notebooks/3.1-Prediction_statistics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predictions statistics\n",
    "\n",
    "Let's analyse the predictions made on our test datasplit.\n",
    "\n",
    "**Note** To run this notebook you previously have to run and save the predictions on a test dataset **with ground-truth labels**. See the Datasplit section in [3.0 notebook](./3.0-Computing_predictions.ipynb#Predicting-a-datasplit-txt-file) for reference.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "import packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import json\n",
    "from planktonclas import test_utils\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score,confusion_matrix,roc_auc_score\n",
    "from planktonclas.data_utils import load_image, load_class_names\n",
    "from planktonclas import paths, plot_utils\n",
    "from planktonclas.plot_utils import create_pred_path,plt_conf_matrix\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import datetime\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn\n",
    "from tqdm import tqdm\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\") # To ignore UndefinedMetricWarning: [Recall/Precision/F-Score] is ill-defined and being set to 0.0 in labels with no [true/predicted] samples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading class names...\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# User parameters to set\n",
    "TIMESTAMP = 'phytoplankton_vliz'           # timestamp of the model\n",
    "SPLIT_NAME = 'DS_split'                   # dataset split to predict\n",
    "MODEL_NAME = 'final_model.h5'         # model to use to make the mediction\n",
    "TOP_K = 5                             # number of top classes predictions to save\n",
    "\n",
    "# Set the timestamp\n",
    "paths.timestamp = TIMESTAMP\n",
    "\n",
    "# Load clas names\n",
    "class_names = load_class_names(splits_dir=paths.get_ts_splits_dir())\n",
    "\n",
    "# Load back the predictions\n",
    "pred_path = os.path.join(paths.get_predictions_dir(), '{}+{}+top{}.json'.format(MODEL_NAME, SPLIT_NAME, TOP_K))\n",
    "with open(pred_path) as f:\n",
    "    pred_dict = json.load(f)\n",
    "data = pd.DataFrame(pred_dict)\n",
    "data[['filenames', 'true_lab','pred_lab']].head(5)\n",
    "data['in_top5'] = data.apply(lambda row: row['true_lab'] in row['pred_lab'], axis=1)\n",
    "data['top1_correct'] = data.apply(lambda row: row['true_lab'] == row['pred_lab'][0], axis=1)\n",
    "data['true_label']=class_names[data['true_lab']]\n",
    "data['top1']=data.apply(lambda row: class_names[row['pred_lab'][0]], axis=1)\n",
    "data['probability']=data.apply(lambda row: row['pred_prob'][0], axis=1)\n",
    "flat_array=np.concatenate(np.array(pred_dict['pred_lab']))\n",
    "data['pred_lab'] = flat_array.reshape(-1, 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create Folder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Computing the metrics\n",
    "\n",
    "Check [sklearn](https://scikit-learn.org/stable/modules/classes.html#classification-metrics) for more classification metrics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top1 accuracy: 84.7 %\n",
      "Top5 accuracy: 98.3 %\n",
      "\n",
      "\n",
      "Micro recall: 84.7 %\n",
      "Macro recall: 76.9 %\n",
      "Macro recall (no labels): 76.9 %\n",
      "Weighted recall: 84.7 %\n",
      "\n",
      "\n",
      "Micro precision: 84.7 %\n",
      "Macro precision: 82.4 %\n",
      "Macro precision (no labels): 82.4 %\n",
      "Weighted precision: 85.3 %\n",
      "\n",
      "\n",
      "Micro F1 score: 84.7 %\n",
      "Macro F1 score: 79.0 %\n",
      "Macro F1 score (no labels): 79.0 %\n",
      "Weighted F1 score: 84.7 %\n"
     ]
    }
   ],
   "source": [
    "true_lab, pred_lab = np.array(pred_dict['true_lab']), np.array(pred_dict['pred_lab'])\n",
    "\n",
    "top1 = test_utils.topK_accuracy(true_lab, pred_lab, K=1)\n",
    "top5 = test_utils.topK_accuracy(true_lab, pred_lab, K=5)\n",
    "\n",
    "print('Top1 accuracy: {:.1f} %'.format(top1 * 100))\n",
    "print('Top5 accuracy: {:.1f} %'.format(top5 * 100))\n",
    "\n",
    "labels = range(len(class_names))\n",
    "print('\\n')\n",
    "print('Micro recall: {:.1f} %'.format(100 * recall_score(true_lab, pred_lab[:, 0], labels=labels, average='micro')))\n",
    "print('Macro recall: {:.1f} %'.format(100 * recall_score(true_lab, pred_lab[:, 0], labels=labels, average='macro')))\n",
    "print('Macro recall (no labels): {:.1f} %'.format(100 * recall_score(true_lab, pred_lab[:, 0], average='macro')))\n",
    "print('Weighted recall: {:.1f} %'.format(100 * recall_score(true_lab, pred_lab[:, 0], labels=labels, average='weighted')))\n",
    "\n",
    "print('\\n')\n",
    "print('Micro precision: {:.1f} %'.format(100 * precision_score(true_lab, pred_lab[:, 0], labels=labels, average='micro')))\n",
    "print('Macro precision: {:.1f} %'.format(100 * precision_score(true_lab, pred_lab[:, 0], labels=labels, average='macro')))\n",
    "print('Macro precision (no labels): {:.1f} %'.format(100 * precision_score(true_lab, pred_lab[:, 0], average='macro')))\n",
    "print('Weighted precision: {:.1f} %'.format(100 * precision_score(true_lab, pred_lab[:, 0], labels=labels, average='weighted')))\n",
    "\n",
    "print('\\n')\n",
    "print('Micro F1 score: {:.1f} %'.format(100 * f1_score(true_lab, pred_lab[:, 0], labels=labels, average='micro')))\n",
    "print('Macro F1 score: {:.1f} %'.format(100 * f1_score(true_lab, pred_lab[:, 0], labels=labels, average='macro')))\n",
    "print('Macro F1 score (no labels): {:.1f} %'.format(100 * f1_score(true_lab, pred_lab[:, 0], average='macro')))\n",
    "print('Weighted F1 score: {:.1f} %'.format(100 * f1_score(true_lab, pred_lab[:, 0], labels=labels, average='weighted')))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Computing the confusion matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Standard confusion matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrices(data, save_path=None, weighted=False,comments=\"normal\"):\n",
    "    \"\"\"\n",
    "    Plot confusion matrices and weighted confusion matrices for different thresholds using the provided data.\n",
    "\n",
    "    Args:\n",
    "        data (pandas.DataFrame): DataFrame containing the necessary data for plotting.\n",
    "        save_path (str): Path where the plots will be saved.\n",
    "        weighted (bool): Flag indicating whether to compute weighted confusion matrices.\n",
    "    \"\"\"\n",
    "\n",
    "    # Define the labels\n",
    "    class_names = np.unique(data[\"top1\"])\n",
    "\n",
    "    # Define the range for iterations\n",
    "    myrange = np.arange(0, 1, 0.25)\n",
    "\n",
    "    # Create the directory for saving the plots\n",
    "    pred_path = create_pred_path(save_path, dir=\"confusion\", weighted=weighted, comments=comments)\n",
    "\n",
    "\n",
    "    # Iterate over the range\n",
    "    for i in myrange:\n",
    "        # Get the current time\n",
    "        current_time = datetime.datetime.now().strftime(\"%Y-%m-%d %H:%M:%S\")\n",
    "        print(f\"[{current_time}] Plotting Confusion Matrix for {i}\")\n",
    "\n",
    "        # Filter the data based on a condition\n",
    "        slice = data[data[\"probability\"] > i]\n",
    "\n",
    "        # Extract true labels and predicted labels\n",
    "        y_true = np.array(slice[\"true_label\"])\n",
    "        y_pred = np.array(slice[\"top1\"])\n",
    "\n",
    "        y_true = y_true.astype(str)\n",
    "        y_pred = y_pred.astype(str)\n",
    "        # Compute the confusion matrix\n",
    "        weights = np.array(slice[\"probability\"]) if weighted else None\n",
    "        conf_mat = confusion_matrix(y_true, y_pred, labels=np.unique(y_pred), sample_weight=weights)\n",
    "\n",
    "        # Normalize the confusion matrix\n",
    "        normed_conf = conf_mat / np.sum(conf_mat, axis=1)[:, np.newaxis]\n",
    "\n",
    "        # Plot the confusion matrix using the custom function plt_conf_matrix\n",
    "        fig, ax = plt_conf_matrix(normed_conf, class_names)\n",
    "\n",
    "        # Save the plot\n",
    "        # Determine the file name suffix based on weighted option\n",
    "        file_suffix = f\"weighted_{i:.2f}\" if weighted else f\"_{i:.2f}\"\n",
    "\n",
    "        # Save the confusion matrix plot\n",
    "        fig.savefig(f\"{pred_path}/confusion_matrix{file_suffix}.jpg\", bbox_inches='tight')\n",
    "        plt.close(fig)\n",
    "    print(f\"Saved in {pred_path}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2024-04-04 19:48:37] Plotting Confusion Matrix for 0.0\n",
      "[2024-04-04 19:48:41] Plotting Confusion Matrix for 0.25\n",
      "[2024-04-04 19:48:44] Plotting Confusion Matrix for 0.5\n",
      "[2024-04-04 19:48:47] Plotting Confusion Matrix for 0.75\n",
      "Saved in /srv/phyto-plankton-classification/models/phytoplankton_vliz/results/confusion/normal\n"
     ]
    }
   ],
   "source": [
    "plot_confusion_matrices(data, save_path=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Weighted confusion matrix\n",
    "We weight each count in the confusion matrix by the probability of the prediction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2024-04-04 19:51:14] Plotting Confusion Matrix for 0.0\n",
      "[2024-04-04 19:51:18] Plotting Confusion Matrix for 0.25\n",
      "[2024-04-04 19:51:21] Plotting Confusion Matrix for 0.5\n",
      "[2024-04-04 19:51:24] Plotting Confusion Matrix for 0.75\n",
      "Saved in /srv/phyto-plankton-classification/models/phytoplankton_vliz/results/confusion/confusion_weighted\n"
     ]
    }
   ],
   "source": [
    "# Create the directory for saving the plots\n",
    "plot_confusion_matrices(data,save_path=False,weighted=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metrics evolution in function of cut-off \n",
    "The evolution of differnt metrics is shown for different cut-off values. All data below this cut-off is discarted when plotting new metrics."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_evolution_cut_off(data, save_path=False,comments=\"\"):\n",
    "    \"\"\"\n",
    "    Plot the evolution of metrics with varying cutoff probability.\n",
    "\n",
    "    Args:\n",
    "        data (DataFrame): Input data containing \"probability\", \"top1\", and \"true_label\" columns.\n",
    "        save_path (str): Path where the plots will be saved.\n",
    "\n",
    "    Returns:\n",
    "        None (plots are displayed and saved)\n",
    "    \"\"\"\n",
    "    myrange = np.arange(0, 1, 0.02)\n",
    "\n",
    "    pred_path = create_pred_path(save_path, dir=\"cut_off_changes\",comments=comments)\n",
    "\n",
    "    for cl in np.unique(data[\"top1\"]):\n",
    "        print(f\"Plotting cutoff probability for {cl}\")\n",
    "\n",
    "        metrics = []\n",
    "        data_slice = []\n",
    "\n",
    "        for i in myrange:\n",
    "            slice_data = data[data[\"probability\"] > i]\n",
    "            data_slice.append(compute_data_slice(slice_data, cl))\n",
    "\n",
    "            TP, FP, FN, TN = calculate_confusion_matrix(slice_data, cl)\n",
    "            precision, recall, outer_precision, outer_recall, f1_score, accuracy, specificity, fpr, fnr, auc_roc = compute_metrics(slice_data, cl, TP, FP, FN, TN)\n",
    "\n",
    "            metrics.append((precision, recall, outer_precision, outer_recall, f1_score, accuracy, specificity, fpr, fnr, auc_roc))\n",
    "\n",
    "        plot_metrics(myrange, metrics, data_slice, cl, pred_path)\n",
    "\n",
    "    print(\"Plots saved in:\", pred_path)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "def compute_data_slice(slice_data, cl):\n",
    "    return np.sum(slice_data['top1'] == cl) / np.sum(data['top1'] == cl)\n",
    "\n",
    "\n",
    "def calculate_confusion_matrix(slice_data, cl):\n",
    "    TP = np.sum((slice_data[\"true_label\"] == cl) & (slice_data[\"top1\"] == cl))\n",
    "    FP = np.sum((slice_data[\"true_label\"] != cl) & (slice_data[\"top1\"] == cl))\n",
    "    FN = np.sum((slice_data[\"true_label\"] == cl) & (slice_data[\"top1\"] != cl))\n",
    "    TN = np.sum((slice_data[\"true_label\"] != cl) & (slice_data[\"top1\"] != cl))\n",
    "    return TP, FP, FN, TN\n",
    "\n",
    "\n",
    "def compute_metrics(slice_data, cl, TP, FP, FN, TN):\n",
    "    precision = TP / (TP + FP)\n",
    "    recall = TP / (TP + FN)\n",
    "    outer_precision = TP / np.sum(slice_data[\"top1\"] == cl)\n",
    "    outer_recall = TP / np.sum(slice_data[\"true_label\"] == cl)\n",
    "\n",
    "    f1_score = 2 * (precision * recall) / (precision + recall)\n",
    "    accuracy = (TP + TN) / (TP + TN + FP + FN)\n",
    "    specificity = TN / (TN + FP)\n",
    "    fpr = FP / (FP + TN)\n",
    "    fnr = FN / (FN + TP)\n",
    "\n",
    "    if np.sum((slice_data[\"true_label\"] == cl)) != 0:\n",
    "        auc_roc = roc_auc_score((slice_data[\"true_label\"] == cl).astype(int), slice_data[\"probability\"])\n",
    "    else:\n",
    "        auc_roc = 0\n",
    "\n",
    "    return precision, recall, outer_precision, outer_recall, f1_score, accuracy, specificity, fpr, fnr, auc_roc\n",
    "\n",
    "\n",
    "def plot_metrics(myrange, metrics, data_slice, cl, pred_path):\n",
    "    plt.figure(figsize=(12, 8))\n",
    "    cmap = plt.get_cmap('tab10')\n",
    "    colors = cmap(np.linspace(0, 1, 12))\n",
    "\n",
    "    plt.plot(myrange, [m[0] for m in metrics], label=\"Precision\", linestyle='-', marker='o', color=colors[0])\n",
    "    plt.plot(myrange, [m[1] for m in metrics], label=\"Recall\", linestyle='--', marker='s', color=colors[1])\n",
    "    plt.plot(myrange, data_slice, label=\"Loss of Predictions\", linestyle='-', marker='*', color=colors[4])\n",
    "\n",
    "    plt.plot(myrange, [m[2] for m in metrics], label=\"Outer Precision\", linestyle=':', marker='^', color=colors[2])\n",
    "    plt.plot(myrange, [m[3] for m in metrics], label=\"Outer Recall\", linestyle='-.', marker='D', color=colors[3])\n",
    "\n",
    "    plt.plot(myrange, [m[4] for m in metrics], label=\"F1 Score\", linestyle='-', marker='x', color=colors[5])\n",
    "    plt.plot(myrange, [m[5] for m in metrics], label=\"Accuracy\", linestyle='--', marker='o', color=colors[6])\n",
    "    plt.plot(myrange, [m[6] for m in metrics], label=\"Specificity\", linestyle=':', marker='s', color=colors[7])\n",
    "    plt.plot(myrange, [m[7] for m in metrics], label=\"False Positive Rate\", linestyle='-.', marker='^', color=colors[8])\n",
    "    plt.plot(myrange, [m[8] for m in metrics], label=\"False Negative Rate\", linestyle='-', marker='D', color=colors[9])\n",
    "    plt.plot(myrange, [m[9] for m in metrics], label=\"AUC-ROC\", linestyle='--', marker='*', color=colors[10])\n",
    "\n",
    "    plt.legend(loc='best', fontsize='medium')\n",
    "    plt.xlabel(\"Cutoff value for probability\", fontsize='large')\n",
    "    plt.ylabel(\"Metric Score\", fontsize='large')\n",
    "    plt.title(cl, fontsize='x-large')\n",
    "\n",
    "    plt.xticks(fontsize='medium')\n",
    "    plt.yticks(fontsize='medium')\n",
    "\n",
    "    plt.grid(True, linestyle='--', alpha=0.7)\n",
    "\n",
    "    plt.savefig(f\"{pred_path}/{cl.replace('/', '+')}.jpg\", bbox_inches='tight')\n",
    "    plt.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plotting cutoff probability for Actinoptychus\n",
      "Plotting cutoff probability for Actinoptychus senarius\n",
      "Plotting cutoff probability for Actinoptychus splendens\n",
      "Plotting cutoff probability for Appendicularia\n",
      "Plotting cutoff probability for Artefact\n",
      "Plotting cutoff probability for Asterionella\n",
      "Plotting cutoff probability for Aulacodiscus argus\n",
      "Plotting cutoff probability for Bacillaria paxillifer\n",
      "Plotting cutoff probability for Bacillariophyceae\n",
      "Plotting cutoff probability for Bacillariophyceae_type1_colony\n",
      "Plotting cutoff probability for Bacteriastrum\n",
      "Plotting cutoff probability for Bellerochea\n",
      "Plotting cutoff probability for Bellerochea horologicalis\n",
      "Plotting cutoff probability for Biddulphia alternans\n",
      "Plotting cutoff probability for Biddulphianae\n",
      "Plotting cutoff probability for Brockmanniella brockmannii\n",
      "Plotting cutoff probability for Bubbles\n",
      "Plotting cutoff probability for Centric Diatom\n",
      "Plotting cutoff probability for Cerataulus granulata\n",
      "Plotting cutoff probability for Ceratium horridum/C. longipes\n",
      "Plotting cutoff probability for Chaetoceros\n",
      "Plotting cutoff probability for Chaetoceros affinis\n",
      "Plotting cutoff probability for Chaetoceros curvisetus/C. pseudocurvisetus\n",
      "Plotting cutoff probability for Chaetoceros danicus\n",
      "Plotting cutoff probability for Chaetoceros socialis\n",
      "Plotting cutoff probability for Ciliophora\n",
      "Plotting cutoff probability for Cnidaria\n",
      "Plotting cutoff probability for Copepoda adult\n",
      "Plotting cutoff probability for Coscinodiscus concinnus\n",
      "Plotting cutoff probability for Coscinodiscus granii\n",
      "Plotting cutoff probability for Crustacea\n",
      "Plotting cutoff probability for Crustaceae:part\n",
      "Plotting cutoff probability for Dactyliosolen/Cerataulina/Guinardia\n",
      "Plotting cutoff probability for Detritus\n",
      "Plotting cutoff probability for Dinoflagellata\n",
      "Plotting cutoff probability for Dinoflagellate cyst\n",
      "Plotting cutoff probability for Diploneis\n",
      "Plotting cutoff probability for Ditylum brightwellii\n",
      "Plotting cutoff probability for Egg/Cyst\n",
      "Plotting cutoff probability for Eucampia\n",
      "Plotting cutoff probability for Faecal pellet\n",
      "Plotting cutoff probability for Favella\n",
      "Plotting cutoff probability for Fibers\n",
      "Plotting cutoff probability for Foraminifera\n",
      "Plotting cutoff probability for Guinardia delicatula\n",
      "Plotting cutoff probability for Guinardia flaccida\n",
      "Plotting cutoff probability for Guinardia striata/Dactyliosolen phuketensis\n",
      "Plotting cutoff probability for Helicotheca tamesis\n",
      "Plotting cutoff probability for Hobaniella longicruris\n",
      "Plotting cutoff probability for Lauderia/Melosira/Detonula\n",
      "Plotting cutoff probability for Leptocylindraceae\n",
      "Plotting cutoff probability for Lithodesmium undulatum\n",
      "Plotting cutoff probability for Melosira\n",
      "Plotting cutoff probability for Meuniera membranacea\n",
      "Plotting cutoff probability for Mollusca\n",
      "Plotting cutoff probability for Nauplii\n",
      "Plotting cutoff probability for Neocalyptrella robusta\n",
      "Plotting cutoff probability for Noctiluca scintillans\n",
      "Plotting cutoff probability for Noctilucales\n",
      "Plotting cutoff probability for Odontella aurita/Ralfsiella minima\n",
      "Plotting cutoff probability for Odontella rhombus f. trigona\n",
      "Plotting cutoff probability for Paralia\n",
      "Plotting cutoff probability for Pennate Diatom\n",
      "Plotting cutoff probability for Pennate Diatom colony\n",
      "Plotting cutoff probability for Peritrichia\n",
      "Plotting cutoff probability for Phytoplankton Colony\n",
      "Plotting cutoff probability for Plagiogrammopsis/Bellerochea malleus\n",
      "Plotting cutoff probability for Pollen\n",
      "Plotting cutoff probability for Polychaeta\n",
      "Plotting cutoff probability for Porifera spicule\n",
      "Plotting cutoff probability for Proboscia alata\n",
      "Plotting cutoff probability for Proboscia indica\n",
      "Plotting cutoff probability for Protoperidinium\n",
      "Plotting cutoff probability for Protoperidinium pentagonum\n",
      "Plotting cutoff probability for Pseudo-nitzschia\n",
      "Plotting cutoff probability for Remnant\n",
      "Plotting cutoff probability for Rhizosolenia\n",
      "Plotting cutoff probability for Rhizosolenia setigera (f. pungens)/R. hebetata f. semispina\n",
      "Plotting cutoff probability for Rotifera\n",
      "Plotting cutoff probability for Skeletonema\n",
      "Plotting cutoff probability for Stellarima stellaris/Podosira/Hyalodiscus\n",
      "Plotting cutoff probability for Stephanopyxis\n",
      "Plotting cutoff probability for Suctoria\n",
      "Plotting cutoff probability for Synedra/Thalassionema\n",
      "Plotting cutoff probability for Thalassiosira/Porosira\n",
      "Plotting cutoff probability for Tintinnina\n",
      "Plotting cutoff probability for Tintinnopsis\n",
      "Plotting cutoff probability for Triceratium favus\n",
      "Plotting cutoff probability for Trieres mobiliensis/T. regia\n",
      "Plotting cutoff probability for Trieres sinensis\n",
      "Plotting cutoff probability for Tripos\n",
      "Plotting cutoff probability for Tripos fusus\n",
      "Plotting cutoff probability for Veliger larvae D-shaped\n",
      "Plotting cutoff probability for Zooplankton\n",
      "Plotting cutoff probability for Zygoceros\n",
      "Plots saved in: /srv/phyto-plankton-classification/models/phytoplankton_vliz/results/cut_off_changes/\n"
     ]
    }
   ],
   "source": [
    "plot_evolution_cut_off(data, save_path=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The Proportional and absolute progression of TRUE class is shown\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_classwise_progression(data,save_path,comments=\"\"):\n",
    "    \"\"\"\n",
    "    Plot the class-wise progression of TRUE class distribution for different cutoff values.\n",
    "\n",
    "    Args:\n",
    "        data (pandas.DataFrame): DataFrame containing the necessary data for plotting.\n",
    "    \"\"\"\n",
    "\n",
    "    # Define the range for iterations\n",
    "    myrange = np.arange(0, 1, 0.02)\n",
    "\n",
    "    # Create directory for saving results\n",
    "    pred_path = create_pred_path(save_path, dir=\"proportional_progression\",comments=comments)\n",
    "\n",
    "\n",
    "    for cl in np.unique(data[\"top1\"]):\n",
    "        print(f\"Plotting proportional progression for {cl}\")\n",
    "        compo = pd.DataFrame(data[data['top1'] == cl][\"true_label\"].value_counts())\n",
    "        for j, i in enumerate(myrange):\n",
    "            slice = data[data[\"probability\"] > i]\n",
    "            x = slice[slice['top1']==cl][\"true_label\"].value_counts()\n",
    "            compo[str(j)] = x\n",
    "\n",
    "\n",
    "        compo = compo.drop(\"true_label\", axis=1)\n",
    "        plotdata = compo / compo.sum()\n",
    "        plotdata1 = plotdata[plotdata > 0.01] * 100\n",
    "        plotdata2 = compo[plotdata > 0.01]\n",
    "        plotdata1.columns = myrange\n",
    "        plotdata2.columns = myrange\n",
    "        fig, axs = plt.subplots(1, 2, figsize=(16, 6))\n",
    "        plotdata1.dropna(how=\"all\").transpose().plot.area(ax=axs[0], legend=False,\n",
    "                                                          title=\"proportional progression of TRUE class distribution\",\n",
    "                                                          xlabel=\"Cutoff value for probability\",\n",
    "                                                          ylabel=\"Class proportion (%)\")\n",
    "        plotdata2.dropna(how=\"all\").transpose().plot.area(ax=axs[1],\n",
    "                                                          title=\"absolute progression of TRUE class distribution\",\n",
    "                                                          xlabel=\"Cutoff value for probability\",\n",
    "                                                          ylabel=\"Class count\")\n",
    "        plt.suptitle(cl)\n",
    "\n",
    "        # Save the plot\n",
    "        plt.savefig(f\"{pred_path}/{cl.replace('/', '+')}.jpg\", bbox_inches='tight')\n",
    "        plt.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plotting proportional progression for Actinoptychus\n",
      "Plotting proportional progression for Actinoptychus senarius\n",
      "Plotting proportional progression for Actinoptychus splendens\n",
      "Plotting proportional progression for Appendicularia\n",
      "Plotting proportional progression for Artefact\n",
      "Plotting proportional progression for Asterionella\n",
      "Plotting proportional progression for Aulacodiscus argus\n",
      "Plotting proportional progression for Bacillaria paxillifer\n",
      "Plotting proportional progression for Bacillariophyceae\n",
      "Plotting proportional progression for Bacillariophyceae_type1_colony\n",
      "Plotting proportional progression for Bacteriastrum\n",
      "Plotting proportional progression for Bellerochea\n",
      "Plotting proportional progression for Bellerochea horologicalis\n",
      "Plotting proportional progression for Biddulphia alternans\n",
      "Plotting proportional progression for Biddulphianae\n",
      "Plotting proportional progression for Brockmanniella brockmannii\n",
      "Plotting proportional progression for Bubbles\n",
      "Plotting proportional progression for Centric Diatom\n",
      "Plotting proportional progression for Cerataulus granulata\n",
      "Plotting proportional progression for Ceratium horridum/C. longipes\n",
      "Plotting proportional progression for Chaetoceros\n",
      "Plotting proportional progression for Chaetoceros affinis\n",
      "Plotting proportional progression for Chaetoceros curvisetus/C. pseudocurvisetus\n",
      "Plotting proportional progression for Chaetoceros danicus\n",
      "Plotting proportional progression for Chaetoceros socialis\n",
      "Plotting proportional progression for Ciliophora\n",
      "Plotting proportional progression for Cnidaria\n",
      "Plotting proportional progression for Copepoda adult\n",
      "Plotting proportional progression for Coscinodiscus concinnus\n",
      "Plotting proportional progression for Coscinodiscus granii\n",
      "Plotting proportional progression for Crustacea\n",
      "Plotting proportional progression for Crustaceae:part\n",
      "Plotting proportional progression for Dactyliosolen/Cerataulina/Guinardia\n",
      "Plotting proportional progression for Detritus\n",
      "Plotting proportional progression for Dinoflagellata\n",
      "Plotting proportional progression for Dinoflagellate cyst\n",
      "Plotting proportional progression for Diploneis\n",
      "Plotting proportional progression for Ditylum brightwellii\n",
      "Plotting proportional progression for Egg/Cyst\n",
      "Plotting proportional progression for Eucampia\n",
      "Plotting proportional progression for Faecal pellet\n",
      "Plotting proportional progression for Favella\n",
      "Plotting proportional progression for Fibers\n",
      "Plotting proportional progression for Foraminifera\n",
      "Plotting proportional progression for Guinardia delicatula\n",
      "Plotting proportional progression for Guinardia flaccida\n",
      "Plotting proportional progression for Guinardia striata/Dactyliosolen phuketensis\n",
      "Plotting proportional progression for Helicotheca tamesis\n",
      "Plotting proportional progression for Hobaniella longicruris\n",
      "Plotting proportional progression for Lauderia/Melosira/Detonula\n",
      "Plotting proportional progression for Leptocylindraceae\n",
      "Plotting proportional progression for Lithodesmium undulatum\n",
      "Plotting proportional progression for Melosira\n",
      "Plotting proportional progression for Meuniera membranacea\n",
      "Plotting proportional progression for Mollusca\n",
      "Plotting proportional progression for Nauplii\n",
      "Plotting proportional progression for Neocalyptrella robusta\n",
      "Plotting proportional progression for Noctiluca scintillans\n",
      "Plotting proportional progression for Noctilucales\n",
      "Plotting proportional progression for Odontella aurita/Ralfsiella minima\n",
      "Plotting proportional progression for Odontella rhombus f. trigona\n",
      "Plotting proportional progression for Paralia\n",
      "Plotting proportional progression for Pennate Diatom\n",
      "Plotting proportional progression for Pennate Diatom colony\n",
      "Plotting proportional progression for Peritrichia\n",
      "Plotting proportional progression for Phytoplankton Colony\n",
      "Plotting proportional progression for Plagiogrammopsis/Bellerochea malleus\n",
      "Plotting proportional progression for Pollen\n",
      "Plotting proportional progression for Polychaeta\n",
      "Plotting proportional progression for Porifera spicule\n",
      "Plotting proportional progression for Proboscia alata\n",
      "Plotting proportional progression for Proboscia indica\n",
      "Plotting proportional progression for Protoperidinium\n",
      "Plotting proportional progression for Protoperidinium pentagonum\n",
      "Plotting proportional progression for Pseudo-nitzschia\n",
      "Plotting proportional progression for Remnant\n",
      "Plotting proportional progression for Rhizosolenia\n",
      "Plotting proportional progression for Rhizosolenia setigera (f. pungens)/R. hebetata f. semispina\n",
      "Plotting proportional progression for Rotifera\n",
      "Plotting proportional progression for Skeletonema\n",
      "Plotting proportional progression for Stellarima stellaris/Podosira/Hyalodiscus\n",
      "Plotting proportional progression for Stephanopyxis\n",
      "Plotting proportional progression for Suctoria\n",
      "Plotting proportional progression for Synedra/Thalassionema\n",
      "Plotting proportional progression for Thalassiosira/Porosira\n",
      "Plotting proportional progression for Tintinnina\n",
      "Plotting proportional progression for Tintinnopsis\n",
      "Plotting proportional progression for Triceratium favus\n",
      "Plotting proportional progression for Trieres mobiliensis/T. regia\n",
      "Plotting proportional progression for Trieres sinensis\n",
      "Plotting proportional progression for Tripos\n",
      "Plotting proportional progression for Tripos fusus\n",
      "Plotting proportional progression for Veliger larvae D-shaped\n",
      "Plotting proportional progression for Zooplankton\n",
      "Plotting proportional progression for Zygoceros\n"
     ]
    }
   ],
   "source": [
    "plot_classwise_progression(data,save_path=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Define cutoff - Nic**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to define the cutoff probability for a given class\n",
    "def define_cutoff(data, cl, p_aim=0.95):\n",
    "    # Slice the data for the specific class and sort it in descending order based on probability\n",
    "    slice = data[data[\"top1\"] == cl].sort_values(by=\"probability\", ascending=False).reset_index()\n",
    "    # Calculate the precision for each position in the slice\n",
    "    prec_list = np.cumsum(slice[\"top1_correct\"]) / (slice.index + 1)\n",
    "    # Find the positions where precision exceeds the desired aim\n",
    "    acceptable_probabs = np.where(prec_list > p_aim)[0]\n",
    "    if len(acceptable_probabs) > 0:\n",
    "        # Return the minimum probability from the acceptable positions\n",
    "        return np.min(slice['probability'][acceptable_probabs])\n",
    "    else:\n",
    "        # If no acceptable positions, return infinity\n",
    "        return np.inf\n",
    "\n",
    "# Function to apply the mask to the data based on the cutoff probabilities\n",
    "def apply_mask(data, mask_data):\n",
    "    # Initialize a mask with False values for each data point\n",
    "    data_mask = np.repeat(False, len(data))\n",
    "    for _, x in mask_data.iterrows():\n",
    "        # Update the mask where the class and probability satisfy the cutoff criteria\n",
    "        data_mask[(data[\"top1\"] == x[\"Class\"]) & (data[\"probability\"] >= x[\"probability_cutoff\"])] = True\n",
    "    # Return the masked data\n",
    "    return data[data_mask]\n",
    "\n",
    "def mask_creater(p_aim,data):\n",
    "    mask = [[x, define_cutoff(data, x, p_aim=p_aim)] for x in np.unique(data['top1'])]\n",
    "    mask_data = pd.DataFrame(mask, columns=['Class', \"probability_cutoff\"])\n",
    "    masked_data = apply_mask(data, mask_data)\n",
    "    return masked_data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_summary_threshold(data, save_path=False,comments=\"\"):\n",
    "    \"\"\"\n",
    "    Plot the scores on a masked DataFrame with aimed precision.\n",
    "\n",
    "    Args:\n",
    "        data (pandas.DataFrame): DataFrame containing the necessary data for plotting.\n",
    "        name (str): Name of the plot.\n",
    "    \"\"\"\n",
    "\n",
    "    # Define the range of precision values to aim for\n",
    "    myrange = np.concatenate((np.arange(0.5, 0.95, 0.01), np.arange(0.95, 0.995, 0.005)))\n",
    "\n",
    "    # Create directory for saving results\n",
    "    pred_path = create_pred_path(save_path, dir=\"summary_threshold\",comments=comments)\n",
    "\n",
    "    # Initialize arrays to store the evaluation metrics\n",
    "    acc= np.zeros(len(myrange))\n",
    "    pr_weighted = np.zeros(len(myrange))\n",
    "    rec_weighted = np.zeros(len(myrange))\n",
    "    f1_weighted = np.zeros(len(myrange))\n",
    "    pr_macro = np.zeros(len(myrange))\n",
    "    rec_macro = np.zeros(len(myrange))\n",
    "    f1_macro = np.zeros(len(myrange))\n",
    "    dataset_prop = np.zeros(len(myrange))\n",
    "    class_prop = np.zeros(len(myrange))\n",
    "\n",
    "    # Calculate evaluation metrics for each precision value\n",
    "    for i, p_aim in enumerate(tqdm(myrange, desc='Processing')):\n",
    "        masked_data =mask_creater(p_aim,data)\n",
    "\n",
    "        dataset_prop[i] = len(masked_data) / len(data)\n",
    "        class_prop[i] = len(np.unique(masked_data[\"top1\"])) / len(np.unique(data[\"top1\"]))\n",
    "        acc[i] = accuracy_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"])#, average=\"weighted\")\n",
    "        pr_weighted[i] = precision_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"weighted\")\n",
    "        rec_weighted[i] = recall_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"weighted\")\n",
    "        f1_weighted[i] = f1_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"weighted\")\n",
    "        # pr_macro[i] = precision_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "        # rec_macro[i] = recall_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "        # f1_macro[i] = f1_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "        # acc[i] = accuracy_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"])\n",
    "        # pr_micro[i] = precision_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"micro\",)\n",
    "        # rec_micro[i] = recall_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"micro\")\n",
    "        # f1_micro[i] = f1_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"micro\")\n",
    "        # pr_macro[i] = precision_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "        # rec_macro[i] = recall_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "        # f1_macro[i] = f1_score(y_true=masked_data[\"true_label\"], y_pred=masked_data[\"top1\"], average=\"macro\", zero_division=1)\n",
    "\n",
    "    print(\"done!\")\n",
    "\n",
    "    # Create a DataFrame to show the evaluation metrics for different precision values\n",
    "    show_data = pd.DataFrame({\n",
    "        \"Accuracy\": acc,\n",
    "        \"Precision (weighted)\": pr_weighted,\n",
    "        \"Recall (weighted)\": rec_weighted,\n",
    "        \"F1 Score (weighted)\": f1_weighted,\n",
    "        \"Dataset Proportion\": dataset_prop,\n",
    "        \"Class Proportion\": class_prop\n",
    "    }, index=myrange)\n",
    "\n",
    "    # Plot the evaluation metrics\n",
    "    plt.figure(figsize=(12, 8))\n",
    "    \n",
    "    # Define custom colors for each metric line\n",
    "    colors = [\"#1f77b4\", \"#ff7f0e\", \"#2ca02c\", \"#d62728\", \"#9467bd\", \"#8c564b\"]#, \"#e377c2\", \"#7f7f7f\", \"#bcbd22\", \"#17becf\"]\n",
    "\n",
    "    # Plot each metric with a different color and line style\n",
    "    for i, metric in enumerate(show_data.columns):\n",
    "        if metric in [\"Dataset Proportion\", \"Class Proportion\"]:\n",
    "            plt.plot(show_data.index, show_data[metric], marker='o', linestyle='--', color=colors[i], label=metric, alpha=0.7)\n",
    "        elif metric==\"Accuracy\":\n",
    "            # print(show_data[metric])\n",
    "            plt.plot(show_data.index, show_data[metric], marker='x', color=colors[i], label=metric, alpha=1,markersize=10)\n",
    "        else:\n",
    "            plt.plot(show_data.index, show_data[metric], marker='o', color=colors[i], label=metric, alpha=0.7)\n",
    "\n",
    "    plt.title(\"Scores on after masking increasing thresholds\")\n",
    "    plt.xlabel(\"Probability Threshold \")\n",
    "    plt.xticks(rotation=45)\n",
    "    # plt.yticks(np.arange(0, 1.1, 0.1))\n",
    "    plt.legend()\n",
    "    plt.tight_layout()\n",
    "\n",
    "    # Save the plot\n",
    "    filename = os.path.join(pred_path, \"summary_threshold_plot.jpg\")\n",
    "    plt.savefig(filename)\n",
    "    print(\"Saved in\", filename)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Processing: 100%|██████████| 55/55 [00:39<00:00,  1.38it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done!\n",
      "Saved in /srv/phyto-plankton-classification/models/phytoplankton_vliz/results/summary_threshold/saved_0.5/summary_threshold_plot.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_summary_threshold(data,comments=\"saved_0.5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Processing:   0%|          | 0/15 [00:00<?, ?it/s]Processing: 100%|██████████| 15/15 [07:11<00:00, 28.75s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done!\n",
      "Saved in /srv/phyto-plankton-classification/models/2023-06-05_143422/results/summary_threshold/saved/summary_threshold_plot.jpg\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_summary_threshold(data,comments=\"saved\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}