{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hide_input": false,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "import json\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "print (\"Importing modules...\")\n",
    "import modules\n",
    "\n",
    "##################################################\n",
    "\n",
    "print (\"Reading data from disk...\", end=' ')\n",
    "sys.stdout.flush()\n",
    "\n",
    "df = pd.read_parquet('./datasets/Metal_all_20180601.parquet')\n",
    "seqs = np.array(df.sequence)\n",
    "metals = np.array(df.ligandId)\n",
    "fingerprints = np.array(df.fingerprint)\n",
    "\n",
    "print (\"Done\")\n",
    "\n",
    "##################################################\n",
    "\n",
    "print (\"Using FOFE encoder...\", end=' ')\n",
    "sys.stdout.flush()\n",
    "metal_dict = {}\n",
    "with open(\"./dictionaries/metal_dict\", 'r') as fp:\n",
    "        metal_dict = json.load(fp)\n",
    "        \n",
    "num_to_metal = {e: k for k, e in metal_dict.items()}\n",
    "print (\"Done\")\n",
    "\n",
    "##################################################\n",
    "print(\"Loading metal_predictor...\", end=' ')\n",
    "from keras.models import model_from_json\n",
    "\n",
    "json_file = open('./models/metal_predict.json', 'r')\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "\n",
    "metal_predictor = model_from_json(loaded_model_json)\n",
    "metal_predictor.load_weights(\"./models/metal_predict.h5\")\n",
    "print (\"Done\")\n",
    "\n",
    "##################################################\n",
    "factor = 2.33\n",
    "def threshold_func(y_in, factor):\n",
    "    y_out = np.zeros_like(y_in)\n",
    "    for i in range(y_in.shape[0]):\n",
    "        th= np.mean(y_in[i]) + factor * np.std(y_in[i])\n",
    "        y_out[i] = (y_in[i] > th)\n",
    "    return y_out\n",
    "\n",
    "print (\"Threshold factor set to \" + str(factor))\n",
    "print (\"--------------------------------------------------\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "choice = np.random.randint(58207)\n",
    "print (\"Choose index [\" + str(choice) + \"]\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "hide_input": false
   },
   "outputs": [],
   "source": [
    "print (\"--------------------------------------------------\")\n",
    "if len(seqs[choice]) > 60:\n",
    "    print (\"The seuqnce is [\" + seqs[choice][:30] + \"...\" + seqs[choice][len(seqs[choice])-30:] + \"]\\n\")\n",
    "else:\n",
    "    print (\"The seuqnce is [\" + seqs[choice] + \"]\\n\")\n",
    "metal_out = metal_predictor.predict(modules.FOFE(seqs[choice]))\n",
    "\n",
    "max_index = np.argmax(metal_out)\n",
    "metal = num_to_metal[max_index]\n",
    "\n",
    "print (\"This sample is binded to [\" + metal + \"]\")\n",
    "print (\"            Ground truth [\" + metals[choice] + \"]\\n\")\n",
    "\n",
    "json_file = open('./models/' + metal + '.json', 'r')\n",
    "loaded_model_json = json_file.read()\n",
    "json_file.close()\n",
    "\n",
    "MBS_predictor = model_from_json(loaded_model_json)\n",
    "MBS_predictor.load_weights('./models/' + metal + '.h5')\n",
    "\n",
    "MBS_out = MBS_predictor.predict(modules.FOFE(seqs[choice]))\n",
    "MBS_OneHot = threshold_func(MBS_out, factor)\n",
    "MBS = [np.where(e==1)[0] for e in MBS_OneHot][0]\n",
    "print (\"This sample has [\", end=\"\")\n",
    "print (*MBS, sep=\",\", end=\"\")\n",
    "print (\"] binding sites\")\n",
    "print (\"   Ground truth [\", end=\"\")\n",
    "print (*fingerprints[choice], sep=\",\", end=\"\")\n",
    "print (\"]\")"
   ]
  }
 ],
 "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": "3.6.3"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}