1j8C1a1sCnp5t8l8Z9b5f6lnZKj/SYLqGK29ZDEH3f99N3QguKX4xlx3Dkn8X//TOLPfSU5Wq+TgIJVbNgRd0fua9AUecL+9stZxI7vuP29xJ/4XhLOr5Mc7TaKf5hnkAUAAAAAAICnbPmeX85nJGeHbZKz0x4Jbfb5TxsdCZLww5cS/7c/ugKpoIr9JaznKQnr+v8kR5v1FjZJ7E2J3feaOBwO+07c0WmuQCq4xgQJ6/E3Cet+TEIaffLTYaP/KbGH3vHFZQEAAAAAAOBJ7SmlS/ZMUJgEvzRR/Px+mlJA0XoSWKazfU6K/Jv9aDAV/8N/2za/3CUlqNpw13GCynQW/yL17HPiP7eJI/aW7y8GAAAAAAAAj08o5bj3L7n3dQu591UjidnR56eNfoESWLqzJN0++9NEcpcWv6BcnpMrUNX1Wb/niLlulVM/jVVyBVhOziV+4kiSpDs/ZPFVAQAAAAAA4PHuKZUUJ0nXjv5vo/PC/9vofKQE/KrGz9/x83twv5S2/Tz4OBZ+AQAAAAAA4HEJpR715jv/vGWtR5RWNjnioz2qpZJunPD4nl9oQRF9s17sTddyPj+34MrZa0rDL/88ZbLykgAAAAAAAJAB2V5aFFi2208f4u9K3J8nicORZH8mXjkoCefW2mf/ApXtR5frBZV5xbY5oi5I/PG5ruPE/7BWEq8csM8Bz7UQv5B8nidyOMSREOP5k5Too6sEAAAAAACAOz+H87V2WeDexg6SdPXgIyulHEnxErOthyT+uPunDaEFLVBy3D6noyLB+exNfM5+UY64O3Jfj/u/VVF+OYvYW/oc0f/46e+wYpKj/SbxD3vWYw4pCa41VYKrDMmKSwcAAAAAAEC295R6BD//IAltsVriT31mb+JLunVaHFH/EL/cxSXg2SYS/MKb4p+r2M/fD85joVP8XxdKwvn1kmThVZL45S0rgcVbS3C1N8QvtEC2XhMAAAAAAACysVIKAAAAAAAAeCx7SgEAAAAAAODpQygFAAAAAAAAnyOUAgAAAAAAgM8RSgEAAAAAAMDnCKUAAAAAAADgc4RSAAAAAAAA8DlCKQAAAAAAAPgcoRQAAAAAAAB8jlAKAAAAAAAAPkcoBQAAAAAAAJ8jlAIAAAAAAIDPEUoBAAAAAADA5wilAAAAAAAA4HOEUtlowYIF0qBBA6lZs6asWLEiQ8fas2ePtG/fXrLT/fv37Xrq1asnjRo1yta5AAAAAACAxxuhVDYaNmyY7N+/X6pXry6/BDly5LDrmTNnTpadQ4M3DeAAAAAAAMCTjVAKAAAAAAAAPhfo+1M+ebQ6p3nz5rJ9+3aJiYmRzp07y5AhQ8TPz8/GN23aJJ9++qlERkZKpUqVZNy4cVKsWDEb023Tp0+Xb7/9VsLCwqRbt27Su3fvDM/J4XDIxx9/LF999ZUEBwdLixYtPMbv3Lkjs2fPlkOHDklAQIB06tRJBgwYYHPW6ylXrpycOHFCevXqJV9++aUULFjQjhcaGpplc7569apMnjxZTp8+LXFxcfLCCy/I2LFj5dlnn7Xx27dvy7vvvit/+ctfbJ7PP/+8fPTRR+Lv7y9jxoyxa9H7P378eNum93jlypUZnhcAAAAAAPA9KqW8dPjwYVm6dKksWbJENmzYIHv37rXtZ86ckWnTplmYoqFV2bJlLZRy0nAnKChINm7caMvadP8DBw5k+MHt2rXL5qHH+/zzz+XgwYMe4xMnTpR79+7J2rVrbd47d+6UrVu3usY1WBs0aJCsXr1a1qxZYyHQ8ePHs3TOsbGx0qZNG1m/fr3s2LFD8uTJI++9955rfNmyZRY6bd68WbZs2SJ9+vRxBX8zZsywpYFFihSRqVOn2mcCKQAAAAAAnlyEUl7q2LGj5M+fX8LDw61qytnXaPfu3VK7dm2pVq2aBTkDBw6UkydPypUrVyQhIUH27dsn/fr1swqkUqVKSatWrSwgyigNxbQ6qmjRojYvrYRyun79uoU2o0aNskqnQoUKWXWUhmZOJUuWlOLFi9v16Ny06ujGjRtZOucSJUpIu3btJFeuXBISEmIBlVZNOWkAdffuXfnxxx+t+ksbpjtDKQAAAAAA8MvC8j0v6fI2pwIFCsi5c+fsswY5Gvo45c6d2wIX3a4hVWJiose4HufixYsZfnA3b960qqyU5qfL5FT37t1d25KSkjy+r8vfdFmf/ij9rXO9detWls1Zjz1r1iw5evSoVXHpeQIDf/4n2LdvX6um0qV6+t3WrVvLyJEjba4AAAAAAOCXhVDKS1p95KQ9lzSYcgY233//vWssKirKghUdz5cvn4U9uq/z+xpWOT87aXiloVFa6DF0Hk56XCdd4qbn1SVwWu2Ulj5V3s75UR52PfPnz5fo6Ghbdpc3b16r9tJlhk5aQaUhlP788MMP1gOrVq1aUr9+fdd3CKgAAAAAAPhloATFS9oHSat3dFmeLoOLiIiw7fpb+01pP6b4+HhreF6+fHlbVqdVQBqoLF682HolnT9/3oKiRo0aeRy7dOnSVj30sGBKlw6uWrXKY1vjxo1l27ZtNh+tmtKG505a5aRLCrXRuYZAWpGkva+OHDmS6nV6O+dH0WWBGszpMkZ3OhcNnnRJoQZdy5cv9xjXvlgXLlywcMwZbOn33em16bUAAAAAAIAnG6GUlzTk0cbbusRM+yJpKKQqVKhgb5DTih/tNaU9krTxubMXko7pm+batm0rw4cPt7fdJQ949Jj6tryGDRum+Ja7S5cu2Zvp3GkY1qFDB3n11Vdtf/dqIqVvudP+UF27dpUmTZrIlClTrF+TN7yZ86NojyutdnrjjTekQYMGrn5UgwcPtmWAeu+GDh0qderU8dhPe0mNGDHC7oOOa1+r6tWre3xHm7NrI3Ttc6WVVAAAAAAA4Mnk59CyFDySNgkfPXp0moIZAAAAAAAAPByVUgAAAAAAAPA5Gp3Da5999pn86U9/eui4LtMLDg7mjgIAAAAAgFSxfA8AAAAAAAA+x/I9AAAAAAAA+ByhFAAAAAAAAHyOUAoAAAAAAAA+RygFAAAAAAAAnyOUAgAAAAAAgM8RSgEAAAAAAMDnCKWy0YIFC6RBgwZSs2ZNWbFiRbqOERMTIyNHjpSIiAg71v379z3Gp02bJnPnzn3s56w/+rlevXrSqFGjTJsvAAAAAAB4PBFKZaNhw4bJ/v37pXr16uk+xs6dO+XSpUuyfft2O1aOHDk8xseNGydvvPGGPO5z1h/9PGfOnHQfd8OGDdKzZ8+Hjrdv31727NmT7uMDAAAAAIDMQyj1hLty5YoUL15cgoOD5UnxJM4ZAAAAAABkLkIpL2iFjS6B09/NmzeXhQsXisPhcI1v2rRJOnfubMvOXnvtNasCcoqMjJS3335bmjZtKh06dJBly5ZlyoPTih9d7vbpp5/KgQMH7LP78r3Nmzfb37Vq1ZJZs2Z57HvkyBGb6+LFi21ebdu2lW+++cY1rnPs1KmT1K9f38aWL1/ukzk/ytWrV+3e6nx1n9dff11+/PFHGzt16pRtmz59upw9e9Z13G3bttn4mDFj7G89xvjx4+1zjx49XMe+c+eOTJo0SVq2bClt2rSRTz75xPV8U7tXAAAAAAAgfQilvHT48GFZunSpLFmyxJaJ7d2717afOXPG+ja9++67thytbNmytmTOSYOSoKAg2bhxoy1N0/01kMkoDUp0udtvf/tbC4/0s/vyvdatW9vf+jsl9+7dk4CAAAtuNGhxXzYXFhYmM2fOlH379smMGTNk/vz5cuLEiSyf86PExsZaYLR+/XrZsWOH5MmTR9577z0bq1Chgh1n7Nixdv+dx23RooWN6zXo30WKFJGpU6fa55UrV7qOPXHiRLsfa9eutWesywu3bt3q1b0CAAAAAADpQyjlpY4dO0r+/PklPDzcqqWcvYl2794ttWvXlmrVqln4NHDgQDl58qQtUUtISLBgp1+/fhIaGiqlSpWSVq1aWejxONBqIQ1bGjZsKBcvXnRt1yqpMmXKiL+/v1SpUkXKlSsnp0+fzta5lihRQtq1aye5cuWSkJAQC6gyY07Xr1+3kGrUqFEWxhUqVMgq4jRg9OZeAQAAAACA9AlM535PnYIFC7o+FyhQQM6dO2efb9y4YUGGU+7cuS000e0aUiUmJnqM63Eeh1AjZ86cEhj40+PXeWolkpNWCWnFkC530/nr8rr4+PhsnK3IrVu3bBni0aNHrXJJ5+Wcf0boNaru3bu7tiUlJVnFlTf3CgAAAAAApA+hVBoqatz7RGkw5QyZvv/+e9dYVFSUhRY6ni9fPquu0X2d39ewyvnZSYMODUIeBxrSTJgwQebNmyc1atSwaqlevXp59NDKyjk/7Li6hDA6OtqW3eXNm9eWT+qyO3d+fn6PPLZeS3K6pE+f0ZYtW6yaDQAAAAAA+AbL97ykvYy0WkeX5enSroiICNuuv7Xf1PHjx62aSJt4ly9fXooWLWrVNdo7SZtkx8TEyPnz5y380N5K7kqXLm0VQA8LeXTp4KpVq8QXtApJAyhncKZVU9o8PLnU5pxe+lY+DfV0CaQ7DaR06Z4usdNgL6Xm64ULF7bno8FgSrRiTXuAJd+myy9nz55t59AKLP2ONjgHAAAAAABZh1DKSxpc9OnTR/r27Wu9jRo3buxqsq0NtrVqR3tNaZ8jbXzurNrRsbi4OGuQPXz4cKs6Sh5K6TH1DXDar6h3794PnFvf5nf79u00PVjtgaRvmdO38K1Zs8Y+a7+r1GjY1L9/fxk8eLA0a9ZMjh07JlWrVn3ge6nNOb20b9fIkSPljTfesDk7+2/pfHTZo973oUOHSp06dR7YVyu76tatK126dLGeU9rvy92gQYPsfmhfrwEDBri2T5482fp/de3aVZo0aSJTpkyRu3fvZto1AQAAAACAB/k5kq/LwgO08fXo0aMfCJMAAAAAAACQPlRKAQAAAAAAwOdodI5Mo8vtHkaX+OkSPAAAAAAAAMXyPQAAAAAAAPgcy/cAAAAAAADgc4RSAAAAAAAA8DlCKQAAAAAAAPgcoRQAAAAAAAB8jlAKAAAAAAAAPkcolY0WLFggDRo0kJo1a8qKFSvSdYyYmBgZOXKkRERE2LHu37/vMT5t2jSZO3fuYzHnzLheAAAAAADwy0AolY2GDRsm+/fvl+rVq6f7GDt37pRLly7J9u3b7Vg5cuTwGB83bpy88cYb8jjMOTOu91E2bNggPXv2zJJjAwAAAACAzEUo9YS7cuWKFC9eXIKDg7N7KgAAAAAAAF4jlPJC+/btbQmc/m7evLksXLhQHA6Ha3zTpk3SuXNnadSokbz22mtWueQUGRkpb7/9tjRt2lQ6dOggy5Ytk8ywZ88eWwr36aefyoEDB+yz+/K9zZs329+1atWSWbNmeex75MgRm+vixYttXm3btpVvvvnGNa5z7NSpk9SvX9/Gli9fLr6Q2nn1Pr/88svSsGFD+71161bbfurUKbvW6dOny9mzZ133Ytu2bT6ZNwAAAAAASLvAdOzzVDp8+LAsXbrUQp8BAwZIxYoVLdg5c+aM9W366KOPbNu8efNsyZx+V2lQEhQUJBs3brSqpkGDBknJkiUteMkIPbcuhVu0aJHNIXnw1Lp1a/t59913U9z/3r17EhAQYMGNHmPOnDmuPk9hYWEyc+ZMKVWqlJw8eVIGDhwo1apVk6pVq0pWetR5tXfWpEmT5MMPP5Q6derI1atX5ebNm7ZfhQoV7F7o8r2VK1fSrwoAAAAAgCcAlVJe6tixo+TPn1/Cw8OtWkorldTu3buldu3aFp5o+KRBigYqGkAlJCTIvn37pF+/fhIaGmphS6tWrawP1OOgR48eFkxp5dHFixdd27VaqUyZMuLv7y9VqlSRcuXKyenTp7N8Po86b1JSks1Vq9Du3r0rRYoUsRAQAAAAAAA8maiU8lLBggVdnwsUKCDnzp2zzzdu3JBChQq5xnLnzi0hISG2XUOqxMREj3E9jnsAlF1y5swpgYE/PX6dZ2xsrGtMl8VppZdWI+n8tTosPj4+y+f0qPPqfLWKatWqVfYWP+2j9c4771h4BQAAAAAAnjxUSnnp+vXrHn2iNJhyhkzuY1FRURbw6Hi+fPmsusd9XMMq575OGgppJdDjQAOhCRMmyIgRI+yNfloRVrZsWY8eWhmdc0r7enPeunXrWm8vHa9cubK8//77Hsfw8/NL13wAAAAAAIDvEUp5af369XLr1i1blqehSEREhG3X39pv6vjx41bVo43Hy5cvL0WLFrVKJO0dpQ3FtSfS+fPnZcuWLdYPyl3p0qXl6NGjDw15dOmgVgj5gvaa0iDIGZxp9ZI2D08utTk/Skr7pnZevfe7du2y6ikNn3SJX65cuTyOW7hwYXs+GgwCAAAAAIDHG6GUl7RvVJ8+faRv377Srl07ady4savJ9tixY2XixInWa0p7IGnjc2fVjo7FxcXZ2+SGDx8uvXr1eiCU0mPeuXPHejv17t37gXNrH6Xbt2+nuV+UvoFO38K3Zs0a+6z9rrwJjPr37y+DBw+WZs2aybFjx1JscJ7anB8lpX1TO68GVqtXr7bm7TqugZUu33NXo0YNq6bq0qWLtGnTxvp9AQAAAACAx5OfI/m6LDygffv2Mnr06AfCJAAAAAAAAKQPlVIAAAAAAADwOd6+h0yjSwQfRpfp6dI8AAAAAAAAxfI9AAAAAAAA+BzL9wAAAAAAAOBzhFIAAAAAAADwOUIpAAAAAAAA+ByhFAAAAAAAAHyOUAoAAAAAAAA+RyiVjRYsWCANGjSQmjVryooVK9J1jJiYGBk5cqRERETYse7fv+8xPm3aNJk7d+5jNWcAAAAAAAA/h8Ph4DZkr0GDBkmjRo2kZ8+ead5348aNsnjxYlm+fLkEBwfLkzBnAAAAAAAAKqWecFeuXJHixYv7NJACAAAAAADIqMAMH+Ep0L59e2nevLls377dlst17txZhgwZIn5+fja+adMm+fTTTyUyMlIqVaok48aNk2LFitmYbps+fbp8++23EhYWJt26dZPevXtneE579uyR3/3udxIfHy9a7KZL6tS2bdskR44csnnzZlu6FxcXZ+ccPXq0a98jR47IW2+9Jf369ZPPP/9cQkNDZcKECVKrVi0bX7ZsmXz55Zdy7do1yZs3r1VD9erVK1PuY7ly5eTEiRN2PD1HwYIF5eOPP7Y53LlzR2bPni2HDh2SgIAA6dSpkwwYMMDu89WrV2Xy5Mly+vRpu6YXXnhBxo4dK88++6yrcuv555+3Y58/f17q1KkjU6dOlcBA/okDAAAAAPA4olLKS4cPH5alS5fKkiVLZMOGDbJ3717bfubMGQt/3n33XQutypYta6GUkwZSQUFBtsxuzpw5tv+BAwcy/OB06dz+/fvlt7/9rdSvX98+648GUqp169b2t/5Oyb179yz40RCrbdu2NjcnDc9mzpwp+/btkxkzZsj8+fMt7MkMGuhpgLR69WpZs2aNBU7Hjx+3sYkTJ9q81q5da/d6586dsnXrVhuLjY2VNm3ayPr162XHjh2SJ08eee+99zyOrWGWzlvDrqNHj2bKfQYAAAAAAFmDUMpLHTt2lPz580t4eLhVTWmlktq9e7fUrl1bqlWrZuHTwIED5eTJk7asLiEhwYIdrUjSSqBSpUpJq1atLGx5HPTo0cOCqYYNG8rFixdd27VCqUyZMuLv7y9VqlSx6iatUMoMJUuWtOWGeh/1nmhF2Y0bN+T69esWoo0aNcpCsUKFCllllQZ9qkSJEtKuXTvJlSuXhISEWECVfE4tWrSQwoUL249WrF24cCFT5gwAAAAAADIfa5u8pMvMnAoUKCDnzp2zzxqoaIDilDt3bgtNdLuGVImJiR7jehz3ACi75MyZ07W0TeeplUhOWp2klUq6ZE7nr2/002WCmUGDLg3C9Efpbz2Hnkt1797d9d2kpCSrPFO3bt2SWbNmWQWUVlPpPsmX5mn1lFPyawIAAAAAAI8XQikvaSWPk/aJ0mDKGTJ9//33rrGoqCgLQ3Q8X758Frrovs7va1jl/OweoGgA8zjQcEj7S82bN09q1KhhIZL2f0r+ksbMnLMeu0iRInavtmzZYhVUyekSwujoaFm5cqX1udLlk7rcDwAAAAAAPJlYvucl7WWk1Tq6LE+XlEVERNh2/a39prQvklYTacPz8uXLS9GiRa2SR/s9LV682BqkawNuDV20H5S70qVLWwXQw0IeXTq4atUq8QWtQtKQyBmcadXU2bNnH/heanNOK60m02WQ2uhcwyethNJ+XdqUXek2XbqnS/s02Fu+fHmmnBcAAAAAAGQPQikvaWDSp08f6du3r/U2aty4sW2vUKGCvQVOq3a015T2OdLG58438+mYvi1Om4kPHz7cqo6Sh1J6TH3znPZ2SunNfJcuXZLbt2+nuV+UvpFP38KnDcX1s/a7So2GTf3795fBgwdLs2bN5NixY1K1atUHvpfanNND366nfbi6du0qTZo0kSlTpsjdu3dtTOejyx71vg8dOtTergcAAAAAAJ5cfo7k67LwAG24PXr06AfCJAAAAAAAAKQPlVIAAAAAAADwORqdI9PoEsGH0SV+ugQPAAAAAABAsXwPAAAAAAAAPsfyPQAAAAAAAPgcoRQAAAAAAAB8jlAKAAAAAAAAPkcoBQAAAAAAAJ8jlAIAAAAAAIDPEUplowULFkiDBg2kZs2asmLFCp/tCwAAAAAAkN0IpbLRsGHDZP/+/VK9enWf7uuNDRs2SM+ePbPk2AAAAAAAAIRSAAAAAAAA8LlA35/yydO+fXtp3ry5bN++XWJiYqRz584yZMgQ8fPzs/FNmzbJp59+KpGRkVKpUiUZN26cFCtWzMZ02/Tp0+Xbb7+VsLAw6datm/Tu3TvL57xs2TL58ssv5dq1a5I3b16reurVq5drXOe8aNEiuXnzpuTPn19ee+01admypZw6dUoGDhwoiYmJkpCQYEsE1e9+9ztp0aJFls8bAAAAAAA8HQilvHT48GFZunSp3L9/XwYMGCAVK1aURo0ayZkzZ2TatGny0Ucf2bZ58+ZZKKXfVRpIBQUFycaNG+XKlSsyaNAgKVmypNSvXz8rn6sFYDNnzpRSpUrJyZMnLWiqVq2aVK1a1YK1SZMmyYcffih16tSRq1evWjilKlSoYMsCdfneypUr6VcFAAAAAACyBMv3vNSxY0erKAoPD7eqqT179tj23bt3S+3atS3w0fBJwx8NgTSA0kqjffv2Sb9+/SQ0NNQColatWsnOnTslq3Xq1EnKlCkj/v7+UqVKFSlXrpycPn3axpKSkiQgIEAuXbokd+/elSJFiligBgAAAAAA4CtUSnmpYMGCrs8FChSQc+fO2ecbN25IoUKFXGO5c+eWkJAQ264hlS6Dcx/X41y8eFGy2tatW61aS6ugdA5a4RUfH29jOXPmtCqqVatW2Vv8ihcvLu+8846FVwAAAAAAAL5ApZSXrl+/7vqsfaI0mHKGTO5jUVFREhsba+P58uWziiT3cQ2rnPs6aXil1UvpkdK+GkRNmDBBRowYYX2wtKqrbNmy4nA4XN+pW7euzJ0718YrV64s77//vscxnP2yAAAAAAAAsgKhlJfWr18vt27dsmV5GuRERETYdv2t/aaOHz9ulUja8Lx8+fJStGhRCQwMtN5Rixcvtj5O58+fly1btlgvKnelS5eWo0ePPjSY0qWDWtWUkpT2vXfvngVQzvBLq6bOnj3rGtfr2LVrl1VPafikS/xy5crlcdzChQvbtWrIBgAAAAAAkNkIpbykfaP69Okjffv2lXbt2knjxo1djcHHjh0rEydOtF5T2rdJG587K410LC4uTtq2bSvDhw+3N+AlD6X0mHfu3JGGDRum+GY+7f10+/btFOeV0r4aVPXv318GDx4szZo1k2PHjlmDcycNrFavXi2tW7e2cQ2sdPmeuxo1alg1VZcuXaRNmzbWOwsAAAAAACCz+Dnc13QhRe3bt5fRo0c/ECYBAAAAAAAgfaiUAgAAAAAAgM8RSgEAAAAAAMDnWL4HAAAAAAAAn6NSCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFLZaMGCBdKgQQOpWbOmrFixIjunAgAAAAAA4FOEUtlo2LBhsn//fqlevXp2TgMAAAAAAMDnCKUAAAAAAADgc4G+P+WTp3379tK8eXPZvn27xMTESOfOnWXIkCHi5+dn45s2bZJPP/1UIiMjpVKlSjJu3DgpVqyYjem26dOny7fffithYWHSrVs36d27d6bMqVy5cnLixAnp1auXfPnll1KwYEH5+OOPJTQ0VO7cuSOzZ8+WQ4cOSUBAgHTq1EkGDBhgc7569apMnjxZTp8+LXFxcfLCCy/I2LFj5dlnn7VjDxo0SJ5//nk79vnz56VOnToydepUCQzknwsAAAAAAMgcVEp56fDhw7J06VJZsmSJbNiwQfbu3Wvbz5w5I9OmTZN3333XQquyZctaKOWkgVRQUJBs3LhR5syZY/sfOHAgUx6ehmMaIK1evVrWrFljgdPx48dtbOLEiXLv3j1Zu3atzXvnzp2ydetWG4uNjZU2bdrI+vXrZceOHZInTx557733PI6tYdbMmTMt7Dp69GimzRkAAAAAAEARSnmpY8eOkj9/fgkPD7eqqT179tj23bt3S+3ataVatWoWPg0cOFBOnjwpV65ckYSEBNm3b5/069fPqpdKlSolrVq1soAoM5QsWVKKFy9uc9Lja3XWjRs35Pr169aratSoUVadVahQIaus0tBMlShRQtq1aye5cuWSkJAQC6i0aspdixYtpHDhwvaj1V8XLlzgvxgAAAAAAJBpWI/lJV0a51SgQAE5d+6cfdYQSEMfp9y5c1vQo9s1pEpMTPQY1+NcvHgxUx6ev7+/Lc3TH6W/9Xy6PE91797d9d2kpCSr4lK3bt2SWbNmWQWUVlPpPsmX5mn1lJNeh1ZXAQAAAAAAZBZCKS9p9ZGT9onSYMoZMn3//feusaioKAtwdDxfvnwWFOm+zu9rWOX87B76aGiUGRwOhxQpUsTOu2XLFqugSm7+/PkSHR0tK1eulLx589pSRF3uBwAAAAAA4Css3/OS9l/SCiNdlqfL4CIiImy7/tZ+U9rLKT4+3hqely9fXooWLWrVR/Xr15fFixdbg3RtGq5BUaNGjTyOXbp0aataelgwpUsHV61a5fVD1cosXVKojc41fNJKKO19deTIERvXbbp0T5f2aUi2fPlyr48NAAAAAACQGQilvKQhT58+faRv377Wj6lx48a2vUKFCvbmOq000l5T2ptJG58738ynY/qGu7Zt28rw4cPtTXnJQyk9pr4tr2HDhim+me/SpUty+/btND1Yfbue9rTq2rWrNGnSRKZMmSJ37961scGDB9sSQr2GoUOH2tv1AAAAAAAAfMnPoeu98EjaJHz06NEPhEkAAAAAAABIHyqlAAAAAAAA4HOEUgAAAAAAAPA5lu8BAAAAAADA56iUAgAAAAAAgM8RSgEAAAAAAMDnCKUAAAAAAADgc4RSAAAAAAAA8DlCKQAAAAAAAPgcoRQAAAAAAAB8jlAKAAAAAAAAPkcoBQAAAAAAAJ8jlAIAAAAAAIDPEUoBAAAAAADA5wilAAAAAAAA4HOEUgAAAAAAAPA5QikAAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD4XmJUHH/z1Jfn2Sozrb38/kYI5AqR8oRDp92J+eaFIDvnjkRvyydGbru/4iUiuYH8pVzBY/r9fF5CaxXJ6HPN2TKIs/e6m7LtwV65EJdgxi+UNkmalc0nPqvkkNOjnnK3D8gtyJTpBfl00VBZ1KJbinNx90KKINCqVy/X3qWsxsvz4LTl2NUYi7yVI7pAAKZkvSFqXy23zennFxVTvwcftw+VyVIJM3vNv19+/Cf/5mg79466s/utt+du1GLkblyR5QwPkhWdCpfcL+aXqM6Gu7204fcd1DJ3Dqq7FJUAvXkR++9U/5a//jvW4TgAAAAAAgKc2lHLSnEiDqLhEh/xwM04O/OOeHL50T/7rZc8A5fmCweIQkXORcRYc/W3rFfnv7sWlSK4gG/9XdIIMWH9JrkYn2N/huQMlPtEh39+Ik+9vRMrOc9Hyxw7FJCzY3+s5ucsTEuD6vO7vt+W9/dck0fFTUBae56db9ZerMRaG1S8eJlV+9fP+52/Gyd14xwPHDXMLyZL79GikLDoSaZ9DA/3kubzBcul2nOw6f1f2Xrgr4yN+Je3L53lgvwu34mXjmSjpUOHBMQAAAAAAgCeBT0KpQjkD5U+dnrPPBy7elZFbrkhiksiW76Mkd8jPoc0HLYtKeO4g+frUHZmy998Sm+CQk/+OdYVSMw782xVITW36jLQom9s+Lz52Uxb8vxty5kacLPzzDXmrXuE0zSm5CzfjXIFU0VyBMrNVUXm+4E9B053YRNnxQ7QUCvPc31mBldJxv4+Me+Acf78W4wqkqv4qROa0CbdKrKvR8fLahsvyzzvxMmP/NaldLKcUDnvwMX1yNFJalcstwQE/VUsBAAAAAAA8SR77nlLP5Ap0hUEH/3HPPv8mPIcrkFJ9q+ezqimlQZfDofVW6bf+1B0LpNTIuoVcgZSzmqpzpbySUZu/j3J9HlarkAVSSgO43/46v32OTXTIjnPRD+yrlVgazv33325neB4AAAAAAAC/2Eqp6/cSrO+Rc/meCvAXaVk2txz6513X997eesV+/xAZJ4H+In1eyC+Vf/VTX6V/3o6XJMfPy/zc+fv5SbmCIda76XZsktyMSZQCOR59adpr6qVFZz22/XlwWft9/tbPlU0vFs0hWeHirXjX5/LJrsc9BLvoNhenQb8pIO9svyJ/OhYpL7OEDwAAAAAAPIF8EkrFJ4k14ta+3AVyBEiF/210XuWZUI9QSpffORUJC5QGJcJSPF5KC9bSuogtpZ5STu6FVj5ZHOfneZb/7V/+UMXzBkmH8nnkq7/fkRXHb2Xt3AAAAAAAAJ7UUEr7Mn3dq2Sq31vfs4QkJDlkxKYrculOvLyz7Yqs/Y8SkiPIX4rlCbKwRqulTruFVyrJ4ZDvI2Ptc94Qf8kf+nPD8vT0lCqdP1gO/fOnpYJ/uXpfIkr+/Ea+zFIiX5Ac+udPn09fj7UliU7698/f86yichrwmwKy6fsoWX78ZqpVYQAAAAAAAI+bx66nVPG8wTKyTiH7fP1eoqtvUt7QAKlXPKd9Pnr5vmw7+3NPpqV/uSU/3vmpAbo2//ZLVnmUVvpWO2f/8D8cui5nb/wcEt2OSZQv/prx6qTW5X7uibXgm+sSFZton7XR+Wff3rTP2sS8WemUA7FfhQVKt8p57Y1/2hQdAAAAAADgSfJYltg0KJFTyhQItt5Sy0/ckm5V8kpIoL+8U7+wfH/jR2vyPX7nv+yNe/GJDrl2L9HVa2rISwXT1OfKXY+q+ayBeqn8wfKfDQrbG/i0T1WvL/9pbwXUnOpydLz8KmegdKuSL0PXWLFwqAyqUUD+eCRSTvw7Vtosu2Dn+OftOFvuqFVhOoeU3rzn9Gr1/LaELzouKUNzAQAAAAAA8LXHMpTSSqe+L+SXibv/JTfuJcrXp6Oka+W89ma6z7s8J5//5absvXhXrkQlWDumcgWCpVmZXNKzaj4J1WZR/+teQpKr4uhhfa7caWWWU8eKea3n1PLjt+TYlRirYAoL9pcqvwqVts//XOWUEQN/U0AqFQ6R1X+9LSf/HSMXb8dJ3pAAeaFIqPSulk+qFXl0k3WtHuv9Qj75+M+RmTIfAAAAAAAAX/FzONzbev8y3I1Lks3fR8mMA9fsbw20tMoKAAAAAAAAj4fHslIqo9otv+Ba0qZVUp0r5cnuKQEAAAAAAOCXHkppP6Y8If5S+VehMuDX+aVsgZDsnhIAAAAAAAB+6cv3AAAAAAAA8Hj7uSs4AAAAAAAA4COEUgAAAAAAAPA5QikAAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFIAAAAAAADwOUKpx0S3bt1k//79Pj3n1atXpUGDBhIdHS3ZYc+ePdK+fftsOTcAAAAAAMhehFKPiS+++MICosy0YcMG6dmz50PHixQpYkFYrly5MvW8AAAAAAAAqSGUAgAAAAAAgM8RSnlBl5jNnTvXfjdv3lwWLlwoDofDNX7kyBFp1KiRbNmyRdq1a2cVT/PmzbOxyMhIefvtt6Vp06bSoUMHWbZsmcex+/fvb99/6aWXbDmbuzt37sikSZOkZcuW0qZNG/nkk088zhsVFSVTp061MT3/m2++KXFxcXLq1Ck75vTp0+Xs2bP2WX+2bdvmsVywfv36UqNGDTuOu0fN2XmtixcvtvG2bdvKN99849U/Np273rsWLVrYfTp+/LjX15vaeW/fvi0jR46Uxo0bS5MmTWTIkCGSlJTkGt+7d69VjekxBgwYIBcuXPBqzgAAAAAAIGsEZtFxf3EOHz4sS5culfv371uoUbFiRQs4nGJiYuTAgQOyfPlyCQ0NtTBIaTAUFBQkGzdulCtXrsigQYOkZMmSFgipzz77zH6n1Ftp4sSJEhwcLGvXrrXzvv766/Lcc89Jq1atbHzChAkSGBgoK1askJw5c1rwokFMhQoVbFmeLt9buXKljae0XPDy5csWOiWX2pzv3bsnAQEBFnItWrRI5syZk+I5ktu1a5fNacmSJXaPNDhKy/U+6rwanOkz2Lx5s/j7+8uf//xn8fPzs7GTJ0/K+PHj5cMPP5QXX3xRvvrqK3nrrbdk9erVdjwAAAAAAOB7VEp5qWPHjpI/f34JDw+3aqnkVU0JCQkyYsQIyZs3r4SEhEjlypVt2759+6Rfv34WwpQqVcoClp07d6Z6vuvXr1uwNGrUKAkLC5NChQpZcLV9+3aP8TFjxki+fPkszNF56Xkywts59+jRwwKdhg0bysWLF706toZmWiVVtGhRu5edOnXy+npTO68GUHfv3pUff/zR7kW9evVcodS6deus+kqrwnTfV155xc7nDA4BAAAAAIDvUSnlpYIFC7o+FyhQQM6dO+cxniNHDilcuLDHtlu3bkliYqIFLO7H8SbE0Tfjqe7du7u2aRVU2bJlXeMavrgfOzN4M2etytIKLaUVVbGxsV4d++bNm675O4/r7fWmdt6+ffva3xrS6TW0bt3alvNp1ZQe++jRox7BWnx8vFy7dk3Kly+fhrsDAAAAAAAyC6GUl7Syxr3nkgZT7lJaBqYVTLpd93V+/8aNGw/s+7A34+m+2qcqpeonHdf+URqsJA/DnJyVQmmRkTmnRo+h985Jj+vt9aZG3yCoIZT+/PDDD7bEslatWrbk8JlnnrF+UsOGDcvwNQAAAAAAgMzB8j0vrV+/3ipwtMeSLimLiIhIdR+t6tFQRJtza7+j8+fPW+ji3ovqYbRSqXbt2jJ79myJjo626qUzZ85Yw2/nuC5RmzFjhs1LK3+0Z5Oex0nDKp1v8kbmWTXn5MsdV61a5bFNm5BrPyidk1ZNaW8nb683NQcPHrTm5doYXauotMpKgyr18ssvy9dffy3fffedbdf7ob2n9J4BAAAAAIDsQaWUlzQw6dOnjwU12gtJAxZvjB07Vt577z17W5wu8evVq5fXAc/kyZPtrX9du3a1Jt/Fixe3CqDk49pnScerV6/uakautIdS3bp1pUuXLhY26Rv1nMHQlClTXG+20zfdKQ249PsZmbPTpUuX7I147jTI0zcDvvrqq9Z3S/tLub8RMLXrfRTtJfX+++9bJVaePHmsJ5beD1WlShUZN26cfPDBB/LPf/7TKrH03mgPLgAAAAAAkD38HM5kAg+lDbdHjx6d5mAmLTQY0jfEafUTAAAAAADALx3L9x4DuuxMK3z0TXcAAAAAAABPA5bvZaN169bJrFmzrAfSwIEDJTw8PDunAwAAAAAA4DMs3wMAAAAAAIDPsXwPAAAAAAAAPkcoBQAAAAAAAJ8jlAIAAAAAAIDPEUoBAAAAAADA5wilAAAAAAAA4HOEUo+Jbt26yf79+316zqtXr0qDBg0kOjpassOePXukffv28riYNm2azJ07N7unAQAAAADAU8HP4XA4snsSyBobNmyQlStXyooVKx7LW6yh1KxZs2yeAAAAAADg6UKlFAAAAAAAAHyOUMoLusRMl3Xp7+bNm8vChQvFvcDsyJEj0qhRI9myZYu0a9fOlsTNmzfPxiIjI+Xtt9+Wpk2bSocOHWTZsmUex+7fv799/6WXXrLKIXd37tyRSZMmScuWLaVNmzbyySefeJw3KipKpk6damN6/jfffFPi4uLk1KlTdszp06fL2bNn7bP+bNu2zWO5YP369aVGjRp2HHePmrPzWhcvXmzjbdu2lW+++carf2w6d713LVq0sPt0/Phx15jOU+dz794917YdO3ZIly5dvDqvzrFTp052DB1bvny5x/MbNWqUPTvdX//u16+fxMTE2PjmzZvt/tSqVcsqt5J72H0GAAAAAADpF5iBfZ8qhw8flqVLl8r9+/dlwIABUrFiRQsonDTgOHDggIUhoaGhFrIoDYaCgoJk48aNcuXKFRk0aJCULFnSwhP12Wef2e+UeitNnDhRgoODZe3atXbe119/XZ577jlp1aqVjU+YMEECAwNteV7OnDll7969kpSUJBUqVLD+VI9avvfFF1/I5cuXLXRKLrU5a3AUEBBgIdeiRYtkzpw5Xi0R3LVrl81pyZIldo+GDBniGitbtqwUL17cgjkNf9TWrVtdn1M7b1hYmMycOVNKlSolJ0+elIEDB0q1atWkatWqNt65c2epU6eO3e+vvvpKhg4daqFYzZo1pXXr1vbz7rvvpjjvh91nAAAAAACQflRKealjx46SP39+CQ8Pt4qb5FVNCQkJMmLECMmbN6+EhIRI5cqVbdu+ffusKkdDGA1MNFDauXNnque7fv26BUta4aOBS6FChSy42r59u8f4mDFjJF++fBZe6bz0PBnh7Zx79OhhAVHDhg3l4sWLXh1bwxytkipatKjdS61scqcVTlptprT5+qFDh2ybN+fVY5UpU0b8/f2lSpUqUq5cOTl9+rRrXEM1Db30+el1FStWTG7cuJHqnLPqPgMAAAAA8LSjUspLBQsWdH0uUKCAnDt3zmM8R44cUrhwYY9tt27dksTERAuU3I/jTYijb8ZT3bt3d23T6hytKHKOa0DifuzM4M2ctVpIK4eUVlTFxsZ6deybN2+65u88rjutVvroo49s+eDBgwctXCpSpIhX59WqKq1k0/ui89fKsvj4eNe4hlUaZumP0t/6vdRk1X0GAAAAAOBpRyjlJa2YcdLQRIMpd86ww51W1uh23df5fa3OSb5vSjSM0X21ciilqhwd175G165deyAMc/Lz8/Pq2jJrzqnRY+i9c0peqaTj2ltLl+fpUsjkVVKPCo50iZ328dIeWRpA9erVy6P/Vkq8efGkN/cZAAAAAACkHcv3vLR+/XqrItIeS7qELiIiItV9tKpH+zBpc23tOXX+/HkLmdx7UT2MVubUrl1bZs+ebUvZtKrnzJkz1vDbOV6vXj2ZMWOGzUurgrRnk7N5t9IQReebvJF5Vs05+XLHVatWeWxr3LixBU46J62a0t5OyWkQ9d///d/W76lJkyZenUt7TWnA5AzOtGrK2dMro7y5zwAAAAAAIO0IpbykAVGfPn2kb9++9uY4DVi8MXbsWKu00bBl+PDhVsHjbcAzefJk6/HUtWtXC2imTJkid+/e9RjXHlbaZ6lZs2YWnGmVkJNWDdWtW9feYKcNw3fv3m3bNRjSt83pG/iUjunf2sMpo3N2unTpkty+fdtjmwZ52lj91VdftfvobJzuTntFaQWV3u9cuXJ5da7SpUvbWwwHDx5s9+HYsWOuBufe0Pun169v4VuzZo191kbp3t5nAAAAAACQdn4Ob9YwPeW0wfjo0aPTHMykhQZD48ePt6qcp52+KU/DMG+DPwAAAAAA8OSh3OMxcOHCBeu1pG+6e9rpWw21GkyrlQAAAAAAwC8Xjc6z0bp162TWrFn2JjldLhYeHi5Ps//4j/+wcE4rxpxv2QMAAAAAAL9MLN8DAAAAAACAz7F8DwAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFIAAAAAAADwOUKpx0S3bt1k//79Pj3n1atXpUGDBhIdHe3T88bExMjIkSMlIiLCzn///n2fnh8AAAAAAGQ/P4fD4cjuSSBrbNiwQVauXCkrVqx4rG7xxo0bZfHixbJ8+XIJDg5O077t27eX0aNHS6NGjbJsfgAAAAAAIOtRKQWfu3LlihQvXjzNgRQAAAAAAPjlIJTysjpn7ty59rt58+aycOFCcS8wO3LkiFXubNmyRdq1a2dL0ubNm2djkZGR8vbbb0vTpk2lQ4cOsmzZMo9j9+/f377/0ksvyZ49ezzG7ty5I5MmTZKWLVtKmzZt5JNPPvE4b1RUlEydOtXG9PxvvvmmxMXFyalTp+yY06dPl7Nnz9pn/dm2bZvHcsH69etLjRo17DjuHjVn57VqpZOOt23bVr755huv/rHp9ek8Pv30Uzlw4IBrXs7le4MGDfKo6tKKqEWLFtnnMWPG2Hd1yeH48ePtc48ePVzffdS+qT2j1O4zAAAAAADIfIFZcMxfpMOHD8vSpUstQBkwYIBUrFjRYwmZ9knSoEWXpIWGhloYpDQYCgoKsiVrWiGk4UnJkiUtEFKfffaZ/dbAK7mJEydaNdHatWvtvK+//ro899xz0qpVKxufMGGCBAYGWhiTM2dO2bt3ryQlJUmFChWsP9Wjlu998cUXcvnyZQudkkttzvfu3ZOAgAALuTT4mTNnjldLBPV+6bx0nzNnzsisWbO8vv8zZszI8PK9hz2j1O4zAAAAAADIfFRKealjx46SP39+CQ8Pt2qp5FVNCQkJMmLECMmbN6+EhIRI5cqVbdu+ffukX79+FoKUKlXKgo6dO3emer7r169bgDNq1CgJCwuTQoUKWSCzfft2j3GtIMqXL5+FKjovPU9GeDtnrVLSYKphw4Zy8eJFeRKk9IxSu88AAAAAACBrUCnlpYIFC7o+FyhQQM6dO+cxniNHDilcuLDHtlu3bkliYqIFHe7H8SbE0WVqqnv37q5tWgVVtmxZ17gGUe7HzgzezFmrsrRCS2lFVWxsrDwJUnpGqd1nAAAAAACQNQilvKQVNe49lzSYcqdVQ8lpBZNu132d379x48YD+6akSJEitq/2QEqp+knHtX/UtWvXHghanPz8/Ly6tsyac0Zp9ZIGYk5379594Dv+/v7p3jelZ5TafQYAAAAAAFmD5XteWr9+vVURaY8lXdoVERGR6j5aTaR9mLQpuPYzOn/+vIUf3vRD0kql2rVry+zZsyU6OtoCF+3DpA27neP16tWzXks6r/j4eNm1a5edx0nDKp1v8kbmWTXn5MsdV61alaZ9SpQoIcePH7fP2u/qxIkTD3xHr1vvQ3r2Tc99BgAAAAAAWYNQyksaXPTp00f69u1rb29r3LixV/uNHTvWKpr0LXXDhw+XXr16eR3wTJ482fogde3aVZo0aSJTpkzxqADSce2PpP2dmjVrZsGZeyWRvlmvbt260qVLF3ur3O7du227NijXt8/pG/iUjunfhw4dyvCcnS5duiS3b99O0z56Hl1Op/PSN+NVq1btge9o0/XNmzdbnyttOJ+WfdN7nwEAAAAAQObzczgcjiw47i9KRt745i0NhsaPH2/VTwAAAAAAAL90VEo9Bi5cuGB9qvRNdwAAAAAAAE8DGp1no3Xr1smsWbPsDXYDBw6U8PDw7JwOAAAAAACAz7B8DwAAAAAAAD7H8j0AAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKPSa6desm+/fv9+k5r169Kg0aNJDo6GifnO/IkSPSqFGjVL83bdo0mTt37iO/s2fPHmnfvv1Dx3v27CkbNmxI1zwBAAAAAEDWC/TBOeCFL774ItPvk4YyK1eulBUrVqQ4XqRIEZ8HYd4YN25cdk8BAAAAAABkMSqlAAAAAAAA4HOEUl7QZWK6nEx/N2/eXBYuXCgOh+OBZWlbtmyRdu3a2ZK4efPm2VhkZKS8/fbb0rRpU+nQoYMsW7bM49j9+/e377/00ku2JM3dnTt3ZNKkSdKyZUtp06aNfPLJJx7njYqKkqlTp9qYnv/NN9+UuLg4OXXqlB1z+vTpcvbsWfusP9u2bfNYLli/fn2pUaOGHcfdo+bsvNbFixfbeNu2beWbb77x/l+cyEP33bx5s82zVq1aMmvWLI999Lr1vrdo0cLu8fHjxz3Gz507J/369ZOGDRvaPUtMTPQY37t3ry3p07kPGDBALly44BobNGiQzJw5U1599VXbf8yYMZKQkJCmawIAAAAAAGlDKOWlw4cPy9KlS2XJkiW2LE5DDncxMTFy4MABWb58uezYsUOaNGli2zUYCgoKko0bN8qcOXNsf/2e02effWZL6HQpXXITJ06Ue/fuydq1a+3cO3fulK1bt7rGJ0yYILdu3bLleRo4aciTlJQkFSpUsGOOHTtWypYta5/1RwMd9+WCD1symNqcdU4BAQGuc+p3vPWofVu3bm3z1N/J7dq1y+67zuXzzz+XgwcPegRWuuSvTp069r0qVarIDz/84Bo/efKkjB8/XkaNGmX3sFWrVvLWW295BFeHDh2yYOrLL7+Uo0ePelwvAAAAAADIfIRSXurYsaPkz59fwsPDrVoqeVWTVtaMGDFC8ubNKyEhIVK5cmXbtm/fPqvgCQ0NlVKlSlkgosFIaq5fv24BjQYpYWFhUqhQIavU2r59u8e4VvXky5dPgoODbV56nozwds49evSwcEkriy5evJimc6RnXw0BNVQrWrSoPYdOnTq5xn788UerCOvdu7cEBgZK586d7Tk4rVu3zqrNtCpMz/vKK6/Y/dN9nPTYhQsXtp9KlSp5VFIBAAAAAIDMR6NzLxUsWND1uUCBArZczF2OHDks0HCnVUxajaOBkvtxvAli9M14qnv37q5tWgWllU/OcQ2i3I+dGbyZc86cOS38UVpRFRsb6/Xx07vvzZs3XdfunJP7mAZoGt4pPz8/e0ZOeq+0+sk9WIuPj5dr165J+fLl7e88efK4xtJ6TQAAAAAAIO0IpbyklTXuPZfcQw+lFTjJaQWTbtd9nd+/cePGA/umRJfz6b7apyql6icd1/5RGqwkD8OcNJxJq4zMOSvp+fW+O+mc3Md0+eTdu3ctmNLlfO7ffeaZZ6yf1LBhw3w+bwAAAAAAkDKW73lp/fr1VkV05coVW0IXERGR6j5aEaTNxLWxt4Ym58+ft5BJm22nRiuVateuLbNnz5bo6GirXjpz5ow1GneO16tXT2bMmGHz0sof7aek53HSsErnm7yReVbNOflyx1WrVklmady4sfWh0uvRyqivvvrKNfbss8/K888/bw3Zdfmh9uC6ffu2a/zll1+Wr7/+Wr777jurNtP7oU3V9Z4BAAAAAIDsQSjlJQ2I+vTpI3379rW3v2lI4g1tNq4VTdrUe/jw4dKrVy+vA57JkydbyNK1a1drnD5lyhSrBnIf195J2qOpWbNmFpz5+//8SLWHUt26daVLly72hr7du3fbdg139C13+gY+pWP6tzb7zuicnS5duuQRDHlDr0PnoYHRmjVr7PPAgQNtTENAfROgviFPn4EGZ+70LYTajF7vkzY2L1OmjGtMG59rI/QPPvjAnpv2lNJG5umpJAMAAAAAAJnDz6FrnfBI2mB89OjRaQ5m0kKDIX1DnFY/AQAAAAAA/NJRKfUY0De9aQ8kfdMdAAAAAADA04BG59lo3bp1MmvWLHvbmy5TCw8Pz87pAAAAAAAA+AzL9wAAAAAAAOBzLN8DAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAAOBzhFIAAAAAAADwOUIpAAAAAAAA+Byh1GOiW7dusn//fp+e8+rVq9KgQQOJjo6Wp8HTdr0AAAAAADzO/BwOhyO7J4GssWHDBlm5cqWsWLHiqZjzk3i9AAAAAAA8raiUAgAAAAAAgM8RSnmhffv2MnfuXPvdvHlzWbhwobgXmB05ckQaNWokW7ZskXbt2tkSsXnz5tlYZGSkvP3229K0aVPp0KGDLFu2zOPY/fv3t++/9NJLsmfPHo+xO3fuyKRJk6Rly5bSpk0b+eSTTzzOGxUVJVOnTrUxPf+bb74pcXFxcurUKTvm9OnT5ezZs/ZZf7Zt2+axXLB+/fpSo0YNO467R83Zea2LFy+28bZt28o333zj9T+4TZs2ycsvvywNGza031u3brXtqc1Zl9699tprdk7d/vrrr8uPP/7o1b6pXa86cOCA9OnTRyIiImxeO3bs8PqaAAAAAABA2gWmY5+n0uHDh2Xp0qVy//59GTBggFSsWNHCGaeYmBgLNpYvXy6hoaEWjigNSoKCgmTjxo1y5coVGTRokJQsWdICEvXZZ5/Zbw28kps4caIEBwfL2rVr7bwaxDz33HPSqlUrG58wYYIEBgbacrWcOXPK3r17JSkpSSpUqGD9qR61nO2LL76Qy5cvW+iUXGpzvnfvngQEBFjos2jRIpkzZ45XS+b0HmnI9uGHH0qdOnUsaLp586aNpTbn2NhYC9/ef/99m5se57333rPwL6PX+7e//U3GjRtnx6tVq5b861//kgsXLqR6PQAAAAAAIP2olPJSx44dJX/+/BIeHm7VUsmrmhISEmTEiBGSN29eCQkJkcqVK9u2ffv2Sb9+/SyoKlWqlAVKO3fuTPV8169ft6Bl1KhREhYWJoUKFbLgavv27R7jY8aMkXz58ll4pfPS82SEt3Pu0aOHBVNa8XTx4kWvjq2Bme5z6dIluXv3rhQpUsTCPW+UKFHCqtBy5cpl91cDqtOnT0tmWLdunVWj1a1b1+anz1g/AwAAAACArEOllJcKFizo+lygQAE5d+6cx3iOHDmkcOHCHttu3boliYmJFii5H8ebEEeriFT37t09Qp2yZcu6xjWIcj92ZvBmzlqVpRVaSquWtIrJG7rfzJkzZdWqVbJgwQIpXry4vPPOO1KlShWv5jVr1iw5evSoVWrpHJ1zyCitjPr1r3+dKccCAAAAAADeIZTyklYmufdc0mDKnVbYJKcVTLpd93V+/8aNGw/smxKtItJ9tU9VStVPOq79o65du/ZAGObk5+fn1bVl1py9oRVI+hMfHy9/+MMfbDmeLotMbc7z58+X6OhoW56n1Wi6VFGXN2b0etUzzzzj6k8FAAAAAAB8g+V7Xlq/fr1V62iPJV1Cpw2xU6OVPNqHSZuCaz+l8+fPW8jk3ovqYbRSqXbt2jJ79mwLY7Qy6MyZM9Zo3Dler149mTFjhs1LQ55du3bZeZw0rNL5ptTYOyvmnHy5o1ZEudN56hy1P5YGSP7+/rYcz93D5qz3QL+rSxk1JNPeXcml53qVNjbXazx48KDdZ62c+p//+Z80HQMAAAAAAKQNoZSXNCDSt7P17dvXehs1btzYq/3Gjh1rFU36lrrhw4dLr169vA54Jk+ebD2eunbtKk2aNJEpU6ZYLyb3ca0a0v5OzZo1s+BMgx4nfdOcViV16dLFejDt3r3btmuDcn07nb6RTumY/n3o0KEMz9lJ+0bdvn3bY5u+OXD16tXSunVrm682g9fle+4eNufBgwfbEkK970OHDrVG6cml93p1+eDvf/97+eijj+w+a2N3DcEAAAAAAEDW8XNoUoBH0gbjo0ePTnMwkxYalIwfP96qnwAAAAAAAH7pqJR6DFy4cMH6VOmb7gAAAAAAAJ4GNDrPRuvWrbM3yukb7AYOHCjh4eHZOR0AAAAAAACfYfkeAAAAAAAAfI7lewAAAAAAAPA5QikAAAAAAAD4HKEUAAAAAAAAfI5QCgAAAAAAAD5HKAUAAAAAAACfI5QCAAAAAACAzxFKAQAAAAAAwOcIpQAAAAAAACC+9v8DMZr6ovLxAfYAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1200x1080 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Group keys by prefix (encoder, decoder, etc.)\n",
    "from collections import defaultdict\n",
    "\n",
    "groups = defaultdict(list)\n",
    "for key in sorted(all_keys):\n",
    "    prefix = key.split('.')[0] if '.' in key else 'other'\n",
    "    groups[prefix].append(key)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(12, max(4, len(all_keys)*0.35 + 1)))\n",
    "ax.axis('off')\n",
    "\n",
    "y = 0.98\n",
    "colors = ['#3498db','#e67e22','#2ecc71','#9b59b6','#e74c3c','#1abc9c','#f39c12']\n",
    "for gi, (grp, keys) in enumerate(sorted(groups.items())):\n",
    "    color = colors[gi % len(colors)]\n",
    "    ax.text(0.01, y, grp.upper(), transform=ax.transAxes,\n",
    "            fontsize=10, fontweight='bold', color=color)\n",
    "    y -= 0.05\n",
    "    for key in keys:\n",
    "        ax.text(0.04, y, f'  {key}', transform=ax.transAxes,\n",
    "                fontsize=9, fontfamily='monospace', color='#333333')\n",
    "        y -= 0.04\n",
    "    y -= 0.02\n",
    "\n",
    "ax.set_title('ComponentRegistry: All Registered Keys', fontsize=13, pad=15)\n",
    "plt.tight_layout(); plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78a47025",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## Summary\n",
    "\n",
    "| Concept | Purpose |\n",
    "|---------|--------|\n",
    "| `BaseAttentiveSpec` | Frozen, JSON-serialisable model blueprint |\n",
    "| `BaseAttentiveComponentSpec` | Declarative selection of component keys |\n",
    "| `ComponentRegistry.register()` | Plug in a custom builder by string key |\n",
    "| `assemble_model()` | Resolve a `BaseAttentiveV2Assembly` from a spec |\n",
    "| `BaseAttentiveV2` | Trainable resolver-driven model scaffold |\n",
    "| `serialize_base_attentive_spec()` | Stable JSON export for saved experiments |\n",
    "| `dataclasses.replace()` | Derive spec variants without mutation |\n",
    "| `component_registry._registry` | Inspect all available component keys |\n",
    "\n",
    "### Key Takeaways\n",
    "\n",
    "- **Spec = reproducibility**: a spec file fully defines the model — no code needed\n",
    "- **Keyword approach** remains valid for quick prototyping\n",
    "- **Spec approach** shines for experiment tracking, hyperparameter sweeps, and deployment\n",
    "- The registry is extensible: register any builder function by a string key"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.20"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}