Training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from src.scrapping.scrapper import AudibleScrapper\n",
    "from src.graph.data_manager import DataManager\n",
    "from src.graph.book_graph import BookGraph\n",
    "from src.link_prediction.link_dataset import LinkDataset\n",
    "from src.link_prediction.link_predictor import LinkPredictor\n",
    "\n",
    "import pytorch_lightning as pl\n",
    "from torch.utils.data import DataLoader\n",
    "from torch_geometric.transforms import RandomLinkSplit\n",
    "\n",
    "%reload_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dump data from database"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dm = DataManager()\n",
    "dump = dm.get_connected_scrapped_subgraph()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a graph from the dump and convert it to usable tensors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "book_graph = BookGraph(dump)\n",
    "data = book_graph.to_torch()\n",
    "book_graph.plot(sample=10, figsize=(20,10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split the data into three sets (Train / Test / Val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = RandomLinkSplit()\n",
    "train_data, val_data, test_data = transform(data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create Dataset objects for each of these sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = LinkDataset(train_data, data)\n",
    "val_dataset = LinkDataset(val_data, data)\n",
    "test_dataset = LinkDataset(test_data, data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create loader objects for each sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loader = DataLoader(train_dataset, batch_size=1)\n",
    "val_loader = DataLoader(val_dataset, batch_size=1)\n",
    "test_loader = DataLoader(test_dataset, batch_size=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Optimization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "study = LinkPredictor.optimize(data, train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Extract best model & save it"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Single model training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create model for predicting links"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create model for predicting links.\n",
    "model = LinkPredictor(\n",
    "        in_channels=4,\n",
    "        hidden_channels=16,\n",
    "        out_channels=1,\n",
    "        embedding_size=16,\n",
    "        num_nodes=data.num_nodes\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# trainer = pl.Trainer(max_epochs=10)\n",
    "# trainer.fit(model, train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save model & data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save_embeddings()\n",
    "train_dataset.save_data()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.13 ('audible')",
   "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": "3.9.13"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "add623dea2dde904fc0e9b9fd05a92b410f6d2f7af6f6548b97060ad30b8253c"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}