experiments-boundary.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "2f323332-bd58-4bd6-b93b-9d3dfe36b4eb",
   "metadata": {},
   "source": [
    "# Experiments: Standard Scenario"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "b9a4ed81-d530-4026-a5a9-7318c2690d06",
   "metadata": {},
   "source": [
    "## Utils: \n",
    "\n",
    "- Density based clustering\n",
    "- Nearest-cluster centroids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a2a42f29-5322-48cb-b3dd-e7bd8c33b701",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import NearestNeighbors\n",
    "from sklearn.neighbors import NearestCentroid\n",
    "from sklearn.cluster import DBSCAN\n",
    "from sklearn import metrics\n",
    "import numpy as np\n",
    "from torchvision import transforms\n",
    "import torch\n",
    "import clip\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle\n",
    "from PIL import Image\n",
    "from rich import print\n",
    "from scipy.stats import trim_mean\n",
    "\n",
    "\n",
    "def find_avg_centroid(x):\n",
    "    return np.mean(x, axis=0)\n",
    "\n",
    "def trimmed_mean_centroid(x, proportion_to_cut=0.1):\n",
    "    \"\"\"\n",
    "    Compute the trimmed mean centroid of an array of points.\n",
    "    \"\"\"\n",
    "    return trim_mean(x, proportiontocut=proportion_to_cut, axis=0)\n",
    "\n",
    "def find_better_centroid(x):\n",
    "    return trimmed_mean_centroid(x)\n",
    "\n",
    "def find_nearest_centroids(x, y):\n",
    "    if not isinstance(x, (np.ndarray, np.generic)):\n",
    "        x = x.detach().cpu().numpy()\n",
    "        \n",
    "    clf = NearestCentroid(metric='manhattan')\n",
    "    clf = clf.fit(x, y)\n",
    "    centroids = clf.centroids_\n",
    "    return centroids\n",
    "\n",
    "def find_knn(x, centroids, n_neighbors=10):\n",
    "    if not isinstance(x, (np.ndarray, np.generic)):\n",
    "        x = x.detach().cpu().numpy()\n",
    "\n",
    "    knn_search = NearestNeighbors(n_neighbors=n_neighbors,\n",
    "                            metric='manhattan', \n",
    "                            algorithm='auto', #'ball_tree'\n",
    "                            n_jobs=-1)  \n",
    "    knn_search.fit(x)\n",
    "\n",
    "    indexes = knn_search.kneighbors(centroids, return_distance=False)\n",
    "    return x[indexes][0]\n",
    "\n",
    "        \n",
    "def get_clusters(x):\n",
    "    db = DBSCAN(eps=0.3, min_samples=10, leaf_size=30, \n",
    "                n_jobs=-1, metric='cosine').fit(x)\n",
    "\n",
    "    # core_samples_mask = np.zeros_like(db.labels_, dtype=bool)\n",
    "    # core_samples_mask[db.core_sample_indices_] = True\n",
    "    labels = db.labels_\n",
    "    # Number of clusters in labels, ignoring noise if present.\n",
    "    n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)\n",
    "    n_noise_ = list(labels).count(-1)\n",
    "\n",
    "    print(\"Estimated number of clusters: %d\" % n_clusters_)\n",
    "    print(\"Estimated number of noise points: %d\" % n_noise_)\n",
    "    if n_clusters_ == 0:\n",
    "        return x, labels, n_clusters_\n",
    "    # Remove outliers to avoid affecting the centroids\n",
    "    x = x[np.where(labels != -1)]\n",
    "    labels = labels[np.where(labels != -1)]\n",
    "    \n",
    "    return x, labels, n_clusters_\n",
    "\n",
    "\n",
    "def get_nearest_embedding_centroids(x, n_neighbors=10):\n",
    "    \n",
    "    #1. Get the CLIP embeddings\n",
    "    if not isinstance(x, (np.ndarray, np.generic)):\n",
    "        x = x.detach().cpu().numpy()\n",
    "    \n",
    "    #2. Cluster the embeddings\n",
    "    x, labels, n_clusters = get_clusters(x)\n",
    "\n",
    "    #3. Find centroids of embeddings\n",
    "    if n_clusters < 2:\n",
    "        # centroids = find_avg_centroid(x)\n",
    "        centroids = find_better_centroid(x)\n",
    "        centroids = centroids.reshape(1, -1)\n",
    "    else:\n",
    "        centroids = find_nearest_centroids(x, labels)\n",
    "\n",
    "    # If some infinite/nan values appear\n",
    "    centroids = np.nan_to_num(centroids)\n",
    "    #4. Find knn of centroids        \n",
    "    x_nearest_centroids = find_knn(x, centroids, n_neighbors)\n",
    "    x_chosen = np.vstack([centroids, x_nearest_centroids])\n",
    "\n",
    "    print(f\"# nearest centroids ({round((x_chosen.shape[0]/x.shape[0])*100, 2)}%):\", x_chosen.shape)\n",
    "    print()\n",
    "\n",
    "    return x_chosen\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0dde7f1a-9b47-48bb-b559-da2c2ff78380",
   "metadata": {},
   "source": [
    "### Semantic embedding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8b98b290-c1c8-49ed-9b5e-a451c703d7fc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Device: cpu\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Device: cpu\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    "print(f\"Device: {device}\")\n",
    "embeder, preprocess = clip.load(\"ViT-B/32\", device=device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "069d7442-3e6e-4db5-9a81-e8dd59a9c1f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "from torch.utils.data import DataLoader\n",
    "from sentence_transformers import SentenceTransformer, util\n",
    "from functools import lru_cache\n",
    "\n",
    "@lru_cache(maxsize=None)\n",
    "def get_embeddings(dataset):\n",
    "    \n",
    "    all_features = []\n",
    "    all_labels = []\n",
    "    all_tasks = []\n",
    "    \n",
    "    with torch.no_grad():\n",
    "        for images, labels, tasks in tqdm(DataLoader(dataset, batch_size=100, num_workers=8)):\n",
    "            features = embeder.encode_image(images.to(device))\n",
    "            # features = embeder_resnet(images.to(device)).squeeze()\n",
    "\n",
    "            all_features.append(features)\n",
    "            all_labels.append(labels)\n",
    "            all_tasks.append(tasks)\n",
    "\n",
    "    return (torch.cat(all_features).cpu().numpy(), \n",
    "            torch.cat(all_labels).cpu().numpy(), torch.cat(all_tasks).cpu().numpy())\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0df5330d-53e2-47b5-88cb-3c09ed63bf09",
   "metadata": {},
   "source": [
    "## Obtain the SODA10M dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "25d16b2d-f317-4554-ba2d-0489602eda15",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] No Detectron installation found, continuing without.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1;36m6\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5157</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1154</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6742</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2560</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4517</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2119</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m[\u001b[0m\u001b[1;36m5157\u001b[0m, \u001b[1;36m1154\u001b[0m, \u001b[1;36m6742\u001b[0m, \u001b[1;36m2560\u001b[0m, \u001b[1;36m4517\u001b[0m, \u001b[1;36m2119\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import sys\n",
    "sys.path.append('./CLAD/')\n",
    "import clad\n",
    "import torchvision\n",
    "\n",
    "\n",
    "root = \"./data\"\n",
    "original_transform = torchvision.transforms.Compose([\n",
    "    torchvision.transforms.ToTensor(),\n",
    "    torchvision.transforms.Normalize((0.3252, 0.3283, 0.3407), (0.0265, 0.0241, 0.0252))\n",
    "\n",
    "])\n",
    "\n",
    "# all_train_sets = clad.get_cladc_train(root, transform=lambda x: x)\n",
    "all_train_sets = clad.get_cladc_train(root, transform=original_transform)\n",
    "# val_sets = clad.get_cladc_val(root, transform=torchvision.transforms.ToTensor())\n",
    "print(len(all_train_sets))\n",
    "print([len(ts) for ts in all_train_sets])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "213d2cbb-668c-4705-a616-49c94403b8f4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1;36m6\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5157</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1154</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6742</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2560</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4517</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2119</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m[\u001b[0m\u001b[1;36m5157\u001b[0m, \u001b[1;36m1154\u001b[0m, \u001b[1;36m6742\u001b[0m, \u001b[1;36m2560\u001b[0m, \u001b[1;36m4517\u001b[0m, \u001b[1;36m2119\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# method_transform = torchvision.transforms.Compose([\n",
    "#     torchvision.transforms.ToTensor(),\n",
    "#     torchvision.transforms.ToPILImage()\n",
    "# ])\n",
    "method_transform = torchvision.transforms.Compose([\n",
    "    preprocess\n",
    "])\n",
    "all_train_sets_method = clad.get_cladc_train(root, transform=method_transform)\n",
    "print(len(all_train_sets_method))\n",
    "print([len(ts) for ts in all_train_sets_method])"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "20ce93c0-ca36-4408-a3c0-e0dd377aa286",
   "metadata": {},
   "source": [
    "### Train/Test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b09d88eb-a438-4e71-bda4-752964149767",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Original data\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Original data\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Training lenghts: \n",
       "<span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4125</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">923</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5393</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2048</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3613</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1695</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Training lenghts: \n",
       "\u001b[1m[\u001b[0m\u001b[1;36m4125\u001b[0m, \u001b[1;36m923\u001b[0m, \u001b[1;36m5393\u001b[0m, \u001b[1;36m2048\u001b[0m, \u001b[1;36m3613\u001b[0m, \u001b[1;36m1695\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Testing lenghts: \n",
       "<span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1032</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">231</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1349</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">512</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">904</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">424</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Testing lenghts: \n",
       "\u001b[1m[\u001b[0m\u001b[1;36m1032\u001b[0m, \u001b[1;36m231\u001b[0m, \u001b[1;36m1349\u001b[0m, \u001b[1;36m512\u001b[0m, \u001b[1;36m904\u001b[0m, \u001b[1;36m424\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Method data\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Method data\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Training lenghts: \n",
       "<span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4125</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">923</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5393</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2048</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3613</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1695</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Training lenghts: \n",
       "\u001b[1m[\u001b[0m\u001b[1;36m4125\u001b[0m, \u001b[1;36m923\u001b[0m, \u001b[1;36m5393\u001b[0m, \u001b[1;36m2048\u001b[0m, \u001b[1;36m3613\u001b[0m, \u001b[1;36m1695\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Testing lenghts: \n",
       "<span style=\"font-weight: bold\">[</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1032</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">231</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1349</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">512</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">904</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">424</span><span style=\"font-weight: bold\">]</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Testing lenghts: \n",
       "\u001b[1m[\u001b[0m\u001b[1;36m1032\u001b[0m, \u001b[1;36m231\u001b[0m, \u001b[1;36m1349\u001b[0m, \u001b[1;36m512\u001b[0m, \u001b[1;36m904\u001b[0m, \u001b[1;36m424\u001b[0m\u001b[1m]\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import torch\n",
    "from torch.utils.data import random_split\n",
    "\n",
    "# Define the split sizes for each dataset\n",
    "train_size = 0.8\n",
    "# test_size = 0.2\n",
    "\n",
    "# Split each dataset in the train_sets list into training and testing sets\n",
    "train_sets = []\n",
    "test_sets = []\n",
    "for dataset in all_train_sets:\n",
    "    dataset_size = len(dataset)\n",
    "    train_dataset_size = int(train_size * dataset_size)\n",
    "    test_dataset_size = dataset_size - train_dataset_size\n",
    "    train_dataset, test_dataset = random_split(dataset, [train_dataset_size, test_dataset_size], \n",
    "                                               generator=torch.Generator().manual_seed(42))\n",
    "    train_sets.append(train_dataset)\n",
    "    test_sets.append(test_dataset)\n",
    "\n",
    "print(\"Original data\")\n",
    "print('Training lenghts: ', [len(ts) for ts in train_sets])\n",
    "print('Testing lenghts: ', [len(ts) for ts in test_sets])\n",
    "\n",
    "train_sets_method = []\n",
    "test_sets_method = []\n",
    "for dataset in all_train_sets_method:\n",
    "    dataset_size = len(dataset)\n",
    "    train_dataset_size = int(train_size * dataset_size)\n",
    "    test_dataset_size = dataset_size - train_dataset_size\n",
    "    train_dataset, test_dataset = random_split(dataset, [train_dataset_size, test_dataset_size], \n",
    "                                               generator=torch.Generator().manual_seed(42))\n",
    "    train_sets_method.append(train_dataset)\n",
    "    test_sets_method.append(test_dataset)\n",
    "\n",
    "print(\"Method data\")\n",
    "print('Training lenghts: ', [len(ts) for ts in train_sets_method])\n",
    "print('Testing lenghts: ', [len(ts) for ts in test_sets_method])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "4274a3a2-e586-4c14-b18a-a98020046a37",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 615 ms, sys: 124 ms, total: 740 ms\n",
      "Wall time: 777 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "from torch.utils.data.dataset import TensorDataset\n",
    "from avalanche.benchmarks.utils import AvalancheDataset\n",
    "from avalanche.benchmarks.utils import make_classification_dataset\n",
    "from torch.utils.data import DataLoader\n",
    "from avalanche.benchmarks.generators import filelist_benchmark, dataset_benchmark\n",
    "\n",
    "\n",
    "\n",
    "# Original benchmark with tasks\n",
    "training_datasets = list()\n",
    "testing_datasets = list()\n",
    "\n",
    "for task, (train_s, test_s) in enumerate(zip(train_sets, test_sets), start=0):\n",
    "    training_datasets.append(make_classification_dataset(train_s, task_labels=np.repeat(task, len(train_s))))\n",
    "    testing_datasets.append(make_classification_dataset(test_s, task_labels=np.repeat(task, len(test_s))))\n",
    "\n",
    "benchmark= dataset_benchmark(\n",
    "    training_datasets,\n",
    "    testing_datasets\n",
    ")\n",
    "\n",
    "# Original benchmark without tasks\n",
    "training_datasets = list()\n",
    "testing_datasets = list()\n",
    "for task, (train_s, test_s) in enumerate(zip(train_sets, test_sets), start=0):\n",
    "    training_datasets.append(make_classification_dataset(train_s, task_labels=np.repeat(0, len(train_s))))\n",
    "    testing_datasets.append(make_classification_dataset(test_s, task_labels=np.repeat(0, len(test_s))))\n",
    "\n",
    "benchmark_no_task= dataset_benchmark(\n",
    "    training_datasets,\n",
    "    testing_datasets\n",
    ")\n",
    "\n",
    "# Method benchmark\n",
    "training_datasets = list()\n",
    "testing_datasets = list()\n",
    "for task, (train_s, test_s) in enumerate(zip(train_sets_method, test_sets_method), start=0):\n",
    "    training_datasets.append(make_classification_dataset(train_s, task_labels=np.repeat(task, len(train_s))))\n",
    "    testing_datasets.append(make_classification_dataset(test_s, task_labels=np.repeat(task, len(test_s))))\n",
    "\n",
    "benchmark_method= dataset_benchmark(\n",
    "    training_datasets,\n",
    "    testing_datasets\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "d637684b-eafb-460f-afd9-03b66fe41332",
   "metadata": {},
   "source": [
    "## Test a simple Task classifier and save in cache"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8a431a5a-8d1e-4a53-ac36-bd7be795704b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Getting embeddings\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Getting embeddings\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 42/42 [02:01<00:00,  2.89s/it]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">x_train shape: \n",
       "<span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4125</span>, <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">512</span><span style=\"font-weight: bold\">)</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "x_train shape: \n",
       "\u001b[1m(\u001b[0m\u001b[1;36m4125\u001b[0m, \u001b[1;36m512\u001b[0m\u001b[1m)\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/11 [00:00<?, ?it/s]"
     ]
    }
   ],
   "source": [
    "train_stream = benchmark_method.train_stream\n",
    "test_stream = benchmark_method.test_stream\n",
    "\n",
    "all_x_train_emb = list()\n",
    "all_y_train = list()\n",
    "all_t_train = list()\n",
    "\n",
    "all_x_test_emb = list()\n",
    "all_y_test = list()\n",
    "all_t_test = list()\n",
    "\n",
    "\n",
    "for task, experience in enumerate(train_stream):\n",
    "    print(\"Getting embeddings\")\n",
    "    x_train_emb, y_train, t_train = get_embeddings(train_stream[task].dataset)\n",
    "\n",
    "    print(\"x_train shape: \", x_train_emb.shape)\n",
    "    all_x_train_emb.append(x_train_emb)\n",
    "    all_y_train.append(y_train)\n",
    "    all_t_train.append(t_train)\n",
    "\n",
    "    x_test_emb, y_test, t_test = get_embeddings(test_stream[task].dataset)\n",
    "    print(\"x_test shape: \", x_test_emb.shape)\n",
    "    all_x_test_emb.append(x_test_emb)\n",
    "    all_y_test.append(y_test)\n",
    "    all_t_test.append(t_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "191177bb-dc70-4566-aeb0-79a801a0a8a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "print(\"Training data knn\")\n",
    "x_train_knn_total = [get_nearest_embedding_centroids(task, n_neighbors=10) for task in all_x_train_emb]\n",
    "print(\"x_train_knn shape: \", [e.shape[0] for e in x_train_knn_total])\n",
    "y_train_knn_total = [all_t_train[i][:x_train_knn_total[i].shape[0]] for i in range(len(all_t_train))]\n",
    "x_train_knn_total = np.vstack(x_train_knn_total)\n",
    "y_train_knn_total = np.hstack(y_train_knn_total)\n",
    "print(x_train_knn_total.shape)\n",
    "print(\"y_train_knn_total \", y_train_knn_total.shape)\n",
    "\n",
    "print()\n",
    "print(\"Testing\")\n",
    "print(\"x_test shape: \", [e.shape[0] for e in all_x_test_emb])\n",
    "x_test_total = np.vstack(all_x_test_emb)\n",
    "y_test_total = np.hstack(all_t_test)\n",
    "print(\"x_test_total \", x_test_total.shape)\n",
    "print(\"y_test_total \", y_test_total.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3507a494-7398-4710-8bca-23874746a2f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# from sklearn.neighbors import NearestCentroid\n",
    "# from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# def scale_data(x_train, x_test):\n",
    "#     scaler = StandardScaler()\n",
    "#     scaler.fit(x_train)\n",
    "#     x_train_scaled = scaler.transform(x_train)\n",
    "#     x_test_scaled = scaler.transform(x_test)\n",
    "#     return x_train_scaled, x_test_scaled\n",
    "\n",
    "# x_train_scaled, x_test_scaled = scale_data(x_train_knn_total, x_test_total)\n",
    "\n",
    "clf = NearestCentroid(metric='manhattan')\n",
    "clf = clf.fit(x_train_knn_total, y_train_knn_total)\n",
    "print(f\"score training knn: {clf.score(x_train_knn_total, y_train_knn_total)}\")\n",
    "print(f\"score testing: {clf.score(x_test_total, y_test_total)}\")\n",
    "# clf = clf.fit(x_train_scaled, y_train_knn_total)\n",
    "# print(f\"score training knn: {clf.score(x_train_scaled, y_train_knn_total)}\")\n",
    "# print(f\"score testing: {clf.score(x_test_scaled, y_test_total)}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "35b830db-c50e-4b83-b4f7-07d7a2c5ce54",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import ConfusionMatrixDisplay, confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "# plot_confusion_matrix(clf, x_val_total, y_val_total)  \n",
    "predictions = clf.predict(x_test_total)\n",
    "cm = confusion_matrix(y_test_total, predictions, labels=clf.classes_)\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm,\n",
    "                             display_labels=clf.classes_)\n",
    "\n",
    "disp.plot()\n",
    "plt.title(f\"SODA(n_tasks=6) | Accuracy: {round(clf.score(x_test_total, y_test_total), 2)}%\")\n",
    "plt.show();"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "592fbc86-dce1-40ac-8821-86bf68f95800",
   "metadata": {},
   "source": [
    "## Utils for the incremental task classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ed74f54e-435a-45e8-a69c-be43ad56a64a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import statistics as st\n",
    "\n",
    "\n",
    "def run_method(benchmark_method, test_stream, new_training, new_testing, task=0):\n",
    "    print(\"Getting embeddings\")\n",
    "    x_train_emb, y_train, t_train = get_embeddings(benchmark_method.train_stream[task].dataset)\n",
    "\n",
    "\n",
    "    print(\"x_train shape: \", x_train_emb.shape)\n",
    "    all_x_train_emb.append(x_train_emb)\n",
    "    all_y_train.append(y_train)\n",
    "    all_t_train.append(t_train)\n",
    "\n",
    "    # Training incremental domain classifier\n",
    "    print(\"Training Domain Classifier model\")\n",
    "    x_train_knn_total = [get_nearest_embedding_centroids(_x_train, n_neighbors=10) for _x_train in all_x_train_emb]\n",
    "    print(\"x_train_knn shape: \", [e.shape[0] for e in x_train_knn_total])\n",
    "    y_train_knn_total = [all_t_train[i][:x_train_knn_total[i].shape[0]] for i in range(len(all_t_train))]\n",
    "    x_train_knn_total = np.vstack(x_train_knn_total)\n",
    "    y_train_knn_total = np.hstack(y_train_knn_total)\n",
    "    print(\"x_train_knn_total shape: \", x_train_knn_total.shape)\n",
    "    print(\"y_train_knn_total shape\", y_train_knn_total.shape)\n",
    "    if len(set(y_train_knn_total)) > 1:\n",
    "        clf = NearestCentroid(metric='manhattan')\n",
    "        clf = clf.fit(x_train_knn_total, y_train_knn_total)\n",
    "        \n",
    "    print(\"y_train_knn_total\", np.random.choice(y_train_knn_total, 10))\n",
    "    \n",
    "    # Testing\n",
    "    # print(\"Testing\")\n",
    "    x_test_emb, y_test, t_test = get_embeddings(benchmark_method.test_stream[task].dataset)\n",
    "\n",
    "    if len(set(y_train_knn_total)) > 1:\n",
    "        print(\"Using the classifier:\")\n",
    "        task_labels = clf.predict(x_test_emb)\n",
    "        print(\"Count predictions: \", dict(zip(*np.unique(task_labels, return_counts=True))))\n",
    "        print(\"Using the mode to get the most voting prediction:\")\n",
    "        mode = st.mode(task_labels)\n",
    "        task_labels = np.repeat(mode, len(task_labels))\n",
    "    else:\n",
    "        print(\"Fist scenario: Task ID=\", task)\n",
    "        task_labels = np.repeat(task, len(benchmark_method.test_stream[task].dataset))\n",
    "    print(f\"Predictions: {np.random.choice(task_labels, 10)}\")\n",
    "    print(f\"Predictions shape: {task_labels.shape}\")\n",
    "\n",
    "    new_training.append(benchmark_method.train_stream[task].dataset)\n",
    "    new_testing.append(make_classification_dataset(test_stream[task].dataset, task_labels=task_labels))\n",
    "\n",
    "    benchmark_custom = dataset_benchmark(\n",
    "        new_training,\n",
    "        new_testing\n",
    "    )\n",
    "\n",
    "    return benchmark_custom"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "df868394-0d48-494c-89e1-4c740e0916bf",
   "metadata": {},
   "source": [
    "## Prepare the strategies to run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4552878-a54d-4ccd-b254-87b4e63b4cb0",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def get_strategy(strategy_name, model, optimizer, criterion, eval_plugin, n_epochs=6):\n",
    "    strategies = {\n",
    "        'EWC': EWC( #Regularization-based method\n",
    "            model=model,\n",
    "            optimizer=optimizer,\n",
    "            criterion=criterion,\n",
    "            train_mb_size=200,\n",
    "            train_epochs=n_epochs,\n",
    "            eval_mb_size=200,\n",
    "            device=device,\n",
    "            evaluator=eval_plugin,\n",
    "            ewc_lambda=0.2,\n",
    "        ),\n",
    "        'Replay': Replay( #Rehearsal-based method\n",
    "            model=model,\n",
    "            optimizer=optimizer,\n",
    "            criterion=criterion,\n",
    "            train_mb_size=200,\n",
    "            train_epochs=n_epochs,\n",
    "            eval_mb_size=200,\n",
    "            device=device,\n",
    "            evaluator=eval_plugin,\n",
    "            mem_size=200,\n",
    "        ),\n",
    "        'LwF': LwF( #Architecture-based method\n",
    "            model,\n",
    "            optimizer,\n",
    "            criterion,\n",
    "            alpha=0.5,\n",
    "            temperature=0.2,\n",
    "            train_epochs=n_epochs,\n",
    "            device=device,\n",
    "            train_mb_size=200,\n",
    "            eval_mb_size=200,\n",
    "            evaluator=eval_plugin,\n",
    "        ),\n",
    "    }\n",
    "\n",
    "    return strategies[strategy_name]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "47899f7f-b1e8-4638-8d40-bdabc5b3e54a",
   "metadata": {},
   "source": [
    "## Prepare and run the experiments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c56ee3ff-2371-4994-a2d2-d3d0d317be0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "root_exp_name = 'Task_boundary'\n",
    "# root_exp_name = 'Agnostic_with_repetition'\n",
    "# sequence = [0, 1, 2, 1, 3, 3, 4, 5]\n",
    "# sequence = [0, 1, 2, 3, 4, 5]\n",
    "n_epochs = 6\n",
    "random_seeds = [7]\n",
    "# random_seeds = [51, 12, 42, 7]\n",
    "num_experiments = len(random_seeds)\n",
    "print(f\"n_exps = {num_experiments}\")\n",
    "print(f\"Setup = {root_exp_name}\")\n",
    "print(f\"Epochs = {n_epochs}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5b2f3998-9d12-46f7-a2e1-5067d21cecc4",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import random\n",
    "import gc\n",
    "from torch.nn import CrossEntropyLoss\n",
    "import copy\n",
    "import statistics as st\n",
    "from torch.optim import Adam\n",
    "from rich import print\n",
    "from avalanche.models import as_multitask\n",
    "from avalanche.benchmarks.classic import SplitMNIST, PermutedMNIST\n",
    "from avalanche.models import MTSimpleMLP, MTSimpleCNN, SimpleCNN, SimpleMLP\n",
    "# from avalanche.training.supervised import EWC, Naive, LFL\n",
    "from avalanche.evaluation.metrics import (\n",
    "    forgetting_metrics,\n",
    "    accuracy_metrics,\n",
    ")\n",
    "from avalanche.training.plugins import EvaluationPlugin\n",
    "from avalanche.logging import InteractiveLogger, TextLogger, TensorboardLogger\n",
    "# from avalanche.training import Naive\n",
    "from torch.nn import CrossEntropyLoss\n",
    "from torch.optim import SGD\n",
    "from avalanche.benchmarks.classic import SplitMNIST, PermutedMNIST, RotatedMNIST, SplitCIFAR10, SplitCIFAR100\n",
    "from avalanche.benchmarks.generators import filelist_benchmark, dataset_benchmark\n",
    "from avalanche.training.utils import adapt_classification_layer\n",
    "\n",
    "from datetime import datetime\n",
    "from avalanche.evaluation.metrics.images_samples import images_samples_metrics\n",
    "from avalanche.evaluation.metrics.labels_repartition import (\n",
    "    labels_repartition_metrics,\n",
    ")\n",
    "from avalanche.evaluation.metrics.mean_scores import mean_scores_metrics\n",
    "from torch.nn import Linear\n",
    "from avalanche.training.supervised import Replay, GEM, LwF, EWC\n",
    "\n",
    "# choose some metrics and evaluation method\n",
    "# loggers = []\n",
    "\n",
    "# loggers.append(InteractiveLogger())\n",
    "\n",
    "# eval_plugin = EvaluationPlugin(\n",
    "#     accuracy_metrics(\n",
    "#         minibatch=False, epoch=True, experience=True\n",
    "#     ),\n",
    "\n",
    "#     forgetting_metrics(experience=True),\n",
    "#     loggers=loggers,\n",
    "# )\n",
    "    \n",
    "# num_experiments = 2\n",
    "# # num_experiments = 3\n",
    "# random_seeds = [42, 123, 456]\n",
    "\n",
    "# results_list = []\n",
    "for exp in range(num_experiments):\n",
    "    \n",
    "    train_stream = benchmark.train_stream\n",
    "    test_stream = benchmark.test_stream\n",
    "\n",
    "    print(f\"Length training: {len(benchmark.train_stream[0].dataset)}\")\n",
    "    print(f\"Length testing: {len(benchmark.test_stream[0].dataset)}\")\n",
    "\n",
    "    print(\"-\"*50)\n",
    "\n",
    "\n",
    "    n_workers = 8\n",
    "\n",
    "    seed = random_seeds[exp]\n",
    "    random.seed(seed)\n",
    "    torch.manual_seed(seed)\n",
    "    print(f\"Setup = {root_exp_name}\")\n",
    "    print(f\"Experiment: {exp+1} / {num_experiments}\")\n",
    "    print(f\"Seed = {seed}\")\n",
    "    # torch.cuda.manual_seed(seed)\n",
    "    all_strategies = ['EWC', 'Replay', 'LwF']\n",
    "    \n",
    "    for strategy_name in all_strategies:\n",
    "        loggers = []\n",
    "\n",
    "        loggers.append(InteractiveLogger())\n",
    "\n",
    "        eval_plugin = EvaluationPlugin(\n",
    "            accuracy_metrics(\n",
    "                minibatch=False, epoch=False, experience=True\n",
    "            ),\n",
    "\n",
    "            # forgetting_metrics(experience=True),\n",
    "            loggers=loggers,\n",
    "        )\n",
    "\n",
    "        model = torchvision.models.resnet18(weights=True)\n",
    "        # model.fc = Linear(model.fc.in_features, 7, bias=True)\n",
    "        model.fc = Linear(model.fc.in_features, 7, bias=True)\n",
    "        mt_model = as_multitask(model, 'fc')\n",
    "\n",
    "        optimizer = Adam(mt_model.parameters(), lr=0.01)\n",
    "        criterion = CrossEntropyLoss()\n",
    "    \n",
    "        strategy = get_strategy(strategy_name, mt_model, optimizer, criterion, eval_plugin, n_epochs)\n",
    "        print(f\"Running strategy: {strategy}, experiment: {exp}\")\n",
    "        results = {key: [] for key in ['ground_truth', 'TADIL', 'normal']}\n",
    "        \n",
    "        all_x_train_emb = list()\n",
    "        all_y_train = list()\n",
    "        all_t_train = list()\n",
    "\n",
    "        all_x_test_emb = list()\n",
    "        all_y_test = list()\n",
    "        all_t_test = list()        \n",
    "        new_training = list()\n",
    "        new_testing = list()\n",
    "\n",
    "        for task, experience in enumerate(train_stream):\n",
    "            curr_experience = experience.current_experience\n",
    "            print(\"Current experience: \", curr_experience)    \n",
    "            print(\"Experience task ID \", experience.task_label)\n",
    "            print('Experience shape:', len(experience.dataset))\n",
    "\n",
    "\n",
    "            # Training\n",
    "            print(\"Training multi-head model\")\n",
    "            strategy.train(experience, num_workers=n_workers)\n",
    "            print('Training multi-head model completed')\n",
    "\n",
    "            # Run the method\n",
    "            benchmark_custom = run_method(benchmark_method, test_stream, new_training, new_testing, task=task)\n",
    "\n",
    "            # Testing\n",
    "            print(\"Evaluation benchmark: Ground Truth\")\n",
    "            results['ground_truth'].append(strategy.eval(benchmark.test_stream[:task+1], num_workers=n_workers))\n",
    "\n",
    "            print(\"Evaluation benchmark: TADIL\")\n",
    "            results['TADIL'].append(strategy.eval(benchmark_custom.test_stream, num_workers=n_workers))\n",
    "\n",
    "            print(\"Evaluation benchmark: Normal\")\n",
    "            results['normal'].append(strategy.eval(benchmark_no_task.test_stream[:task+1], num_workers=n_workers))\n",
    "\n",
    "            print(\"-\"*40)\n",
    "            print()\n",
    "\n",
    "        \n",
    "        benchmark_custom = None\n",
    "            \n",
    "        with open(f'{root_exp_name}_{strategy_name}_{seed}.pickle', 'wb') as handle:\n",
    "            pickle.dump(results, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
    "            print(f\"Saved results for {root_exp_name}_{strategy_name}, seed: {seed}\")\n",
    "\n",
    "\n",
    "        print(\"-\"*40)\n",
    "        print()\n",
    "    print(\"-\"*80)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "0bd28139-56e2-4a98-b53f-888833740a34",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Plot the comparison using model's accuracies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1117baad-9896-4275-806d-ace9b10507a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "strategy_name = \"EWC\"\n",
    "n_exp = 0\n",
    "# strategy_name = \"LwF\"\n",
    "with open(f'{strategy_name}_{n_exp}.pickle', 'rb') as handle:\n",
    "    results = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "3bd89072-a743-493e-89db-598fadd1a1be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['ground_truth', 'TADIL', 'normal'])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "76469831-b6c5-482d-a038-3f8d0e2bc306",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt\n",
    "\n",
    "# # Extract the eval_phase values\n",
    "# def get_eval_phase_values(data, keyword):\n",
    "#     eval_values = []\n",
    "#     for exp in data:\n",
    "#         eval_values.append({k: v for k, v in exp.items() if ('eval_phase/test_stream' in k) and (keyword in k)})\n",
    "#     return eval_values\n",
    "\n",
    "# results_clean = {k:get_eval_phase_values(results[k], \"Top1_Acc_Exp\") for k in results.keys()}"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "3de7a14e-bf91-411e-88c0-f00cb208098c",
   "metadata": {},
   "source": [
    "### Plot the numbers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "8c2e499c-caef-4143-9d4f-fa5a11c95730",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import pickle\n",
    "import numpy as np\n",
    "\n",
    "strategy_name = \"EWC\"\n",
    "n_exp = 1\n",
    "# strategy_name = \"LwF\"\n",
    "with open(f'{strategy_name}_{n_exp}.pickle', 'rb') as handle:\n",
    "    results = pickle.load(handle)\n",
    "\n",
    "\n",
    "metric_root_name = \"Top1_Acc_Exp/eval_phase/test_stream\"\n",
    "\n",
    "# Create a dictionary to store the scores for each method\n",
    "method_scores = {\n",
    "    'ground_truth': [],\n",
    "    'TADIL': [],\n",
    "    'normal': [],\n",
    "}\n",
    "\n",
    "for domain in range(len(results['ground_truth'])):\n",
    "    scores_task = []\n",
    "    scores_method = []\n",
    "    scores_third_method = [] # Initialize an empty list for the scores of the third method\n",
    "    iters_ = []\n",
    "    for index, r in enumerate(results['ground_truth']):\n",
    "        try:\n",
    "            score_task = r[f\"{metric_root_name}/Task00{domain}/Exp00{domain}\"]\n",
    "            score_method = results['TADIL'][index][f\"{metric_root_name}/Task00{domain}/Exp00{domain}\"]\n",
    "            score_third_method = results['normal'][index][f\"{metric_root_name}/Task00{0}/Exp00{domain}\"] # Get the score of the third method\n",
    "            if score_task != 0:\n",
    "                scores_task.append(score_task)\n",
    "                scores_method.append(score_method)\n",
    "                scores_third_method.append(score_third_method) # Append the score of the third method to the list\n",
    "                iters_.append(index + 1)\n",
    "        except:\n",
    "            pass\n",
    "    method_scores['ground_truth'].append((iters_, scores_task))\n",
    "    method_scores['TADIL'].append((iters_, scores_method))\n",
    "    method_scores['normal'].append((iters_, scores_third_method)) # Append the scores of the third method to the dictionary\n",
    "\n",
    "\n",
    "max_iter = max([max(scores, key=lambda s: s[0][-1])[0][-1] for scores in method_scores.values()])  # Get the maximum iteration index for all methods\n",
    "\n",
    "fig, axs = plt.subplots(1, 3, figsize=(15, 5)) # Create 3 subplots arranged horizontally\n",
    "plt.suptitle(f\"Strategy: {strategy_name}, Dataset: Soda10-class with {len(results['ground_truth'])} tasks\")\n",
    "\n",
    "for i, (method_name, method_scores_list) in enumerate(method_scores.items()):\n",
    "    \n",
    "    avg_list = list()\n",
    "    for j, (x, y) in enumerate(method_scores_list):\n",
    "        average_accuracy = np.mean(y)  # Calculate the average accuracy\n",
    "        avg_list.append(average_accuracy)\n",
    "        axs[i].plot(x, y, '*--', label=f'{method_name}-Task{j+1} (Avg Acc: {average_accuracy:.2f})')  # Include the average accuracy in the label\n",
    "        axs[i].set_yticks(np.arange(0, 1.1, 0.1))\n",
    "        axs[i].set_yticklabels([\"{:.2f}\".format(val) for val in np.arange(0, 1.1, 0.1)])\n",
    "        axs[i].legend()\n",
    "        \n",
    "    axs[i].set_title(f\"Approach: {method_name} | Avg Acc: {round(np.mean(avg_list), 2)}\")\n",
    "    \n",
    "    axs[i].set_xlabel(\"Number of Tasks\")\n",
    "    axs[i].set_ylabel(\"Performance\")\n",
    "    axs[i].set_xticks(np.arange(1, max_iter+1))\n",
    "\n",
    "plt.tight_layout() \n",
    "plt.savefig(f'{strategy_name}.pdf', bbox_inches='tight', pad_inches=0)\n",
    "plt.show()\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "eb5e2bb8-6619-4faa-a238-4edfd933378d",
   "metadata": {},
   "source": [
    "## Table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "a1fcfa48-a894-4df8-b71f-9b79913ec902",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">EWC\n",
       "</pre>\n"
      ],
      "text/plain": [
       "EWC\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">  Task ID Ground Truth        TADIL       Normal\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.08</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.08</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.07</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.07</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.56</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.89</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.89</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.68</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.73</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.73</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.71</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "  Task ID Ground Truth        TADIL       Normal\n",
       "\u001b[1;36m0\u001b[0m  Task \u001b[1;36m1\u001b[0m  \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m1\u001b[0m  Task \u001b[1;36m2\u001b[0m  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.08\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.08\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m2\u001b[0m  Task \u001b[1;36m3\u001b[0m  \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.07\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.07\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.56\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m3\u001b[0m  Task \u001b[1;36m4\u001b[0m   \u001b[1;36m0.89\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.89\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m4\u001b[0m  Task \u001b[1;36m5\u001b[0m  \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.68\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m5\u001b[0m  Task \u001b[1;36m6\u001b[0m   \u001b[1;36m0.73\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.73\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.71\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">LwF\n",
       "</pre>\n"
      ],
      "text/plain": [
       "LwF\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">  Task ID Ground Truth        TADIL       Normal\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.70</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.70</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.56</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.87</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.87</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.85</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.65</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.65</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.58</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "  Task ID Ground Truth        TADIL       Normal\n",
       "\u001b[1;36m0\u001b[0m  Task \u001b[1;36m1\u001b[0m  \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m1\u001b[0m  Task \u001b[1;36m2\u001b[0m  \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m2\u001b[0m  Task \u001b[1;36m3\u001b[0m  \u001b[1;36m0.70\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.70\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.56\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m3\u001b[0m  Task \u001b[1;36m4\u001b[0m  \u001b[1;36m0.87\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.87\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.85\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m4\u001b[0m  Task \u001b[1;36m5\u001b[0m   \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m5\u001b[0m  Task \u001b[1;36m6\u001b[0m   \u001b[1;36m0.65\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.65\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.58\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">Replay\n",
       "</pre>\n"
      ],
      "text/plain": [
       "Replay\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">  Task ID Ground Truth        TADIL       Normal\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.93</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.93</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.88</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.81</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.81</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.90</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.90</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.83</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.76</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span>\n",
       "<span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span>  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.80</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.80</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "  Task ID Ground Truth        TADIL       Normal\n",
       "\u001b[1;36m0\u001b[0m  Task \u001b[1;36m1\u001b[0m  \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m1\u001b[0m  Task \u001b[1;36m2\u001b[0m  \u001b[1;36m0.93\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.93\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.88\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m2\u001b[0m  Task \u001b[1;36m3\u001b[0m  \u001b[1;36m0.81\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.81\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m3\u001b[0m  Task \u001b[1;36m4\u001b[0m  \u001b[1;36m0.90\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.90\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.83\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m4\u001b[0m  Task \u001b[1;36m5\u001b[0m  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m  \u001b[1;36m0.76\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m\n",
       "\u001b[1;36m5\u001b[0m  Task \u001b[1;36m6\u001b[0m   \u001b[1;36m0.80\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.80\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m   \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df_tables = dict()\n",
    "strategy_names = ['EWC', 'LwF', 'Replay']\n",
    "for strategy_name in strategy_names:\n",
    "    # strategy_name = \"LwF\"\n",
    "    n_exp = 1\n",
    "    # strategy_name = \"LwF\"\n",
    "    with open(f'{strategy_name}_{n_exp}.pickle', 'rb') as handle:\n",
    "        results = pickle.load(handle)\n",
    "\n",
    "    method_scores = {\n",
    "        'ground_truth': [],\n",
    "        'TADIL': [],\n",
    "        'normal': [],\n",
    "    }\n",
    "\n",
    "    for domain in range(len(results['ground_truth'])):\n",
    "        scores_task = []\n",
    "        scores_method = []\n",
    "        scores_third_method = []\n",
    "        iters_ = []\n",
    "        for index, r in enumerate(results['ground_truth']):\n",
    "            try:\n",
    "                score_task = r[f\"{metric_root_name}/Task00{domain}/Exp00{domain}\"]\n",
    "                score_method = results['TADIL'][index][f\"{metric_root_name}/Task00{domain}/Exp00{domain}\"]\n",
    "                score_third_method = results['normal'][index][f\"{metric_root_name}/Task00{0}/Exp00{domain}\"] # Get the score of the third method\n",
    "                if score_task != 0:\n",
    "                    scores_task.append(score_task)\n",
    "                    scores_method.append(score_method)\n",
    "                    scores_third_method.append(score_third_method) # Append the score of the third method to the list\n",
    "                    iters_.append(index + 1)\n",
    "            except:\n",
    "                pass\n",
    "        method_scores['ground_truth'].append((iters_, scores_task))\n",
    "        method_scores['TADIL'].append((iters_, scores_method))\n",
    "        method_scores['normal'].append((iters_, scores_third_method)) # Append the scores of the third method to the dictionary\n",
    "\n",
    "\n",
    "    max_iter = max([max(scores, key=lambda s: s[0][-1])[0][-1] for scores in method_scores.values()])  # Get the maximum iteration index for all methods\n",
    "\n",
    "    avg_acc_df = pd.DataFrame(columns=method_scores.keys(), index=range(1, len(method_scores_list) + 1))\n",
    "\n",
    "\n",
    "    for method_name, method_scores_list in method_scores.items():\n",
    "        for j, (x, y) in enumerate(method_scores_list):\n",
    "            average_accuracy = np.mean(y)\n",
    "            std_dev = np.std(y)\n",
    "            avg_acc_df.at[j+1, method_name] = f\"{average_accuracy:.2f} ({round(std_dev, 2)})\"\n",
    "\n",
    "    # Display the DataFrame\n",
    "    avg_acc_df.index = [f\"Task {i+1}\" for i in range(len(method_scores_list))]\n",
    "    avg_acc_df = avg_acc_df.reset_index()\n",
    "    avg_acc_df = avg_acc_df.rename(columns={'index': 'Task ID', 'ground_truth': 'Ground Truth', 'normal': 'Normal'})\n",
    "    print(strategy_name)\n",
    "    print(avg_acc_df)\n",
    "    df_tables[strategy_name] = avg_acc_df"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7e25f87f-ba24-4d2f-8e45-b2242c4f0f94",
   "metadata": {},
   "source": [
    "## Table to latex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "11129d2d-9599-4626-843b-344b52416668",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-113-0ca883fb85e9>:9: FutureWarning: The frame.append method is deprecated and will be removed from pandas in a future version. Use pandas.concat instead.\n",
      "  final_df = final_df.append(df, ignore_index=True)\n",
      "<ipython-input-113-0ca883fb85e9>:9: FutureWarning: The frame.append method is deprecated and will be removed from pandas in a future version. Use pandas.concat instead.\n",
      "  final_df = final_df.append(df, ignore_index=True)\n",
      "<ipython-input-113-0ca883fb85e9>:9: FutureWarning: The frame.append method is deprecated and will be removed from pandas in a future version. Use pandas.concat instead.\n",
      "  final_df = final_df.append(df, ignore_index=True)\n",
      "<ipython-input-113-0ca883fb85e9>:16: FutureWarning: In future versions `DataFrame.to_latex` is expected to utilise the base implementation of `Styler.to_latex` for formatting and rendering. The arguments signature may therefore change. It is recommended instead to use `DataFrame.style.to_latex` which also contains additional functionality.\n",
      "  latex_table = final_df.to_latex(index=False, float_format=\"%.2f\", bold_rows=True, column_format=\"|l|l|c|c|c|\", multicolumn_format=\"c\")\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">\\begin<span style=\"font-weight: bold\">{</span>tabular<span style=\"font-weight: bold\">}{</span>|l|l|c|c|c|<span style=\"font-weight: bold\">}</span>\n",
       "\\hline\n",
       "Strategy &amp; Task ID &amp; Ground Truth &amp;       TADIL &amp;      Normal \\\\\n",
       "\\hline\n",
       "     EWC &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.59</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.08</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.08</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.07</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.69</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.07</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.56</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.04</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> &amp;   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.89</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.89</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.68</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span> &amp;   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.73</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.73</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.71</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "     LwF &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.62</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.05</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.86</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.70</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.70</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.56</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.87</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.87</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.85</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> &amp;   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.74</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span> &amp;   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.65</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.65</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.58</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "  Replay &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">1</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.03</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">2</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.93</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.93</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.88</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">3</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.81</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.81</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.75</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">4</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.90</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.90</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.83</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.02</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">5</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.82</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> &amp; <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.76</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.01</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "         &amp;  Task <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">6</span> &amp;   <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.80</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.80</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> &amp;  <span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.77</span> <span style=\"font-weight: bold\">(</span><span style=\"color: #008080; text-decoration-color: #008080; font-weight: bold\">0.0</span><span style=\"font-weight: bold\">)</span> \\\\\n",
       "\\hline\n",
       "\\end<span style=\"font-weight: bold\">{</span>tabular<span style=\"font-weight: bold\">}</span>\n",
       "\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\\begin\u001b[1m{\u001b[0mtabular\u001b[1m}\u001b[0m\u001b[1m{\u001b[0m|l|l|c|c|c|\u001b[1m}\u001b[0m\n",
       "\\hline\n",
       "Strategy & Task ID & Ground Truth &       TADIL &      Normal \\\\\n",
       "\\hline\n",
       "     EWC &  Task \u001b[1;36m1\u001b[0m &  \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.59\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m2\u001b[0m &  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.08\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.08\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m3\u001b[0m &  \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.07\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.69\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.07\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.56\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.04\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m4\u001b[0m &   \u001b[1;36m0.89\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.89\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m5\u001b[0m &  \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.68\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m6\u001b[0m &   \u001b[1;36m0.73\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.73\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.71\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "     LwF &  Task \u001b[1;36m1\u001b[0m &  \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.62\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.05\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m2\u001b[0m &  \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.86\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m3\u001b[0m &  \u001b[1;36m0.70\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.70\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.56\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m4\u001b[0m &  \u001b[1;36m0.87\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.87\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.85\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m5\u001b[0m &   \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.74\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m6\u001b[0m &   \u001b[1;36m0.65\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.65\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.58\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "  Replay &  Task \u001b[1;36m1\u001b[0m &  \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.03\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m2\u001b[0m &  \u001b[1;36m0.93\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.93\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.88\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m3\u001b[0m &  \u001b[1;36m0.81\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.81\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.75\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m4\u001b[0m &  \u001b[1;36m0.90\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.90\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.83\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.02\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m5\u001b[0m &  \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.82\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m & \u001b[1;36m0.76\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.01\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "         &  Task \u001b[1;36m6\u001b[0m &   \u001b[1;36m0.80\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.80\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m &  \u001b[1;36m0.77\u001b[0m \u001b[1m(\u001b[0m\u001b[1;36m0.0\u001b[0m\u001b[1m)\u001b[0m \\\\\n",
       "\\hline\n",
       "\\end\u001b[1m{\u001b[0mtabular\u001b[1m}\u001b[0m\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "\n",
    "final_df = pd.DataFrame()\n",
    "for key, df in df_tables.items():\n",
    "\n",
    "    df = df.reset_index(drop=True)\n",
    "    df.insert(0, \"Strategy\", key)\n",
    "    final_df = final_df.append(df, ignore_index=True)\n",
    "    \n",
    "unique_strategies = final_df[\"Strategy\"].unique()\n",
    "for i, strategy in enumerate(unique_strategies):\n",
    "    final_df.loc[(final_df[\"Strategy\"] == strategy) & (final_df.index != i * len(df)), \"Strategy\"] = \"\"\n",
    "\n",
    "\n",
    "latex_table = final_df.to_latex(index=False, float_format=\"%.2f\", bold_rows=True, column_format=\"|l|l|c|c|c|\", multicolumn_format=\"c\")\n",
    "\n",
    "\n",
    "# Add grid lines to the table\n",
    "latex_table = latex_table.replace('\\\\toprule', '\\\\hline')\n",
    "latex_table = latex_table.replace('\\\\midrule', '\\\\hline')\n",
    "latex_table = latex_table.replace('\\\\bottomrule', '\\\\hline')\n",
    "\n",
    "# Print the LaTeX table\n",
    "print(latex_table)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:pytorch] *",
   "language": "python",
   "name": "conda-env-pytorch-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}