{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3e37f9e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# what biolink:category does my identifier represent\n",
    "# how to find the predicates used for gene to disease mappings\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "169c9e40",
   "metadata": {},
   "outputs": [],
   "source": [
    "from linkml_runtime.utils.schemaview import SchemaView\n",
    "import requests \n",
    "# note you can also use a path on a local filesystem\n",
    "view = SchemaView(\"https://raw.githubusercontent.com/biolink/biolink-model/master/biolink-model.yaml\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "54424ac5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# what biolink:category does my identifier represent?\n",
    "# id_prefixes\n",
    "# mappings\n",
    "\n",
    "element = view.get_element('disease')\n",
    "print(type(element))\n",
    "\n",
    "element = view.get_element('phenotype of')\n",
    "print(element.name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "61eeb009",
   "metadata": {},
   "outputs": [],
   "source": [
    "# find inverses of a predicate\n",
    "print(\"inverse is: \" + view.inverse(element.name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "91212ad9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# id_prefixes\n",
    "prefixed_categories = view.get_elements_applicable_by_identifier(\"DOID:4\")\n",
    "print(prefixed_categories)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "49d364e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# mappings \n",
    "\n",
    "mapped_categories = view.get_category_by_mapping('SO:0001583')\n",
    "print(mapped_categories)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "a7a0fa3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# object = 'gene'\n",
    "# object = 'disease'\n",
    "# object = 'phenotypic feature'\n",
    "object = 'sequence variant'\n",
    "\n",
    "query_prefix = f'https://www.ebi.ac.uk/ols/api/ontologies/_ontology/terms/'\n",
    "mappings = view.get_mappings(object)\n",
    "if len(mappings) == 0:\n",
    "    print(\"no exact mappings found for: \" + object)\n",
    "for exact_mapping in mappings.get('exact'):\n",
    "    url = query_prefix.replace(\"_ontology\", exact_mapping.split(\":\")[0]) + \"http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252F\" + exact_mapping.replace(\":\", \"_\")\n",
    "    response = requests.get(url)\n",
    "    if response.status_code == 200:\n",
    "        print(exact_mapping)\n",
    "        element = view.get_element(object)\n",
    "        if element.description is None:\n",
    "            print(\"biolink description is empty\")\n",
    "        else:\n",
    "            print(\"description: \" + element.description +\"\\n\")\n",
    "        if response.json().get('description') is None:\n",
    "            print (\"no definition found via OLS\")\n",
    "        else: \n",
    "            for description in response.json().get('description'):\n",
    "                print(\"OLS description: \" + description)\n",
    "    else:\n",
    "        print(exact_mapping + \": can't find any matching terms in OLS that don't return 404 errors\")\n",
    "     "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5fbbc1c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# is my element a mixin?\n",
    "\n",
    "e = view.get_element('gene or gene product')\n",
    "view.is_mixin(e.name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "5ab2f5c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# view poly hierarchy - a gene is a chemical and biological entity\n",
    "\n",
    "ancestors = view.class_ancestors('gene')\n",
    "for a in ancestors:\n",
    "    print(a)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2f9e0c55",
   "metadata": {},
   "outputs": [],
   "source": [
    "# how to find the predicates used for gene to disease mappings\n",
    "# association:   \n",
    "#  exact_mappings:\n",
    "#      - OBAN:association\n",
    "#      - rdf:Statement\n",
    "#      - owl:Axiom\n",
    "# check out the biolink-model onion again\n",
    "\n",
    "associations = [c for c in view.all_classes().keys() if view.is_relationship(c)][0:100]\n",
    "print(len(associations))\n",
    "for a in associations:\n",
    "    if a.startswith('gene'):\n",
    "        print(a)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "2779d0d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "for association in associations:\n",
    "    domain_element = view.get_element(view.induced_slot('subject', association).range)\n",
    "    range_element = view.get_element(view.induced_slot('object', association).range)\n",
    "    \n",
    "    if 'gene' in view.class_ancestors(domain_element.name) and 'disease' in view.class_ancestors(range_element.name):\n",
    "        print(association)\n",
    "        #print(view.induced_slot('subject', association))\n",
    "    if 'gene or gene product' in view.class_ancestors(domain_element.name) and 'disease' in view.class_ancestors(range_element.name):\n",
    "        print(association)\n",
    "        print(view.induced_slot('subject', association))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "04093dc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# find predicates for those associations\n",
    "# at this point, navigating the online doc might be easiest if you just want answers. \n",
    "# programatically, we can get the predicates that have equivalent domain and range constraints to find which \n",
    "# coudl be used for associations above.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c8ade22",
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}