analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/pb/_0h8x45x091307x0pwzykw8h0000gn/T/ipykernel_79873/2433061539.py:42: FutureWarning: \n",
      "\n",
      "The `ci` parameter is deprecated. Use `errorbar=None` for the same effect.\n",
      "\n",
      "  sns.barplot(\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Detailed Results:\n",
      "metric                                               Hits@1  Hits@10  Hits@5  \\\n",
      "base_ontology embedding_type model                                             \n",
      "go-basic      anc2vec        torch_mlp                  NaN      NaN     NaN   \n",
      "              biobert        torch_mlp                  NaN      NaN     NaN   \n",
      "              owl2vec        MLPClassifier            0.043    0.303   0.184   \n",
      "go-full       anc2vec        DecisionTreeClassifier   0.000    0.001   0.000   \n",
      "              biobert        MLPClassifier            0.016    0.155   0.077   \n",
      "              owl2vec        MLPClassifier            0.042    0.301   0.183   \n",
      "\n",
      "metric                                                 MRR   epoch  \\\n",
      "base_ontology embedding_type model                                   \n",
      "go-basic      anc2vec        torch_mlp                 NaN   3.333   \n",
      "              biobert        torch_mlp                 NaN  34.000   \n",
      "              owl2vec        MLPClassifier           0.125     NaN   \n",
      "go-full       anc2vec        DecisionTreeClassifier  0.001     NaN   \n",
      "              biobert        MLPClassifier           0.066     NaN   \n",
      "              owl2vec        MLPClassifier           0.123     NaN   \n",
      "\n",
      "metric                                               learning_rate  \\\n",
      "base_ontology embedding_type model                                   \n",
      "go-basic      anc2vec        torch_mlp                       0.001   \n",
      "              biobert        torch_mlp                       0.000   \n",
      "              owl2vec        MLPClassifier                     NaN   \n",
      "go-full       anc2vec        DecisionTreeClassifier            NaN   \n",
      "              biobert        MLPClassifier                     NaN   \n",
      "              owl2vec        MLPClassifier                     NaN   \n",
      "\n",
      "metric                                               train_loss  val_loss  \n",
      "base_ontology embedding_type model                                         \n",
      "go-basic      anc2vec        torch_mlp                    0.087     0.094  \n",
      "              biobert        torch_mlp                    0.025     0.085  \n",
      "              owl2vec        MLPClassifier                  NaN       NaN  \n",
      "go-full       anc2vec        DecisionTreeClassifier         NaN       NaN  \n",
      "              biobert        MLPClassifier                  NaN       NaN  \n",
      "              owl2vec        MLPClassifier                  NaN       NaN  \n"
     ]
    }
   ],
   "source": [
    "import wandb\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from typing import List, Dict\n",
    "\n",
    "# Initialize wandb\n",
    "api = wandb.Api()\n",
    "\n",
    "# Get runs from your project\n",
    "runs = api.runs(\"go-embedding-evaluation\")  # Replace with your actual project name\n",
    "\n",
    "# Collect data\n",
    "data = []\n",
    "for run in runs:\n",
    "    config = run.config\n",
    "    # Get the last logged metrics\n",
    "    history = run.history()\n",
    "    if not history.empty:\n",
    "        metrics = history.iloc[-1].to_dict()\n",
    "\n",
    "        # Extract metrics for each embedding type and model\n",
    "        for key, value in metrics.items():\n",
    "            if isinstance(value, (int, float)):  # Filter out non-numeric values\n",
    "                # Parse the metric key (format: \"embedding_type/model/metric\")\n",
    "                parts = key.split('/')\n",
    "                if len(parts) == 3:\n",
    "                    embedding_type, model, metric = parts\n",
    "                    data.append({\n",
    "                        'base_ontology': config.get('base_ontology', 'unknown'),\n",
    "                        'embedding_type': embedding_type,\n",
    "                        'model': model,\n",
    "                        'metric': metric,\n",
    "                        'value': value\n",
    "                    })\n",
    "\n",
    "# Convert to DataFrame\n",
    "df = pd.DataFrame(data)\n",
    "\n",
    "# Create visualization\n",
    "plt.figure(figsize=(15, 8))\n",
    "sns.barplot(\n",
    "    data=df[df['metric'].isin(['MRR', 'Hits@1', 'Hits@5', 'Hits@10'])],\n",
    "    x='embedding_type',\n",
    "    y='value',\n",
    "    hue='base_ontology',\n",
    "    ci=None\n",
    ")\n",
    "plt.title('Embedding Performance Comparison by Ontology')\n",
    "plt.xticks(rotation=45)\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "# Create summary table\n",
    "summary_table = df.pivot_table(\n",
    "    index=['base_ontology', 'embedding_type', 'model'],\n",
    "    columns='metric',\n",
    "    values='value',\n",
    "    aggfunc='mean'\n",
    ").round(3)\n",
    "\n",
    "# Display the table\n",
    "print(\"\\nDetailed Results:\")\n",
    "print(summary_table)\n",
    "\n",
    "# Optionally save to CSV\n",
    "summary_table.to_csv('embedding _evaluation_results.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "default",
   "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": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}