preprocess_HQS.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import spacy\n",
    "import scispacy\n",
    "\n",
    "from scispacy.linking import EntityLinker\n",
    "import json\n",
    "\n",
    "nlp = spacy.load(\"en_core_sci_md\")\n",
    "\n",
    "# This line takes a while, because we have to download ~1GB of data\n",
    "# and load a large JSON file (the knowledge base). Be patient!\n",
    "# Thankfully it should be faster after the first time you use it, because\n",
    "# the downloads are cached.\n",
    "# NOTE: The resolve_abbreviations parameter is optional, and requires that\n",
    "# the AbbreviationDetector pipe has already been added to the pipeline. Adding\n",
    "# the AbbreviationDetector pipe and setting resolve_abbreviations to True means\n",
    "# that linking will only be performed on the long form of abbreviations.\n",
    "nlp.add_pipe(\"scispacy_linker\", config={\"resolve_abbreviations\": True, \"linker_name\": \"umls\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_test = []\n",
    "with open(\"test.txt\", 'r') as input:\n",
    "    for jsonObj in input:\n",
    "        patientDict = json.loads(jsonObj)\n",
    "        patientDict[\"1_medical\"] = []\n",
    "        patientDict[\"2_medical\"] = []\n",
    "        doc = nlp(patientDict[\"summary\"])\n",
    "        for ent in doc.ents:\n",
    "            linker = nlp.get_pipe(\"scispacy_linker\")\n",
    "            name = ent.text.lower()\n",
    "            if name in [\"information\", \"treatment\", \"treatments\"] or len(ent._.kb_ents) == 0:\n",
    "                continue\n",
    "            \n",
    "            if linker.kb.cui_to_entity[ent._.kb_ents[0][0]][1].lower() == name:\n",
    "                if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                    patientDict[\"2_medical\"].append(ent.text)\n",
    "                else:\n",
    "                    patientDict[\"1_medical\"].append(ent.text)\n",
    "            else:\n",
    "                aliases_set = set()\n",
    "                for e in linker.kb.cui_to_entity[ent._.kb_ents[0][0]].aliases:\n",
    "                    aliases_set.add(e.lower())\n",
    "                \n",
    "                if name in aliases_set:\n",
    "                    if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                        patientDict[\"2_medical\"].append(ent.text)\n",
    "                    else:\n",
    "                        patientDict[\"1_medical\"].append(ent.text)\n",
    "        \n",
    "        data_test.append(patientDict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_train = []\n",
    "with open(\"train.txt\", 'r') as input:\n",
    "    for jsonObj in input:\n",
    "        patientDict = json.loads(jsonObj)\n",
    "        patientDict[\"1_medical\"] = []\n",
    "        patientDict[\"2_medical\"] = []\n",
    "        doc = nlp(patientDict[\"summary\"])\n",
    "        for ent in doc.ents:\n",
    "            linker = nlp.get_pipe(\"scispacy_linker\")\n",
    "            name = ent.text.lower()\n",
    "            if name in [\"information\", \"treatment\", \"treatments\"] or len(ent._.kb_ents) == 0:\n",
    "                continue\n",
    "            \n",
    "            if linker.kb.cui_to_entity[ent._.kb_ents[0][0]][1].lower() == name:\n",
    "                if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                    patientDict[\"2_medical\"].append(ent.text)\n",
    "                else:\n",
    "                    patientDict[\"1_medical\"].append(ent.text)\n",
    "            else:\n",
    "                aliases_set = set()\n",
    "                for e in linker.kb.cui_to_entity[ent._.kb_ents[0][0]].aliases:\n",
    "                    aliases_set.add(e.lower())\n",
    "                \n",
    "                if name in aliases_set:\n",
    "                    if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                        patientDict[\"2_medical\"].append(ent.text)\n",
    "                    else:\n",
    "                        patientDict[\"1_medical\"].append(ent.text)\n",
    "        \n",
    "        data_train.append(patientDict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_validation = []\n",
    "with open(\"validation.txt\", 'r') as input:\n",
    "    for jsonObj in input:\n",
    "        patientDict = json.loads(jsonObj)\n",
    "        patientDict[\"1_medical\"] = []\n",
    "        patientDict[\"2_medical\"] = []\n",
    "        doc = nlp(patientDict[\"summary\"])\n",
    "        for ent in doc.ents:\n",
    "            linker = nlp.get_pipe(\"scispacy_linker\")\n",
    "            name = ent.text.lower()\n",
    "            if name in [\"information\", \"treatment\", \"treatments\"] or len(ent._.kb_ents) == 0:\n",
    "                continue\n",
    "            \n",
    "            if linker.kb.cui_to_entity[ent._.kb_ents[0][0]][1].lower() == name:\n",
    "                if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                    patientDict[\"2_medical\"].append(ent.text)\n",
    "                else:\n",
    "                    patientDict[\"1_medical\"].append(ent.text)\n",
    "            else:\n",
    "                aliases_set = set()\n",
    "                for e in linker.kb.cui_to_entity[ent._.kb_ents[0][0]].aliases:\n",
    "                    aliases_set.add(e.lower())\n",
    "                \n",
    "                if name in aliases_set:\n",
    "                    if name in patientDict[\"question\"].lower() or singularize(name) in patientDict[\"question\"].lower():\n",
    "                        patientDict[\"2_medical\"].append(ent.text)\n",
    "                    else:\n",
    "                        patientDict[\"1_medical\"].append(ent.text)\n",
    "        \n",
    "        data_validation.append(patientDict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('HQS_dataset/train.txt', 'w') as file:\n",
    "    for e in data_train:\n",
    "        file.write(json.dumps(e, ensure_ascii=False))\n",
    "        file.write('\\n')\n",
    "        \n",
    "with open('HQS_dataset/validation.txt', 'w') as file:\n",
    "    for e in data_validation:\n",
    "        file.write(json.dumps(e, ensure_ascii=False))\n",
    "        file.write('\\n')\n",
    "        \n",
    "with open('HQS_dataset/test.txt', 'w') as file:\n",
    "    for e in data_test:\n",
    "        file.write(json.dumps(e, ensure_ascii=False))\n",
    "        file.write('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}