dirichlet_model_training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dirichlet model training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook we will train Dirichlet model for atrial fibrillation detection."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Table of contents"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* [Dataset initialization](#Dataset-initialization)\n",
    "* [Training pipeline](#Training-pipeline)\n",
    "* [Saving the model](#Saving-the-model)\n",
    "* [Testing pipeline](#Testing-pipeline)\n",
    "* [Predicting pipeline](#Predicting-pipeline)\n",
    "* [Analyzing the uncertainty](#Analyzing-the-uncertainty)\n",
    "* [Visualizing predictions](#Visualizing-predictions)\n",
    "    * [Certain prediction](#Certain-prediction)\n",
    "    * [Uncertain prediction](#Uncertain-prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ubuntu/anaconda3/envs/tensorflow_p36/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n",
      "/home/ubuntu/anaconda3/envs/tensorflow_p36/lib/python3.6/site-packages/dicom/__init__.py:53: UserWarning: \n",
      "This code is using an older version of pydicom, which is no longer \n",
      "maintained as of Jan 2017.  You can access the new pydicom features and API \n",
      "by installing `pydicom` from PyPI.\n",
      "See 'Transitioning to pydicom 1.x' section at pydicom.readthedocs.org \n",
      "for more information.\n",
      "\n",
      "  warnings.warn(msg)\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import sys\n",
    "from functools import partial\n",
    "\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy.stats import beta\n",
    "\n",
    "sys.path.append(os.path.join(\"..\", \"..\", \"..\"))\n",
    "import cardio.dataset as ds\n",
    "from cardio import EcgDataset\n",
    "from cardio.dataset import B, V, F\n",
    "from cardio.models.dirichlet_model import DirichletModel, concatenate_ecg_batch\n",
    "from cardio.models.metrics import f1_score, classification_report, confusion_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Seaborn plotting parameters setting:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set(\"talk\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default, TensorFlow attempts to allocate almost the entire memory on all of the available GPUs. Executing this instruction makes only the GPU with id 0 visible for TensorFlow in this process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "env: CUDA_VISIBLE_DEVICES=0\n"
     ]
    }
   ],
   "source": [
    "#%env CUDA_VISIBLE_DEVICES=0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset initialization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we need to specify paths to ECG signals and their labels:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "SIGNALS_PATH = os.path.expanduser(\"~\")+\"/coding/cnn/datasets/training2017/\"\n",
    "SIGNALS_MASK = SIGNALS_PATH + \"*.hea\"\n",
    "LABELS_PATH = SIGNALS_PATH + \"REFERENCE.csv\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's create an ECG dataset and perform a train/test split:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "eds = EcgDataset(path=SIGNALS_MASK, no_ext=True, sort=True)\n",
    "eds.cv_split(0.8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training pipeline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dirichlet model builder expects model config to contain input signals' shape and class names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.5, allow_growth=True)\n",
    "\n",
    "model_config = {\n",
    "    \"session\": {\"config\": tf.ConfigProto(gpu_options=gpu_options)},\n",
    "    \"input_shape\": F(lambda batch: batch.signal[0].shape[1:]),\n",
    "    \"class_names\": F(lambda batch: batch.label_binarizer.classes_),\n",
    "    \"loss\": None,\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "N_EPOCH = 1000\n",
    "BATCH_SIZE = 256"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model training pipeline is composed of:\n",
    "* model initialization with the config defined above\n",
    "* data loading, preprocessing (e.g. flipping) and augmentation (e.g. resampling)\n",
    "* train step execution\n",
    "\n",
    "Let's create a template pipeline, then link it to our training dataset and run:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "template_train_ppl = (\n",
    "    ds.Pipeline()\n",
    "      .init_model(\"dynamic\", DirichletModel, name=\"dirichlet\", config=model_config)\n",
    "      .init_variable(\"loss_history\", init_on_each_run=list)\n",
    "      .load(components=[\"signal\", \"meta\"], fmt=\"wfdb\")\n",
    "      .load(components=\"target\", fmt=\"csv\", src=LABELS_PATH)\n",
    "      .drop_labels([\"~\"])\n",
    "      .rename_labels({\"N\": \"NO\", \"O\": \"NO\"})\n",
    "      .flip_signals()\n",
    "      .random_resample_signals(\"normal\", loc=300, scale=10)\n",
    "      .random_split_signals(2048, {\"A\": 9, \"NO\": 3})\n",
    "      .binarize_labels()\n",
    "      .train_model(\"dirichlet\", make_data=concatenate_ecg_batch,\n",
    "                   fetches=\"loss\", save_to=V(\"loss_history\"), mode=\"a\")\n",
    "      .run(batch_size=BATCH_SIZE, shuffle=True, drop_last=True, n_epochs=N_EPOCH, lazy=True)\n",
    ")\n",
    "\n",
    "train_ppl = (eds.train >> template_train_ppl).run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training loss is stored in \"loss_history\" pipeline variable. Let's take a look at its plot:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA44AAAESCAYAAACyzIuFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec7278c320>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_loss = [np.mean(l) for l in np.array_split(train_ppl.get_variable(\"loss_history\"), N_EPOCH)]\n",
    "\n",
    "fig = plt.figure(figsize=(15, 4))\n",
    "plt.plot(train_loss)\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Training loss\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, training loss almost reaches a plateau by the end of the training."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Saving the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#MODEL_PATH = \"D:\\\\Projects\\\\data\\\\ecg\\\\dirichlet_model\"\n",
    "MODEL_PATH = os.path.expanduser(\"~\")+\"/coding/cnn/DeepECG/dirichlet_model\"\n",
    "train_ppl.save_model(\"dirichlet\", path=MODEL_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Testing pipeline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Testing pipeline is almost identical to the training one. The differences lie in the absence of signal resampling and the modified segmentation procedure. Notice, that the model is imported from the training pipeline, rather than being constructed from scratch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "template_test_ppl = (\n",
    "    ds.Pipeline()\n",
    "      .import_model(\"dirichlet\", train_ppl)\n",
    "      .init_variable(\"predictions_list\", init_on_each_run=list)\n",
    "      .load(components=[\"signal\", \"meta\"], fmt=\"wfdb\")\n",
    "      .load(components=\"target\", fmt=\"csv\", src=LABELS_PATH)\n",
    "      .drop_labels([\"~\"])\n",
    "      .rename_labels({\"N\": \"NO\", \"O\": \"NO\"})\n",
    "      .flip_signals()\n",
    "      .split_signals(2048, 2048)\n",
    "      .binarize_labels()\n",
    "      .predict_model(\"dirichlet\", make_data=concatenate_ecg_batch,\n",
    "                     fetches=\"predictions\", save_to=V(\"predictions_list\"), mode=\"e\")\n",
    "      .run(batch_size=BATCH_SIZE, shuffle=False, drop_last=False, n_epochs=1, lazy=True)\n",
    ")\n",
    "\n",
    "test_ppl = (eds.test >> template_test_ppl).run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now \"predictions_list\" pipeline variable stores model predictions and true targets for all signals labeled with \"A\", \"O\" and \"N\" in the testing dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use [F1-score](https://en.wikipedia.org/wiki/F1_score) with macro averaging to measure classification performance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = test_ppl.get_variable(\"predictions_list\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.89170797140566993"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          A       0.84      0.77      0.80       171\n",
      "         NO       0.97      0.98      0.98      1484\n",
      "\n",
      "avg / total       0.96      0.96      0.96      1655\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Also take a look at the more detailed report - the confusion matrix:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>True</th>\n",
       "      <th>A</th>\n",
       "      <th>NO</th>\n",
       "      <th>All</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Pred</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A</th>\n",
       "      <td>132</td>\n",
       "      <td>25</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NO</th>\n",
       "      <td>39</td>\n",
       "      <td>1459</td>\n",
       "      <td>1498</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>All</th>\n",
       "      <td>171</td>\n",
       "      <td>1484</td>\n",
       "      <td>1655</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "True    A    NO   All\n",
       "Pred                 \n",
       "A     132    25   157\n",
       "NO     39  1459  1498\n",
       "All   171  1484  1655"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model misclassifies 33 patients with atrial fibrillation and 25 patients with normal and other rhythms. All other patients were classified correctly."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We’ve already obtained good classification performance. Let’s see if we can do even better."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analyzing the uncertainty"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In addition to class probabilities the model returns its uncertainty in the prediction, which varies from 0 (absolutely sure) to 1 (absolutely unsure). You can see the uncertainty histogram on the plot below:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA20AAAEPCAYAAADRUSyyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec216f1128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "uncertainty = [d[\"uncertainty\"] for d in predictions]\n",
    "\n",
    "fig = plt.figure(figsize=(15, 4))\n",
    "sns.distplot(uncertainty, hist=True, norm_hist=True, kde=False)\n",
    "plt.xlabel(\"Model uncertainty\")\n",
    "plt.xlim(-0.05, 1.05)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The figure above shows, that the model is almost always certain in its predictions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compare the metrics for the full testing dataset above with the same metrics for 90% most certain predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "q = 90\n",
    "thr = np.percentile(uncertainty, q)\n",
    "certain_predictions = [d for d in predictions if d[\"uncertainty\"] <= thr]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.94263368777931889"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_score(certain_predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          A       0.93      0.86      0.89       107\n",
      "         NO       0.99      0.99      0.99      1382\n",
      "\n",
      "avg / total       0.98      0.99      0.98      1489\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(certain_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>True</th>\n",
       "      <th>A</th>\n",
       "      <th>NO</th>\n",
       "      <th>All</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Pred</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A</th>\n",
       "      <td>92</td>\n",
       "      <td>7</td>\n",
       "      <td>99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NO</th>\n",
       "      <td>15</td>\n",
       "      <td>1375</td>\n",
       "      <td>1390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>All</th>\n",
       "      <td>107</td>\n",
       "      <td>1382</td>\n",
       "      <td>1489</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "True    A    NO   All\n",
       "Pred                 \n",
       "A      92     7    99\n",
       "NO     15  1375  1390\n",
       "All   107  1382  1489"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(certain_predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can observe a significant increase in precision, recall and F1-score for the atrial fibrillation class. Now only 16 signals were misclassified."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Predicting pipeline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's predict class probabilities for a new, unobserved ECG signal.<br>\n",
    "Besides, we will load pretrained model from MODEL_PATH directory instead of importing it from another pipeline:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SIGNALS_PATH = \"D:\\\\Projects\\\\data\\\\ecg\\\\training2017\\\\\"\n",
    "SIGNALS_PATH = os.path.expanduser(\"~\")+\"/coding/cnn/datasets/training2017/\"\n",
    "# MODEL_PATH = \"D:\\\\Projects\\\\data\\\\ecg\\\\dirichlet_model\"\n",
    "MODEL_PATH = os.path.expanduser(\"~\")+\"/coding/cnn/DeepECG/dirichlet_model\"\n",
    "\n",
    "BATCH_SIZE = 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.5, allow_growth=True)\n",
    "\n",
    "model_config = {\n",
    "    \"session\": {\"config\": tf.ConfigProto(gpu_options=gpu_options)},\n",
    "    \"build\": False,\n",
    "    \"load\": {\"path\": MODEL_PATH},\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from /home/ubuntu/coding/cnn/DeepECG/dirichlet_model/model-26000\n"
     ]
    }
   ],
   "source": [
    "template_predict_ppl = (\n",
    "    ds.Pipeline()\n",
    "      .init_model(\"static\", DirichletModel, name=\"dirichlet\", config=model_config)\n",
    "      .init_variable(\"predictions_list\", init_on_each_run=list)\n",
    "      .load(fmt=\"wfdb\", components=[\"signal\", \"meta\"])\n",
    "      .flip_signals()\n",
    "      .split_signals(2048, 2048)\n",
    "      .predict_model(\"dirichlet\", make_data=partial(concatenate_ecg_batch, return_targets=False),\n",
    "                     fetches=\"predictions\", save_to=V(\"predictions_list\"), mode=\"e\")\n",
    "      .run(batch_size=BATCH_SIZE, shuffle=False, drop_last=False, n_epochs=1, lazy=True)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to create a dataset with a single ECG in it, then link it to the template predicting pipeline defined above and run it. Model prediction will be stored in the \"predictions_list\" variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "signal_name = \"A00001.hea\"\n",
    "signal_path = SIGNALS_PATH + signal_name\n",
    "predict_eds = EcgDataset(path=signal_path, no_ext=True, sort=True)\n",
    "predict_ppl = (predict_eds >> template_predict_ppl).run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'target_pred': {'A': 0.021257795, 'NO': 0.97874224},\n",
       "  'uncertainty': 0.0075736045837402344}]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_ppl.get_variable(\"predictions_list\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The length of the resulting list equals the length of the index of the dataset (1 in out case)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's look at the target Dirichlet mixture density for a given signal. The pipeline below stores the signal and Dirichlet distribution parameters in its variables in addition to the predicted class probabilities."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from /home/ubuntu/coding/cnn/DeepECG/dirichlet_model/model-26000\n"
     ]
    }
   ],
   "source": [
    "template_full_predict_ppl = (\n",
    "    ds.Pipeline()\n",
    "      .init_model(\"static\", DirichletModel, name=\"dirichlet\", config=model_config)\n",
    "      .init_variable(\"signals\", init_on_each_run=list)\n",
    "      .init_variable(\"predictions_list\", init_on_each_run=list)\n",
    "      .init_variable(\"parameters_list\", init_on_each_run=list)\n",
    "      .load(fmt=\"wfdb\", components=[\"signal\", \"meta\"])\n",
    "      .update_variable(\"signals\", value=B(\"signal\"))\n",
    "      .flip_signals()\n",
    "      .split_signals(2048, 2048)\n",
    "      .predict_model(\"dirichlet\", make_data=partial(concatenate_ecg_batch, return_targets=False),\n",
    "                     fetches=[\"predictions\", \"parameters\"],\n",
    "                     save_to=[V(\"predictions_list\"), V(\"parameters_list\")], mode=\"e\")\n",
    "      .run(batch_size=BATCH_SIZE, shuffle=False, drop_last=False, n_epochs=1, lazy=True)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_and_visualize(signal_path):\n",
    "    predict_eds = EcgDataset(path=signal_path, no_ext=True, sort=True)\n",
    "    \n",
    "    full_predict_ppl = (predict_eds >> template_full_predict_ppl).run()\n",
    "    signal = full_predict_ppl.get_variable(\"signals\")[0][0][0][:2000].ravel()\n",
    "    predictions = full_predict_ppl.get_variable(\"predictions_list\")[0]\n",
    "    parameters = full_predict_ppl.get_variable(\"parameters_list\")[0]\n",
    "    \n",
    "    print(predictions)\n",
    "\n",
    "    x = np.linspace(0.001, 0.999, 1000)\n",
    "    y = np.zeros_like(x)\n",
    "    for alpha in parameters:\n",
    "        y += beta.pdf(x, *alpha)\n",
    "    y /= len(parameters)\n",
    "    \n",
    "    fig, (ax1, ax2) = plt.subplots(1, 2, gridspec_kw={\"width_ratios\": [2.5, 1]}, figsize=(15, 4))\n",
    "\n",
    "    ax1.plot(signal)\n",
    "\n",
    "    ax2.plot(x, y)\n",
    "    ax2.fill_between(x, y, alpha=0.3)\n",
    "    ax2.set_ylim(ymin=0)\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Certain prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, let’s look at the healthy person’s ECG. The signal is shown on the left plot. Note that it has a clear quasi periodic structure. The right plot shows the pdf of the mixture distributions with atrial fibrillation probability plotted on the horizontal axis. The model is absolutely certain in the absence of AF: almost all the probability density is concentrated around 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'target_pred': {'A': 0.021271603, 'NO': 0.97872835}, 'uncertainty': 0.0074365139007568359}\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4AAAAD9CAYAAADtYBnVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIygAJIQQQgghhBAf8f8BgxnzV7qG/oYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec2138f470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "predict_and_visualize(SIGNALS_PATH + \"A00150.hea\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Uncertain prediction"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And now comes an ECG with irregular structure, which may be caused by a disease or some measurement errors. The probability density on the right plot is almost equally concentrated around 0 and 1. This is an example of an uncertain prediction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'target_pred': {'A': 0.19596583, 'NO': 0.80403417}, 'uncertainty': 0.4200446605682373}\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3QAAAECCAYAAAC2dJVOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQgghJE79f46dhvzuxYCaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fec21377780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "predict_and_visualize(SIGNALS_PATH + \"A01505.hea\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}