testGenerateEmbeddings.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from embeddingsModule.generateEmbeddings import BERTEmbeddingsGenerator\n",
    "\n",
    "texts = [\n",
    "        \"bank\", \"he eventually sold the shares back to the bank at a premium.\", \n",
    "        \"the bank strongly resisted cutting interest rates.\"\n",
    "        # Add more sentences \n",
    "    ]\n",
    "\n",
    "    # choose your output directory in order to save your embeddings\n",
    "output_directory = 'embeddings_stored'\n",
    "    \n",
    "embeddings_generator = BERTEmbeddingsGenerator()\n",
    "embeddings_generator.generate_embeddings(texts, output_directory)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### t-SNE algorithm\n",
    "\n",
    "Implementing t-SNE (t-Distributed Stochastic Neighbor Embedding) for visualizing BERT embeddings involves several steps. t-SNE is commonly used to reduce high-dimensional data into a lower-dimensional space (usually 2D or 3D) while preserving the local structure of the data points as much as possible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Embeddings Shape: (13, 768)\n",
      "Encoded Labels Shape: (13,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Load .tsv differently\n",
    "tsv_embeddings = 'embeddings_stored/embeddings.tsv'\n",
    "tsv_metadata = 'embeddings_stored/metadata.tsv'\n",
    "\n",
    "# Load embeddings and labels\n",
    "bert_embeddings = np.loadtxt(tsv_embeddings, delimiter='\\t')\n",
    "# Sample metadata labels\n",
    "metadata_labels = [\n",
    "    'bank', 'eventually', 'sold', 'shares', 'back', \n",
    "    'bank', 'premium', 'bank', 'strongly', 'resisted', \n",
    "    'cutting', 'interest', 'rates'\n",
    "]\n",
    "\n",
    "# Create a dictionary to map labels to numerical values\n",
    "label_map = {label: i for i, label in enumerate(set(metadata_labels))}\n",
    "encoded_labels = np.array([label_map[label] for label in metadata_labels])\n",
    "\n",
    "# Check shapes\n",
    "print(\"Embeddings Shape:\", bert_embeddings.shape)\n",
    "print(\"Encoded Labels Shape:\", encoded_labels.shape)\n",
    "\n",
    "# Define the algorithm with reduced perplexity\n",
    "perplexity = min(30, bert_embeddings.shape[0] - 1)  # Adjust to a reasonable value\n",
    "tsne = TSNE(n_components=2, perplexity=perplexity, random_state=42)\n",
    "tsne_embeddings = tsne.fit_transform(bert_embeddings)\n",
    "\n",
    "# Visualize t-SNE embeddings\n",
    "plt.figure(figsize=(10, 8))\n",
    "plt.scatter(tsne_embeddings[:, 0], tsne_embeddings[:, 1], c=encoded_labels, cmap='viridis')\n",
    "plt.colorbar()\n",
    "plt.title('t-SNE Visualization of BERT Embeddings')\n",
    "plt.xlabel('t-SNE Dimension 1')\n",
    "plt.ylabel('t-SNE Dimension 2')\n",
    "plt.show()\n"
   ]
  }
 ],
 "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}