{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from scipy.stats import pearsonr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sorted_datasets = [\"sst2\", \"agnews\", \"dbpedia\", \"20newsgroups\", \"banking77\"]\n",
    "dataset2name = {\n",
    "    \"sst2\": {\"label\": \"SST-2\", \"color\": \"tab:blue\"},\n",
    "    \"agnews\": {\"label\": \"AG News\", \"color\": \"tab:orange\"},\n",
    "    \"dbpedia\": {\"label\": \"DBpedia\", \"color\": \"tab:green\"},\n",
    "    \"20newsgroups\": {\"label\": \"20 Newsgroups\", \"color\": \"tab:red\"},\n",
    "    \"banking77\": {\"label\": \"Banking77\", \"color\": \"tab:purple\"},\n",
    "}\n",
    "models = [(\"llama3.2-1b-instruct\", \"Llama3.2-1B\", \"o\"), (\"qwen2.5-7b-instruct\", \"Qwen2.5-7B\", \"x\")]\n",
    "\n",
    "def read_results(model, metric, cal_method):\n",
    "    df = pd.read_json(f\"../outputs/results_paper/{model}/{metric}.jsonl\", lines=True)\n",
    "    df_cal = df[(df[\"size\"] == 256) & (df[\"method\"] == cal_method)].drop(columns=[\"size\", \"min_result\"]).copy()\n",
    "    df_cal[\"uncal\"] = 0.\n",
    "    for dataset in sorted_datasets:\n",
    "        df_cal.loc[df_cal[\"dataset\"] == dataset, \"uncal\"] = df.loc[(df[\"dataset\"] == dataset) & (df[\"method\"] == \"no_adaptation\"), \"result\"].values[0]\n",
    "    df_cal[\"delta\"] = df_cal[\"uncal\"] - df_cal[\"result\"]\n",
    "    df_cal[\"delta_perc\"] = 100 * df_cal[\"delta\"] / df_cal[\"uncal\"]\n",
    "    df_cal = df_cal.rename(columns={\"result\": \"Calibrated\", \"uncal\": \"NoA\", \"delta\": \"Calibration Loss\", \"delta_perc\": \"Relative Calibration Loss\"})\n",
    "    return df_cal\n",
    "\n",
    "def plot(metric, method, y=\"cal_loss\"):\n",
    "\n",
    "    x = \"NoA\"\n",
    "    fig, axes = plt.subplots(1, 2, figsize=(10, 3))\n",
    "    for model, model_name, ax in [(\"llama3.2-1b-instruct\", \"Llama3.2-1B\", axes[0]), (\"qwen2.5-7b-instruct\", \"Qwen2.5-7B\", axes[1])]:\n",
    "        df = read_results(model, metric, method)\n",
    "        corr, p = pearsonr(df[x].values, df[y].values)\n",
    "        for dataset in sorted_datasets:\n",
    "            df_dataset = df[df[\"dataset\"] == dataset]\n",
    "            ax.plot(df_dataset[x], df_dataset[y], marker=\"o\", label=dataset2name[dataset])\n",
    "        ax.set_title(f\"{model_name}\\nCorrelation: {corr:.3f} (p={p:.2e})\")\n",
    "        ax.set_xlabel(x)\n",
    "        ax.grid(True)\n",
    "\n",
    "    axes[0].set_ylabel(y)    \n",
    "    axes[1].legend(title=\"Dataset\", bbox_to_anchor=(1.05, 1), loc='upper left')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    fig.savefig(f\"../outputs/results_paper/corr_{metric}_{method}.png\")\n",
    "\n",
    "def plot_all_in_one(method, y):\n",
    "\n",
    "    x = \"NoA\"\n",
    "    fig, axes = plt.subplots(1, 2, figsize=(10, 4))\n",
    "    for metric, ax in zip([\"nce\", \"ner\"], axes):\n",
    "        x_values, y_values = [], []\n",
    "        for model, _, marker in models:\n",
    "            df = read_results(model, metric, method)\n",
    "            x_values.append(df[x].values)\n",
    "            y_values.append(df[y].values)\n",
    "            for dataset in sorted_datasets:\n",
    "                df_dataset = df[df[\"dataset\"] == dataset]\n",
    "                ax.plot(df_dataset[x], df_dataset[y], linestyle=\"none\", marker=marker, color=dataset2name[dataset][\"color\"])\n",
    "        x_values = np.concatenate(x_values)\n",
    "        y_values = np.concatenate(y_values)\n",
    "        corr, p = pearsonr(x_values, y_values)\n",
    "        ax.set_title(f\"{metric.upper()}\")\n",
    "        # text box with correlation at upper left\n",
    "        p_text = \"p  < 0.00001\" if p < 0.00001 else f\"p = {p:.2e}\"\n",
    "        ax.text(0.05, 0.95, f\"Correlation: {corr:.2f}\\n({p_text})\", transform=ax.transAxes, fontsize=10, verticalalignment='top', bbox=dict(facecolor='white', alpha=0.5))\n",
    "        ax.set_xlabel(x)\n",
    "        ax.grid(True)\n",
    "\n",
    "    axes[0].set_ylabel(y)\n",
    "\n",
    "    # Gather handles and labels from all axes\n",
    "    custom_handles = []\n",
    "    for dataset in sorted_datasets:\n",
    "        custom_handles.append(\n",
    "            plt.Line2D([0], [0], color=dataset2name[dataset][\"color\"], linestyle=\"-\", linewidth=3, label=dataset2name[dataset][\"label\"])\n",
    "        ) \n",
    "    for model, model_name, marker in models:\n",
    "        custom_handles.append(\n",
    "            plt.Line2D([0], [0], color=\"black\", linestyle=\"none\", marker=marker, markersize=5, label=model_name)\n",
    "        )\n",
    "    fig.legend(handles=custom_handles, loc='upper right', bbox_to_anchor=(1.17, .95), title_fontsize=12, fontsize=10)\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    fig.savefig(f\"../outputs/results_paper/corr_allinone_{method}.png\")\n",
    "\n",
    "\n",
    "# metric = \"nce\"\n",
    "metric = \"ner\"\n",
    "# plot(metric, \"dp_calibration\", \"Relative Calibration Loss\")\n",
    "# plot_all_in_one(\"dp_calibration\", \"Relative Calibration Loss\")\n",
    "plot_all_in_one(\"dp_calibration\", \"Calibrated\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sst2. Min count: 196. Class: 0\n",
      "agnews. Min count: 92. Class: 0\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dbpedia. Min count: 35. Class: 11\n",
      "20newsgroups. Min count: 25. Class: 0\n",
      "banking77. Min count: 6. Class: 1\n"
     ]
    }
   ],
   "source": [
    "datset2test = {\n",
    "    \"sst2\": 400,\n",
    "    \"agnews\": 400,\n",
    "    \"dbpedia\": 700,\n",
    "    \"20newsgroups\": 800,\n",
    "    \"banking77\": 1000\n",
    "}\n",
    "\n",
    "for dataset, test_size in datset2test.items():\n",
    "    df = pd.read_csv(f\"../data/{dataset}/all.csv\", header=0, index_col=0)\n",
    "    test_list = np.loadtxt(f\"../lists/{dataset}/test_{test_size}.txt\", dtype=int)\n",
    "    labels = df.loc[test_list,\"label\"].values\n",
    "    counts = np.bincount(labels)\n",
    "    print(f\"{dataset}. Min count: {np.min(counts)}. Class: {np.argmin(counts)}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "llmcal",
   "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": "3.11.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}