GRPM_01_main_dataset_build.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GRPM Dataset Builder 2.0 (LitVar2 Patch)\n",
    "\n",
    "This notebook is designed to retrieve genetic polymorphism data from multiple sources. It uses the LitVar API to extract polymorphisms for each human gene within the LitVar2 database along with all associated PubMed Identifiers (PMIDs). These PMIDs are then employed as queries on PubMed to obtain MEDLINE data (parsed though the 'nbib' package). All collected data are ultimately consolidated into a single CSV file, known as the \"GRPM Dataset\", which serves as the primary source against which MeSH term queries can be launched to retrieve genes and polymorphisms associated with specific contexts.  \n",
    "\n",
    "**Updates** : \n",
    "\n",
    "The GRPM Dataset available on Zenodo is a snapshot of [LitVar1](https://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/LitVar/help.html). LitVar1 is now <u>**deprecated**</u> and has been fully replaced by [LitVar2](https://www.ncbi.nlm.nih.gov/research/litvar2/). This module \n",
    "([Dataset Builder](https://github.com/johndef64/GRPM_system/blob/main/GRPM_01_dataset_builder.ipynb)) has been updated to retrieve data from LitVar2. The subsequent modules in the pipeline remain functional and can be tested using the original version of the GRPM Dataset available on Zenodo."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import Requirements and Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:16.401148400Z",
     "start_time": "2025-02-16T20:40:06.645117500Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Utente\\anaconda3\\envs\\newenv\\lib\\site-packages\\sentence_transformers\\cross_encoder\\CrossEncoder.py:13: TqdmExperimentalWarning: Using `tqdm.autonotebook.tqdm` in notebook mode. Use `tqdm.tqdm` instead to force console mode (e.g. in jupyter console)\n",
      "  from tqdm.autonotebook import tqdm, trange\n"
     ]
    }
   ],
   "source": [
    "import importlib\n",
    "import subprocess\n",
    "\n",
    "try:\n",
    "    importlib.import_module('pygrpm')\n",
    "except ImportError:\n",
    "    subprocess.check_call([\"pip\", \"install\", \"git+https://github.com/johndef64/GRPM_system.git\"])\n",
    "\n",
    "from pygrpm import *\n",
    "from pygrpm.data_retrieval import *\n",
    "from pygrpm.utils import *\n",
    "\n",
    "# requirements = ['nbib', 'requests', \"biopython\", \"matplotlib\", \"bs4\", \"pandas\"]\n",
    "# check_and_install_requirements(requirements)\n",
    "\n",
    "\n",
    "# Import Requirements ================\n",
    "import os\n",
    "import ast\n",
    "import time\n",
    "import nbib\n",
    "import pandas as pd\n",
    "import requests as rq\n",
    "import matplotlib.pyplot as plt\n",
    "from datetime import datetime\n",
    "from io import StringIO\n",
    "from Bio import Entrez\n",
    "\n",
    "Entrez.email = \"your_email@example.com\"\n",
    "\n",
    "request_counter = 0\n",
    "gene_counter = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "# Get required data from Zenodo"
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ensembl genes:\n",
      "protein_coding_genes_list = 19318\n",
      "RNA_genes_list            = 11452\n",
      "pseudo_genes_list         = 9866\n"
     ]
    }
   ],
   "source": [
    "if not os.path.exists('human_genes.csv'):\n",
    "    get_from_zenodo('human_genes.csv')\n",
    "\n",
    "# Load Human Gene list ---------------------------------\n",
    "h_genes= pd.read_csv('human_genes.csv', index_col=0)\n",
    "\n",
    "protein_coding_genes = h_genes[h_genes['Type code']==\"PCG\"  ].reset_index(drop=True)\n",
    "IG_TR_genes          = h_genes[h_genes['Type code']==\"IG_TR\"].reset_index(drop=True)\n",
    "RNA_genes            = h_genes[h_genes['Type code']==\"RNA\"  ].reset_index(drop=True)\n",
    "pseudo_genes         = h_genes[h_genes['Type code']==\"PSD\"  ].reset_index(drop=True)\n",
    "misc_genes           = h_genes[h_genes['Type code']==\"MISC\" ].reset_index(drop=True)\n",
    "\n",
    "#First: protein coding genes:\n",
    "protein_coding_genes_list = protein_coding_genes['Gene name'].dropna().tolist()\n",
    "RNA_genes_list            = RNA_genes['Gene name'].dropna().tolist()\n",
    "pseudo_genes_list         = pseudo_genes['Gene name'].dropna().tolist()\n",
    "\n",
    "print('Ensembl genes:')\n",
    "print('protein_coding_genes_list =',len(protein_coding_genes_list))\n",
    "print('RNA_genes_list            =',len(RNA_genes_list))\n",
    "print('pseudo_genes_list         =',len(pseudo_genes_list))"
   ],
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:22.715148400Z",
     "start_time": "2025-02-16T20:40:22.479146700Z"
    }
   },
   "execution_count": 2
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Whole genome forecast"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:25.790148Z",
     "start_time": "2025-02-16T20:40:25.726148100Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Forecast:\n",
      "max genes/day=  2352\n",
      "for 19318 genes:\n",
      "    request counter = 82101.5 requests\n",
      "    whole genome runtime = 34.13 hours\n",
      "    whole genome runtime overnight = 40.03\n",
      "    db table size 7669.25 MB\n"
     ]
    }
   ],
   "source": [
    "#Statistics based on 150 random genes:\n",
    "time_sleep = 0.4\n",
    "runtime_gene = 6.36 #sec/gene\n",
    "genes_hour = 566 #genes/hour\n",
    "request_counter_gene = 4.25 #request/gene (with base sleep (0.4))\n",
    "sleep_request_base = 0.4 #time sleep each request\n",
    "sleep_request_overnight_plus = 1.1 # for an overnight job\n",
    "\n",
    "print('Forecast:')\n",
    "max_genes = int(10000/request_counter_gene)\n",
    "table_size_db_gene = 0.496 #MB\n",
    "table_size_gene = 0.397 #MB\n",
    "png_size_db_gene = 0.47 #KB\n",
    "\n",
    "#Forecast:\n",
    "genes = protein_coding_genes_list\n",
    "nruntime = len(genes) * runtime_gene\n",
    "#print('runtime, '+str(len(genes)), nruntime)\n",
    "nrequest_counter = len(genes) * request_counter_gene\n",
    "\n",
    "tempo_ore = round(nruntime/3600, 2)\n",
    "tempo_ore_overnight = round((nruntime+(sleep_request_overnight_plus*len(genes)))/3600, 2)\n",
    "\n",
    "print('max genes/day= ',max_genes)\n",
    "print('for',str(int(len(genes))),'genes:')\n",
    "print('    request counter =', nrequest_counter,'requests')\n",
    "print('    whole genome runtime =', tempo_ore,'hours')\n",
    "print('    whole genome runtime overnight =', tempo_ore_overnight)\n",
    "\n",
    "db_table_size = len(genes) * table_size_gene\n",
    "print('    db table size', round(db_table_size,2),'MB')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Human Genes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:28.305147900Z",
     "start_time": "2025-02-16T20:40:28.243147800Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "protein_coding_genes 19318 \n",
      "IG_TR_genes 641 \n",
      "RNA_genes 11452 \n",
      "pseudo_genes 9866 \n",
      "misc_genes 22\n",
      "\n",
      "recommended job lenght for pcg: 1073\n"
     ]
    }
   ],
   "source": [
    "# Load Human Gene list ---------------------------------\n",
    "\n",
    "# create gene lists:\n",
    "protein_coding_genes_list = protein_coding_genes['Gene name'].dropna().tolist()\n",
    "rna_genes_list = RNA_genes['Gene name'].dropna().tolist()\n",
    "pseudo_genes_list = pseudo_genes['Gene name'].dropna().tolist()\n",
    "\n",
    "\n",
    "# Split job packages:----------------------------------\n",
    "\n",
    "# (1) protein coding genes:\n",
    "gene_range = int(len(protein_coding_genes_list)/18)\n",
    "genes = [protein_coding_genes_list[i * gene_range : (i + 1) * gene_range] for i in range(0, 18)]\n",
    "pcg_chunks = genes[:18]\n",
    "\n",
    "# (2) RNA genes:\n",
    "rna_gene_range = int(len(rna_genes_list)/5)\n",
    "genes = [rna_genes_list[i * rna_gene_range : (i + 1) * rna_gene_range] for i in range(0, 8)]\n",
    "rna_chunks = genes[:5]\n",
    "\n",
    "# (3) pseudo genes:\n",
    "pseudo_gene_range = int(len(pseudo_genes_list)/2)\n",
    "genes = [rna_genes_list[i * pseudo_gene_range : (i + 1) * pseudo_gene_range] for i in range(0, 8)]\n",
    "pseudo_chunks = genes[:2]\n",
    "\n",
    "print('protein_coding_genes',len(protein_coding_genes['Gene name'].dropna()),\n",
    "      '\\nIG_TR_genes',len(IG_TR_genes['Gene name'].dropna()),\n",
    "      '\\nRNA_genes',len(RNA_genes['Gene name'].dropna()),\n",
    "      '\\npseudo_genes',len(pseudo_genes['Gene name'].dropna()),\n",
    "      '\\nmisc_genes',len(misc_genes['Gene name'].dropna()))\n",
    "\n",
    "print('\\nrecommended job lenght for pcg:',int(len(protein_coding_genes_list)/18))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Set options and import building dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:33.951145700Z",
     "start_time": "2025-02-16T20:40:33.890149300Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time load 0:00:00.004001\n",
      "empty dataset\n",
      "empty dataset\n"
     ]
    }
   ],
   "source": [
    "# set main path:--------------------------------\n",
    "db_path = 'grpm_dataset/'\n",
    "if not os.path.exists(db_path):\n",
    "    os.makedirs(db_path)\n",
    "#------------------------------------------------\n",
    "\n",
    "\n",
    "#import checkpoint datasets:\n",
    "time_a = datetime.now()\n",
    "if os.path.isfile(db_path+'/grpm_dataset.parquet'):\n",
    "    complete_df = pd.read_parquet(db_path+'/grpm_dataset.parquet', engine='auto', columns=None)\n",
    "    restart = True\n",
    "else:\n",
    "    complete_df = pd.DataFrame()\n",
    "    restart = False\n",
    "\n",
    "if os.path.isfile(db_path+'/grpm_nbib.parquet'):\n",
    "    grpm_nbib = pd.read_parquet(db_path+'/grpm_nbib.parquet', engine='auto', columns=None)\n",
    "\n",
    "else:\n",
    "    grpm_nbib = pd.DataFrame()\n",
    "time_b = datetime.now()\n",
    "\n",
    "print('time load',time_b-time_a)\n",
    "\n",
    "## check saved data:\n",
    "if os.path.isfile(db_path+'/grpm_dataset.parquet'):\n",
    "    gene_db_count =  complete_df.gene.nunique()\n",
    "    print('complete_df gene count:',gene_db_count,'on', len(protein_coding_genes_list))\n",
    "    if gene_db_count >= 15519:\n",
    "        print('grpm db already contains all available genes on litvar1')\n",
    "\n",
    "    print('\\ngrpm_dataset.csv size'  ,round(os.path.getsize(db_path+'/grpm_dataset.parquet')/(1024*1024),3),'MB')\n",
    "    print('grpm_nbib.csv size',round(os.path.getsize(db_path+'/grpm_nbib.parquet')/(1024*1024),3),'MB')\n",
    "    print('memory_usage_complete_df'     ,round(complete_df.memory_usage().sum()/(1024*1024),3))\n",
    "    print('memory_usage_complete_nbib_df',round(grpm_nbib.memory_usage().sum()/(1024*1024),3))\n",
    "else:\n",
    "    print('empty dataset')\n",
    "    print('empty dataset')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run Job"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Set gene-range for this job"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:40.203145600Z",
     "start_time": "2025-02-16T20:40:36.407147800Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "sample size: 2\n",
      "saving checkpoint each 10 genes\n"
     ]
    }
   ],
   "source": [
    "# Set the gene list ----------------\n",
    "\n",
    "#1. Set gene-range [whole genome build]\n",
    "gene_chunk = pcg_chunks[0]\n",
    "            # pcg_chunks[0:17]\n",
    "            # rna_chunks[0:4]\n",
    "            # pseudo_chunks[0:1]\n",
    "\n",
    "# run a sample?\n",
    "run_sample = simple_bool('Do you want to run a chunk sample for testing?')\n",
    "if run_sample:\n",
    "    sample_size = int(input('sample size? \\nnum:'))\n",
    "    print(f'sample size: {sample_size}')\n",
    "\n",
    "\n",
    "#2. place here your custom gene list [custom build]\n",
    "if not run_sample:\n",
    "    custom_genes = ['APOA1', 'FFC1', 'ERH', 'USP53']\n",
    "    custom_list = simple_bool('Do you want to run the custom gene list?')\n",
    "else:\n",
    "    custom_list = False\n",
    "\n",
    "#if stucked, store skipped_genes and run custom list later:\n",
    "skipped_genes =  []\n",
    "\n",
    "# Set save interval checkpoint frequency\n",
    "checkpoint = 10\n",
    "print(f'saving checkpoint each {checkpoint} genes')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run Job"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:50.378145100Z",
     "start_time": "2025-02-16T20:40:42.743148100Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Start at  2025-02-16 21:40:42.746147\n",
      "ADAMTS12_runtime: 0:00:03.634034\n",
      "LILRA6_runtime: 0:00:03.634001\n",
      "saved checkpoint\n",
      "\n",
      "Job Statistics:\n",
      "gene batch: 2\n",
      "time batch: 0:00:07\n",
      "runtime/gene: 0:00:03.803499\n",
      "request_counter: 4  (limit: 10.000/day)\n",
      "requests/gene: 2.0\n",
      "2025-02-16 21:40:50.353145\n"
     ]
    }
   ],
   "source": [
    "# set options:--------------------------------\n",
    "save_plot           = False\n",
    "save_studytype_data = False\n",
    "save_accessory_data = False\n",
    "\n",
    "\n",
    "# Run Job -------------------------\n",
    "if run_sample:\n",
    "    genes = pd.Series(gene_chunk).sample(sample_size).reset_index(drop=True).to_list()\n",
    "    restart = False\n",
    "else:\n",
    "    genes = pd.Series(gene_chunk).to_list()\n",
    "\n",
    "if restart:\n",
    "    restart_from = complete_df.gene.nunique()\n",
    "    gene_start = restart_from\n",
    "    print('search restarted from '+str(restart_from))\n",
    "else:\n",
    "    gene_start = 0\n",
    "\n",
    "if custom_list:\n",
    "    genes = custom_genes\n",
    "    gene_start = 0\n",
    "\n",
    "\n",
    "#=========================================================================\n",
    "time_start = datetime.now()\n",
    "print('Start at ',time_start)\n",
    "\n",
    "for gene in genes[gene_start:]:\n",
    "\n",
    "    #LitVar2 \"Variants for Gene\" API request\n",
    "    if request_counter > 9950:\n",
    "        print('Request limit reached. Please, wait \\'till tomorrow!')\n",
    "\n",
    "    time_alpha = datetime.now()\n",
    "    url = \"https://www.ncbi.nlm.nih.gov/research/litvar2-api/variant/search/gene/\" + gene\n",
    "    response = (rq.get(url)).text\n",
    "\n",
    "    # Build Dataframe from response\n",
    "    data = \"[\" + transform_string(response) + \"]\"\n",
    "    df = pd.read_json(StringIO(data))\n",
    "\n",
    "    if 'rsid' in df.columns and len(df.rsid)>1:\n",
    "        # Creating a df without the clingen entry\n",
    "        dfb = df[['_id','pmids_count','rsid']]\n",
    "        dfa = dfb[~dfb['_id'].str.contains('@CA')].drop_duplicates().reset_index(drop=True)\n",
    "        dfn = dfa.dropna(subset=['rsid'])\n",
    "\n",
    "        #handle = Entrez.esearch(db=\"snp\", term=gene)\n",
    "        #record = Entrez.read(handle)\n",
    "        #request_counter += 1\n",
    "\n",
    "        NCBI_dbSNP = 'na' #record[\"Count\"]\n",
    "        lit2_variant = len(dfa['_id'].drop_duplicates())\n",
    "        lit2_variant_norsid = len(dfa.loc[df['rsid'].isna()])\n",
    "        lit2_rsid = len(dfn.rsid.drop_duplicates())\n",
    "\n",
    "        # remove rsid with pmid_count = 1\n",
    "        df2 = dfn.loc[df.pmids_count !=1]\n",
    "        lit2_rsid_f = len(df2)\n",
    "\n",
    "        # accessory data\n",
    "        dfsort = df.sort_values(by='pmids_count',ascending=False).reset_index(drop=True)\n",
    "        df2sort = df2.sort_values(by='pmids_count',ascending=False).reset_index(drop=True)\n",
    "      \n",
    "\n",
    "        for rsid in df2.rsid[0:1]:\n",
    "            url=\"https://www.ncbi.nlm.nih.gov/research/litvar2-api/variant/get/litvar@\"+rsid+\"%23%23/publications\"\n",
    "            response = (rq.get(url)).text\n",
    "            response_dict = ast.literal_eval(response)\n",
    "            rsid_pubs = pd.DataFrame({key: pd.Series(value) for key, value in response_dict.items()})\n",
    "            \n",
    "            #Display-------------------------------------------------------\n",
    "            pmid_mesh_spost = pd.DataFrame(rsid_pubs)\n",
    "            if 'pmids' in pmid_mesh_spost.columns and len(pmid_mesh_spost.pmids_count)>1:\n",
    "                lit1_rsid = 0\n",
    "    \n",
    "                # Creating the simple list [rsid-pmid]========================\n",
    "                rsidpmid = pmid_mesh_spost[['pmids']].copy()\n",
    "                rsidpmid['rsid'] = str(rsid)\n",
    "                rsidpmid['pmids'] = rsidpmid['pmids'].astype('str')\n",
    "                \n",
    "                #report data:\n",
    "                lit1_rsid_pmid = len(rsidpmid)\n",
    "                lit1_pmid = len(rsidpmid.drop_duplicates(subset='pmids'))\n",
    "    \n",
    "    \n",
    "                #### [MODULE: groupby.describe]\n",
    "                # applicare groupby ad rsidpmid per avere tabella pmid count\n",
    "                rsidpmidcount = rsidpmid.groupby('rsid').describe().reset_index()\n",
    "                rsidpmidcount.columns = rsidpmidcount.columns.to_flat_index()\n",
    "                new_column_names = ['rsid', 'pmid_count', 'pmid_unique','pmid_top','pmid_freq']\n",
    "                rsidpmidcount.columns = new_column_names\n",
    "                rsidpmidcountf = rsidpmidcount[['rsid','pmid_unique']]\n",
    "    \n",
    "                #report data:-------------------------------------------------------------\n",
    "                lit1_rsid_f = len(rsidpmidcountf[rsidpmidcountf.pmid_unique!=1])\n",
    "                lit1_rsid_m = len(rsidpmidcountf[rsidpmidcountf.pmid_unique==1])\n",
    "    \n",
    "                rsidpmidcountfsort = rsidpmidcountf.sort_values('pmid_unique',ascending=False).reset_index(drop=True)\n",
    "    \n",
    "    \n",
    "                #Filter pmid for rsid with pmid>1------------------------------------------\n",
    "                outless = rsidpmidcountfsort[rsidpmidcountfsort.pmid_unique>1]\n",
    "                #creare una mask isin su rsidpmid con outless.rsid\n",
    "                mask = rsidpmid['rsid'].isin(outless.rsid)\n",
    "                rsidpmidless = rsidpmid[mask]\n",
    "                lit1_pmid_f = len(rsidpmidless.pmids.drop_duplicates())\n",
    "\n",
    "\n",
    "                # PubMed Request =====================================================\n",
    "                \n",
    "                # PubMed queries Build:\n",
    "                pmid_l = rsidpmid.pmids.drop_duplicates().tolist()\n",
    "                query = build_pubmed_query(pmid_l, limit = 1300)\n",
    "                \n",
    "                ## Get PubMed data:     \n",
    "                time1 = datetime.now()\n",
    "                pages = ((len(pmid_l)//200)+1)+1\n",
    "                if len(pmid_l) % 200 == 0:\n",
    "                    pages = pages -1\n",
    "                fullnbib = str()\n",
    "                for d in query:\n",
    "                    for i in range(1, pages):\n",
    "                        nbib01 = GetPubmedData(i, d)                        \n",
    "                        fullnbib += nbib01\n",
    "                        request_counter += pages\n",
    "                        time.sleep(1.5)\n",
    "    \n",
    "                time2 = datetime.now()\n",
    "                timestamp = time2.strftime('%Y%m%d%H%M%S')\n",
    "                runtime = time2-time1\n",
    "                duration = str(runtime).split('.')[0]\n",
    "                hours, minutes, seconds = duration.split(':')\n",
    "                compact_duration = '{}:{}:{}'.format(hours, minutes, seconds)\n",
    "    \n",
    "                # nbib parsing:\n",
    "                timea = datetime.now()\n",
    "                ref = nbib.read(fullnbib)\n",
    "                dfbib = pd.DataFrame(ref)\n",
    "                if 'descriptors' in dfbib.columns and len(dfbib['descriptors'])>1:\n",
    "                    dfbibdes = dfbib[['pubmed_id','descriptors']].dropna().reset_index(drop=True)\n",
    "                    nbib_objects = len(dfbib)\n",
    "                    nbib_objects_withdescriptors = len(dfbibdes)\n",
    "                    #print('nibib objects:',nbib_objects)\n",
    "                    #print('nibib objects with descriptors:',len(dfbibdes))\n",
    "                    timeb = datetime.now()\n",
    "                    #print('runtime:', timeb-timea)\n",
    "    \n",
    "                    # Statistics:\n",
    "                    pubmed_pmid_nbib = len(dfbib.pubmed_id.drop_duplicates())\n",
    "                    pubmed_pmid_nbib_yesmesh = len(dfbibdes.pubmed_id.drop_duplicates())\n",
    "                    pubmed_pmid_nbib_nomesh = len(dfbib.pubmed_id.drop_duplicates())-len(dfbibdes.pubmed_id.drop_duplicates())\n",
    "    \n",
    "                    # refine MESH\n",
    "                    pmid_mesh = []\n",
    "                    for i in range(len(dfbibdes)):\n",
    "                        for mesh in dfbibdes['descriptors'][i]:\n",
    "                            out = dfbibdes['pubmed_id'][i], mesh\n",
    "                            pmid_mesh.append(out)\n",
    "                    MESH = pd.DataFrame(pmid_mesh).rename(columns={0: 'pmid',1:'mesh'})\n",
    "    \n",
    "                    # dataframe parsing splitting three fields\n",
    "                    MESH_split =[]\n",
    "                    for i in range(len(MESH)):\n",
    "                        mg = MESH.mesh[i].get('descriptor')\n",
    "                        mg2 = MESH.mesh[i].get('qualifier')\n",
    "                        mg3 = MESH.mesh[i].get('major')\n",
    "                        mgg = MESH.pmid[i], mg, mg2, mg3\n",
    "                        MESH_split.append(mgg)\n",
    "    \n",
    "                    dfmesh = pd.DataFrame(MESH_split).rename(columns={0: 'pmid',1:'mesh',2:'qualifier',3:'major'}).drop_duplicates()\n",
    "    \n",
    "                    # statistics\n",
    "                    pubmed_pmidmesh = len(dfmesh[['pmid','mesh']].drop_duplicates())\n",
    "                    pubmed_mesh_qualifier_major = len(MESH.mesh.drop_duplicates())\n",
    "                    pubmed_mesh = len(dfmesh.mesh.drop_duplicates())\n",
    "    \n",
    "                    pmidmesh = dfmesh[['pmid','mesh']].drop_duplicates()\n",
    "                    pmidmesh['pmid'] = pmidmesh['pmid'].astype(str) #convert pmid type in str\n",
    "    \n",
    "    \n",
    "                    #Analyze enrichment with groupby.describe method-------------------------------\n",
    "                    #Add rsid column con merge\n",
    "                    rsidpmid = rsidpmid.rename(columns={\"pmids\": \"pmid\"})\n",
    "                    rspmidmesh_merge = pd.merge(pmidmesh, rsidpmid, on= 'pmid', how='inner').drop_duplicates().reindex(columns=['pmid', 'rsid', 'mesh'])\n",
    "                    #rspmidmesh_merge['pmid'] = rspmidmesh_merge['pmid'].astype(str)\n",
    "    \n",
    "                    ### groupby.describe analysis by mesh\n",
    "                    meshrspmidmerge_count = rspmidmesh_merge.groupby('mesh').describe().reset_index()\n",
    "                    meshrspmidmerge_count.columns = meshrspmidmerge_count.columns.to_flat_index()\n",
    "                    #to handle generate df.groupby.describe, convert Multicolumn to single column https://datascientyst.com/flatten-multiindex-in-pandas/\n",
    "                    new_column_names = ['mesh', 'pmid-count', 'pmid-unique','pmid-top','pmid-freq','rsid-count', 'rsid-unique','rsid-top','rsid-freq']\n",
    "                    meshrspmidmerge_count.columns = new_column_names\n",
    "    \n",
    "                    meshrspmidmerge_count_short = meshrspmidmerge_count[['mesh','pmid-unique','rsid-unique']]\n",
    "\n",
    "    \n",
    "                    # add frequency\n",
    "                    totalpmid_count = len(pmidmesh.pmid.drop_duplicates())\n",
    "                    meshrspmidmerge_count_short_freq = meshrspmidmerge_count_short.copy()\n",
    "                    meshb_frq = meshrspmidmerge_count_short_freq.loc[:,'pmid-unique'].astype(float)/totalpmid_count\n",
    "                    meshrspmidmerge_count_short_freq.loc[:,'mesh frequency'] = round(meshb_frq,3)#*100\n",
    "                    meshrspmidmerge_count_short_freq_sort = meshrspmidmerge_count_short_freq.sort_values(by='pmid-unique',ascending=False).reset_index(drop=True)\n",
    "    \n",
    "                    top10mesh_all = meshrspmidmerge_count_short_freq_sort['mesh'][:10].tolist()\n",
    "                    #display(meshrspmidmerge_count_short_freq_sort.head(20))\n",
    "    \n",
    "                    ### groupby.describe analysis by rsid------------------\n",
    "                    rspmidmeshmerge_count = rspmidmesh_merge.groupby('rsid').describe().reset_index()\n",
    "                    rspmidmeshmerge_count.columns = rspmidmeshmerge_count.columns.to_flat_index()\n",
    "                    new_column_names = ['rsid', 'pmid-count', 'pmid-unique','pmid-top','pmid-freq','mesh-count', 'mesh-unique','mesh-top','mesh-freq']\n",
    "                    rspmidmeshmerge_count.columns = new_column_names\n",
    "    \n",
    "                    rsid_pmid10 = len(rspmidmeshmerge_count[rspmidmeshmerge_count['pmid-unique']>10])\n",
    "                    rsid_pmid50 = len(rspmidmeshmerge_count[rspmidmeshmerge_count['pmid-unique']>50])\n",
    "                    rsid_pmid100 = len(rspmidmeshmerge_count[rspmidmeshmerge_count['pmid-unique']>100])\n",
    "                                        \n",
    "    \n",
    "                    rspmidmeshmerge_count_short = rspmidmeshmerge_count[['rsid','pmid-unique','mesh-unique']]\n",
    "                    rspmidmeshmerge_count_short_sort = rspmidmeshmerge_count_short.sort_values(by='pmid-unique', ascending= False).reset_index(drop=True)\n",
    "                    top10rsid_all = rspmidmeshmerge_count_short_sort['rsid'].iloc[:10].tolist()\n",
    "\n",
    "                    # create a scatter plot-----------------------------------------\n",
    "                    if save_plot:\n",
    "                        plot_retrieval_data(meshrspmidmerge_count_short_freq_sort, timestamp, db_path, gene)\n",
    " \n",
    "                    # GET STUDY TYPE from NBIB----------------------------------------\n",
    "                    if save_studytype_data:\n",
    "                        get_studytype_data(ref, gene, db_path)\n",
    "    \n",
    "    \n",
    "                    #SAVE DATA========================================================\n",
    "                    timestamp = time2.strftime('%Y%m%d%H%M%S')\n",
    "                    # save accessory data:\n",
    "                    if save_accessory_data:\n",
    "                        dfsort[[\"_id\",\"rsid\",\"pmids_count\"]].to_csv(db_path+'/'+gene+'_litvar2_variants4gene.csv')\n",
    "                        rsidpmid.to_csv(db_path+'/'+gene+'_litvar1_rsids2pmids.csv') #lit1 [rsid-pmid]\n",
    "                        #rsidpmidcountfsort #lit1 pmid count\n",
    "    \n",
    "                        meshrspmidmerge_count_short_freq_sort.to_csv(db_path+'/'+gene+'_mesh_pmidrsid_count.csv')\n",
    "    \n",
    "                    #complete_df with concat:\n",
    "                    dfmesh['pmid'] = dfmesh['pmid'].astype(str)\n",
    "                    rsidpmid['pmid'] = rsidpmid['pmid'].astype(str)\n",
    "    \n",
    "                    # add a rsid-merger to dfmesh\n",
    "                    gene_rsidpmidmesh = pd.merge(rsidpmid, dfmesh, on='pmid')\n",
    "                    gene_rsidpmidmesh['gene'] = gene\n",
    "    \n",
    "                    gene_df = pd.DataFrame(gene_rsidpmidmesh)\n",
    "                    complete_df = pd.concat([complete_df, gene_rsidpmidmesh])\n",
    "    \n",
    "                    #grpm_nbib with concat:\n",
    "                    dfbib['gene'] = gene\n",
    "                    grpm_nbib = pd.concat([grpm_nbib, dfbib])\n",
    "    \n",
    "                    # save checkpoint----------------------\n",
    "                    if genes.index(gene) > 1 and genes.index(gene) % checkpoint == 0:\n",
    "                        save_checkpoint(complete_df, grpm_nbib, db_path)\n",
    "                    else:\n",
    "                        pass\n",
    "\n",
    "    \n",
    "                    # Build REPORT-------------------------------------------------\n",
    "                    time_omega = datetime.now()\n",
    "                    full_runtime = time_omega - time_alpha\n",
    "                    print(gene + '_runtime:', full_runtime)\n",
    "                    nbib_seconds = runtime.total_seconds()\n",
    "                    total_seconds = full_runtime.total_seconds()\n",
    "                    full_runtime_str = str(full_runtime).split('.')[0]\n",
    "    \n",
    "                    report = {'type': h_genes.loc[h_genes['Gene name'] == gene, 'Type code'].values[0],\n",
    "                              'ncbi_dbsnp': NCBI_dbSNP,\n",
    "                              'lit2_variant': lit2_variant,\n",
    "                              'lit2_variant_norsid': lit2_variant_norsid,\n",
    "                              'lit2_rsid': lit2_rsid,\n",
    "                              'lit2_rsid_plus1': lit2_rsid_f,\n",
    "                              'lit1_rsid': lit1_rsid,\n",
    "                              #'lit1_raw_pmid': lit1_raw_pmid,\n",
    "                              #'lit1_rsid_pmid': lit1_rsid_pmid,\n",
    "                              'lit1_rsid_pmid_plus1': lit1_rsid_f,\n",
    "                              #lit1_rsid_pmid=1': lit1_rsid_m,\n",
    "                              'lit1_pmid': lit1_pmid,\n",
    "                              'lit1_pmid_pmid_plus1': lit1_pmid_f,\n",
    "                              'pubmed_pmid_query': len(pmid_l),\n",
    "                              'nbib_objects': nbib_objects,\n",
    "                              'nbib_objects_withdescriptors': nbib_objects_withdescriptors,\n",
    "                              'pubmed_pmid': pubmed_pmid_nbib,\n",
    "                              'pubmed_pmid_withmesh': pubmed_pmid_nbib_yesmesh,\n",
    "                              #'pubmed_pmid_nomesh':pubmed_pmid_nbib_nomesh,\n",
    "                              'pubmed_pmidmesh': pubmed_pmidmesh,\n",
    "                              'pubmed_mesh_qualifier_major': pubmed_mesh_qualifier_major,\n",
    "                              'pubmed_mesh': pubmed_mesh,\n",
    "                              'rsid_pmid10': rsid_pmid10,\n",
    "                              'rsid_pmid50': rsid_pmid50,\n",
    "                              'rsid_pmid100': rsid_pmid100,\n",
    "                              'top10mesh_all': str(top10mesh_all),\n",
    "                              'top10rsid_all': str(top10rsid_all),\n",
    "                              'pubmed_runtime': duration,\n",
    "                              'total_runtime': full_runtime_str,\n",
    "                              'time_stamp': time2\n",
    "                              }\n",
    "    \n",
    "                    df_report = pd.DataFrame(report, index=[gene])#.transpose()\n",
    "    \n",
    "                    # generate fist report.csv\n",
    "                    if os.path.isfile(db_path+'/grpm_report.parquet'):\n",
    "                        #dfL = pd.read_csv(db_path+'/grpm_report.csv', index_col=0)\n",
    "                        dfL = pd.read_parquet(db_path+'/grpm_report.parquet')#, index_col=0)\n",
    "                        dfL = pd.concat([dfL, df_report])#, axis=1)\n",
    "                        dfL.to_parquet(db_path+'/grpm_report.parquet')\n",
    "                    else:\n",
    "                        df_report.to_parquet(db_path+'/grpm_report.parquet')  # solo la prima volta\n",
    "    \n",
    "                    #Update gene values\n",
    "                    #grpm_report = pd.read_csv(db_path+'/grpm_report.csv', index_col=0)\n",
    "                    grpm_report =  pd.read_parquet(db_path+'/grpm_report.parquet', engine='pyarrow')\n",
    "                    if gene + '.1' in grpm_report.columns:\n",
    "                        grpm_report = grpm_report.drop(columns=gene)\n",
    "                        grpm_report = grpm_report.rename(columns={gene + '.1': gene})\n",
    "                        grpm_report.to_csv(db_path+'/grpm_report.csv')\n",
    "                        print(gene,'already in db')\n",
    "                else:\n",
    "                    print(gene + ' no descriptors in NBIB')\n",
    "                    time.sleep(0.8)\n",
    "                    pass\n",
    "            else:\n",
    "                print(gene + ' no results on litvar2 (rsid2pmids)')\n",
    "                time.sleep(0.8)\n",
    "                pass\n",
    "    else:\n",
    "        print(gene + ' no results on litvar2 (gene2pmidcount)')\n",
    "        pass\n",
    "\n",
    "    if request_counter > 9000:\n",
    "        dada = 2\n",
    "        #print('Allert! Reaching pubmed request limit')\n",
    "    if request_counter > 9950:\n",
    "        #print('Request limit reached. Wait \\'till tomorrow!')\n",
    "        time_finish = datetime.now()\n",
    "        time_batch = time_finish - time_start\n",
    "        time_batch_str = str(time_batch).split('.')[0]\n",
    "        #print('time batch:', time_batch_str)\n",
    "        #break\n",
    "\n",
    "# Save Data\n",
    "save_checkpoint(complete_df, grpm_nbib, db_path)\n",
    "\n",
    "# Print Job Statistics\n",
    "time_finish = datetime.now()\n",
    "time_batch = time_finish - time_start\n",
    "time_batch_str = str(time_batch).split('.')[0]\n",
    "print('\\nJob Statistics:')\n",
    "print('gene batch:', len(genes))\n",
    "print('time batch:', time_batch_str)\n",
    "print('runtime/gene:', time_batch/len(genes))\n",
    "print('request_counter:', request_counter,' (limit: 10.000/day)')\n",
    "gene_counter += len(genes)\n",
    "print('requests/gene:', request_counter/gene_counter)\n",
    "print(time_finish)\n",
    "\n",
    "### notes:\n",
    "# LIMITS PubMed Programming Utilities (PMU)\n",
    "# 10 requests/second\n",
    "# 10,000 requests/day"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:40:52.831145900Z",
     "start_time": "2025-02-16T20:40:52.709148400Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GRPM report:\n"
     ]
    },
    {
     "data": {
      "text/plain": "      index type ncbi_dbsnp  lit2_variant  lit2_variant_norsid  lit2_rsid  \\\n0  ADAMTS12  PCG         na           954                    4        943   \n1    LILRA6  PCG         na           292                    1        291   \n\n   lit2_rsid_plus1  lit1_rsid  lit1_rsid_pmid_plus1  lit1_pmid  ...  \\\n0              346          0                     1          2  ...   \n1              132          0                     1         14  ...   \n\n   pubmed_mesh_qualifier_major  pubmed_mesh  rsid_pmid10  rsid_pmid50  \\\n0                           33           28            0            0   \n1                          132          106            1            0   \n\n   rsid_pmid100                                      top10mesh_all  \\\n0             0  ['Genetic Variation', 'Humans', 'Age Factors',...   \n1             0  ['Humans', 'Mutation', 'Male', 'High-Throughpu...   \n\n   top10rsid_all  pubmed_runtime  total_runtime                 time_stamp  \n0   ['rs997598']         0:00:01        0:00:03 2025-02-16 21:40:46.286157  \n1  ['rs8105096']         0:00:02        0:00:03 2025-02-16 21:40:49.770180  \n\n[2 rows x 27 columns]",
      "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></th>\n      <th>index</th>\n      <th>type</th>\n      <th>ncbi_dbsnp</th>\n      <th>lit2_variant</th>\n      <th>lit2_variant_norsid</th>\n      <th>lit2_rsid</th>\n      <th>lit2_rsid_plus1</th>\n      <th>lit1_rsid</th>\n      <th>lit1_rsid_pmid_plus1</th>\n      <th>lit1_pmid</th>\n      <th>...</th>\n      <th>pubmed_mesh_qualifier_major</th>\n      <th>pubmed_mesh</th>\n      <th>rsid_pmid10</th>\n      <th>rsid_pmid50</th>\n      <th>rsid_pmid100</th>\n      <th>top10mesh_all</th>\n      <th>top10rsid_all</th>\n      <th>pubmed_runtime</th>\n      <th>total_runtime</th>\n      <th>time_stamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>ADAMTS12</td>\n      <td>PCG</td>\n      <td>na</td>\n      <td>954</td>\n      <td>4</td>\n      <td>943</td>\n      <td>346</td>\n      <td>0</td>\n      <td>1</td>\n      <td>2</td>\n      <td>...</td>\n      <td>33</td>\n      <td>28</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>['Genetic Variation', 'Humans', 'Age Factors',...</td>\n      <td>['rs997598']</td>\n      <td>0:00:01</td>\n      <td>0:00:03</td>\n      <td>2025-02-16 21:40:46.286157</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>LILRA6</td>\n      <td>PCG</td>\n      <td>na</td>\n      <td>292</td>\n      <td>1</td>\n      <td>291</td>\n      <td>132</td>\n      <td>0</td>\n      <td>1</td>\n      <td>14</td>\n      <td>...</td>\n      <td>132</td>\n      <td>106</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>['Humans', 'Mutation', 'Male', 'High-Throughpu...</td>\n      <td>['rs8105096']</td>\n      <td>0:00:02</td>\n      <td>0:00:03</td>\n      <td>2025-02-16 21:40:49.770180</td>\n    </tr>\n  </tbody>\n</table>\n<p>2 rows × 27 columns</p>\n</div>"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "GRPM genes gathered: 2\n",
      "\n",
      "NBIB table:\n"
     ]
    },
    {
     "data": {
      "text/plain": "       gene  pubmed_id citation_owner          nlm_status last_revision_date  \\\n0  ADAMTS12   25326100            NLM             MEDLINE         2022-12-07   \n1  ADAMTS12   24058506            NLM             MEDLINE         2021-10-21   \n0    LILRA6   37012232            NLM             MEDLINE         2024-05-15   \n1    LILRA6   29118384            NLM             MEDLINE         2022-12-07   \n2    LILRA6   36075891            NLM  PubMed-not-MEDLINE         2023-01-11   \n\n  electronic_issn linking_issn journal_volume journal_issue publication_date  \\\n0       1744-4292    1744-4292             10            10      2014 Oct 16   \n1       1932-6203    1932-6203              8             9             2013   \n0       2041-1723    2041-1723             14             1       2023 Apr 3   \n1       2045-2322    2045-2322              7             1       2017 Nov 8   \n2       2054-345X    2054-345X              9             1       2022 Sep 8   \n\n   ... revised_time accepted_time         pubmed_time        medline_time  \\\n0  ...          NaT           NaT 2014-10-19 06:00:00 2015-02-11 06:00:00   \n1  ...          NaT    2013-07-23 2013-09-24 06:00:00 2014-06-04 06:00:00   \n0  ...          NaT    2023-03-20 2023-04-04 06:00:00 2023-04-05 06:42:00   \n1  ...          NaT    2017-10-23 2017-11-10 06:00:00 2019-07-19 06:00:00   \n2  ...   2022-07-18    2022-07-20 2022-09-09 06:00:00 2022-09-09 06:01:00   \n\n          entrez_time              pii                           doi  \\\n0 2014-10-19 06:00:00             None         10.15252/msb.20145114   \n1 2013-09-24 06:00:00  PONE-D-13-18467  10.1371/journal.pone.0073950   \n0 2023-04-03 23:13:00            37547    10.1038/s41467-023-37547-0   \n1 2017-11-10 06:00:00            15183    10.1038/s41598-017-15183-1   \n2 2022-09-08 23:14:00              208    10.1038/s41439-022-00208-7   \n\n  publication_status print_issn   pages  \n0           epublish  1744-4292     754  \n1           epublish       None  e73950  \n0           epublish       None    1851  \n1           epublish       None   14992  \n2           epublish  2054-345X      31  \n\n[5 rows x 36 columns]",
      "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></th>\n      <th>gene</th>\n      <th>pubmed_id</th>\n      <th>citation_owner</th>\n      <th>nlm_status</th>\n      <th>last_revision_date</th>\n      <th>electronic_issn</th>\n      <th>linking_issn</th>\n      <th>journal_volume</th>\n      <th>journal_issue</th>\n      <th>publication_date</th>\n      <th>...</th>\n      <th>revised_time</th>\n      <th>accepted_time</th>\n      <th>pubmed_time</th>\n      <th>medline_time</th>\n      <th>entrez_time</th>\n      <th>pii</th>\n      <th>doi</th>\n      <th>publication_status</th>\n      <th>print_issn</th>\n      <th>pages</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>ADAMTS12</td>\n      <td>25326100</td>\n      <td>NLM</td>\n      <td>MEDLINE</td>\n      <td>2022-12-07</td>\n      <td>1744-4292</td>\n      <td>1744-4292</td>\n      <td>10</td>\n      <td>10</td>\n      <td>2014 Oct 16</td>\n      <td>...</td>\n      <td>NaT</td>\n      <td>NaT</td>\n      <td>2014-10-19 06:00:00</td>\n      <td>2015-02-11 06:00:00</td>\n      <td>2014-10-19 06:00:00</td>\n      <td>None</td>\n      <td>10.15252/msb.20145114</td>\n      <td>epublish</td>\n      <td>1744-4292</td>\n      <td>754</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>ADAMTS12</td>\n      <td>24058506</td>\n      <td>NLM</td>\n      <td>MEDLINE</td>\n      <td>2021-10-21</td>\n      <td>1932-6203</td>\n      <td>1932-6203</td>\n      <td>8</td>\n      <td>9</td>\n      <td>2013</td>\n      <td>...</td>\n      <td>NaT</td>\n      <td>2013-07-23</td>\n      <td>2013-09-24 06:00:00</td>\n      <td>2014-06-04 06:00:00</td>\n      <td>2013-09-24 06:00:00</td>\n      <td>PONE-D-13-18467</td>\n      <td>10.1371/journal.pone.0073950</td>\n      <td>epublish</td>\n      <td>None</td>\n      <td>e73950</td>\n    </tr>\n    <tr>\n      <th>0</th>\n      <td>LILRA6</td>\n      <td>37012232</td>\n      <td>NLM</td>\n      <td>MEDLINE</td>\n      <td>2024-05-15</td>\n      <td>2041-1723</td>\n      <td>2041-1723</td>\n      <td>14</td>\n      <td>1</td>\n      <td>2023 Apr 3</td>\n      <td>...</td>\n      <td>NaT</td>\n      <td>2023-03-20</td>\n      <td>2023-04-04 06:00:00</td>\n      <td>2023-04-05 06:42:00</td>\n      <td>2023-04-03 23:13:00</td>\n      <td>37547</td>\n      <td>10.1038/s41467-023-37547-0</td>\n      <td>epublish</td>\n      <td>None</td>\n      <td>1851</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>LILRA6</td>\n      <td>29118384</td>\n      <td>NLM</td>\n      <td>MEDLINE</td>\n      <td>2022-12-07</td>\n      <td>2045-2322</td>\n      <td>2045-2322</td>\n      <td>7</td>\n      <td>1</td>\n      <td>2017 Nov 8</td>\n      <td>...</td>\n      <td>NaT</td>\n      <td>2017-10-23</td>\n      <td>2017-11-10 06:00:00</td>\n      <td>2019-07-19 06:00:00</td>\n      <td>2017-11-10 06:00:00</td>\n      <td>15183</td>\n      <td>10.1038/s41598-017-15183-1</td>\n      <td>epublish</td>\n      <td>None</td>\n      <td>14992</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>LILRA6</td>\n      <td>36075891</td>\n      <td>NLM</td>\n      <td>PubMed-not-MEDLINE</td>\n      <td>2023-01-11</td>\n      <td>2054-345X</td>\n      <td>2054-345X</td>\n      <td>9</td>\n      <td>1</td>\n      <td>2022 Sep 8</td>\n      <td>...</td>\n      <td>2022-07-18</td>\n      <td>2022-07-20</td>\n      <td>2022-09-09 06:00:00</td>\n      <td>2022-09-09 06:01:00</td>\n      <td>2022-09-08 23:14:00</td>\n      <td>208</td>\n      <td>10.1038/s41439-022-00208-7</td>\n      <td>epublish</td>\n      <td>2054-345X</td>\n      <td>31</td>\n    </tr>\n  </tbody>\n</table>\n<p>5 rows × 36 columns</p>\n</div>"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('GRPM report:')\n",
    "report = pd.read_parquet(db_path+'/grpm_report.parquet', engine='pyarrow').reset_index()\n",
    "display(report.head())\n",
    "dataset = pd.read_parquet(db_path+'/grpm_dataset.parquet', engine='pyarrow')\n",
    "\n",
    "print('\\nGRPM genes gathered:', len(dataset.gene.drop_duplicates()))\n",
    "#display(dataset.gene.drop_duplicates())\n",
    "\n",
    "print('\\nNBIB table:')\n",
    "nbib_data = pd.read_parquet(db_path+'/grpm_nbib.parquet', engine='pyarrow' )\n",
    "display(nbib_data.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Show saved report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T20:41:01.484145400Z",
     "start_time": "2025-02-16T20:41:01.426146300Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "      index type ncbi_dbsnp  lit2_variant  lit2_variant_norsid  lit2_rsid  \\\n1    LILRA6  PCG         na           292                    1        291   \n0  ADAMTS12  PCG         na           954                    4        943   \n\n   lit2_rsid_plus1  lit1_rsid  lit1_rsid_pmid_plus1  lit1_pmid  ...  \\\n1              132          0                     1         14  ...   \n0              346          0                     1          2  ...   \n\n   pubmed_mesh_qualifier_major  pubmed_mesh  rsid_pmid10  rsid_pmid50  \\\n1                          132          106            1            0   \n0                           33           28            0            0   \n\n   rsid_pmid100                                      top10mesh_all  \\\n1             0  ['Humans', 'Mutation', 'Male', 'High-Throughpu...   \n0             0  ['Genetic Variation', 'Humans', 'Age Factors',...   \n\n   top10rsid_all  pubmed_runtime  total_runtime                 time_stamp  \n1  ['rs8105096']         0:00:02        0:00:03 2025-02-16 21:40:49.770180  \n0   ['rs997598']         0:00:01        0:00:03 2025-02-16 21:40:46.286157  \n\n[2 rows x 27 columns]",
      "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></th>\n      <th>index</th>\n      <th>type</th>\n      <th>ncbi_dbsnp</th>\n      <th>lit2_variant</th>\n      <th>lit2_variant_norsid</th>\n      <th>lit2_rsid</th>\n      <th>lit2_rsid_plus1</th>\n      <th>lit1_rsid</th>\n      <th>lit1_rsid_pmid_plus1</th>\n      <th>lit1_pmid</th>\n      <th>...</th>\n      <th>pubmed_mesh_qualifier_major</th>\n      <th>pubmed_mesh</th>\n      <th>rsid_pmid10</th>\n      <th>rsid_pmid50</th>\n      <th>rsid_pmid100</th>\n      <th>top10mesh_all</th>\n      <th>top10rsid_all</th>\n      <th>pubmed_runtime</th>\n      <th>total_runtime</th>\n      <th>time_stamp</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1</th>\n      <td>LILRA6</td>\n      <td>PCG</td>\n      <td>na</td>\n      <td>292</td>\n      <td>1</td>\n      <td>291</td>\n      <td>132</td>\n      <td>0</td>\n      <td>1</td>\n      <td>14</td>\n      <td>...</td>\n      <td>132</td>\n      <td>106</td>\n      <td>1</td>\n      <td>0</td>\n      <td>0</td>\n      <td>['Humans', 'Mutation', 'Male', 'High-Throughpu...</td>\n      <td>['rs8105096']</td>\n      <td>0:00:02</td>\n      <td>0:00:03</td>\n      <td>2025-02-16 21:40:49.770180</td>\n    </tr>\n    <tr>\n      <th>0</th>\n      <td>ADAMTS12</td>\n      <td>PCG</td>\n      <td>na</td>\n      <td>954</td>\n      <td>4</td>\n      <td>943</td>\n      <td>346</td>\n      <td>0</td>\n      <td>1</td>\n      <td>2</td>\n      <td>...</td>\n      <td>33</td>\n      <td>28</td>\n      <td>0</td>\n      <td>0</td>\n      <td>0</td>\n      <td>['Genetic Variation', 'Humans', 'Age Factors',...</td>\n      <td>['rs997598']</td>\n      <td>0:00:01</td>\n      <td>0:00:03</td>\n      <td>2025-02-16 21:40:46.286157</td>\n    </tr>\n  </tbody>\n</table>\n<p>2 rows × 27 columns</p>\n</div>"
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Visualize grpm_report.csv\n",
    "grpm_report = pd.read_parquet(db_path+'/grpm_report.parquet', engine='pyarrow').reset_index()\n",
    "\n",
    "# Sort Values\n",
    "grpm_report.sort_values(by='lit1_pmid',ascending = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2025-02-16T19:22:02.658309400Z",
     "start_time": "2025-02-16T19:22:01.394310400Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 400x800 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show Bar Graph\n",
    "grpm_report_sort = grpm_report.sort_values(by= 'pubmed_pmid',ascending=False)\n",
    "\n",
    "x = grpm_report_sort['index'].iloc[:40]\n",
    "y = grpm_report_sort['pubmed_pmid'].iloc[:40]\n",
    "plt.figure(figsize=(4, 8))\n",
    "plt.title('PMIDs in Dataset', loc='center',pad=10)\n",
    "\n",
    "plt.barh(x,y)\n",
    "plt.gca().invert_yaxis()\n",
    "plt.tick_params(axis='x', which='both', top=True, bottom=False, labeltop=True, labelbottom=False)\n",
    "#plt.xlabel('pmid count', position=(0.5, 1.08))\n",
    "plt.ylabel('gene')\n",
    "plt.xlabel('pmid count', position=(0.5, 1.08))\n",
    "ax = plt.gca()\n",
    "ax.xaxis.set_label_position('top')\n",
    "#plt.savefig('PMID_filtered.png',dpi=300, bbox_inches = \"tight\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Debugging"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## debugging: litvar data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2024-03-14T10:52:32.895096Z"
    }
   },
   "outputs": [],
   "source": [
    "# if stucked check 'data' variable:\n",
    "# replace manually malformed lines\n",
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## debugging: nbib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2024-03-14T10:52:32.897093200Z"
    }
   },
   "outputs": [],
   "source": [
    "#NBIB PROBLEM SOLVER---------------------------------\n",
    "# replace malformed lines\n",
    "fullnbib= fullnbib.replace('2007/09/31','2007/09/30') # <= some dates are mispelled in pubmed\n",
    "ref = nbib.read(fullnbib)\n",
    "dfbib = pd.DataFrame(ref)\n",
    "dfbib"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "NBIB PROBLEM SOLVER [History] ---------------------------------\n",
    "with open('nbib report '+gene+'.txt', 'w', encoding='utf-8') as file:\n",
    "    file.write(fullnbib)\n",
    "with open('nbib report '+gene+'_FIXED.txt', 'r', encoding='utf-8') as file:\n",
    "    fullnbib = file.read() # --> not work"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# extras"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Eutils: get study type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2024-03-14T10:52:32.899094600Z"
    }
   },
   "outputs": [],
   "source": [
    "check_and_install_module('biopython')\n",
    "from Bio import Entrez\n",
    "Entrez.email = \"your_email@example.com\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "start_time": "2024-03-14T10:52:32.901094700Z"
    }
   },
   "outputs": [],
   "source": [
    "### EUTILS GET STUDY TYPE MODULE\n",
    "#https://biopython.org/docs/1.76/api/Bio.Entrez.html\n",
    "def get_study_type(pmids):\n",
    "    Entrez.email = 'your_email@your_domain.com'\n",
    "    handle = Entrez.esummary(db='pubmed', id=','.join(pmids), retmode='xml')\n",
    "    records = Entrez.parse(handle)\n",
    "    study_types = []\n",
    "    for record in records:\n",
    "        article_types = record['PubTypeList']\n",
    "        if 'Randomized Controlled Trial' in article_types:\n",
    "            study_types.append('Randomized Controlled Trial')\n",
    "        elif 'Controlled Clinical Trial' in article_types:\n",
    "            study_types.append('Controlled Clinical Trial')\n",
    "        elif 'Cohort Studies' in article_types:\n",
    "            study_types.append('Cohort Study')\n",
    "        elif 'Case-Control Studies' in article_types:\n",
    "            study_types.append('Case-Control Study')\n",
    "        elif 'Review' in article_types:\n",
    "            study_types.append('Review')\n",
    "        elif 'Clinical Trial' in article_types:\n",
    "            study_types.append('Clinical Trial')\n",
    "        elif 'Meta-Analysis' in article_types:\n",
    "            study_types.append('Meta-Analysis')\n",
    "        elif 'Multicenter Study' in article_types:\n",
    "            study_types.append('Multicenter Study')\n",
    "        else:\n",
    "            study_types.append('Unknown')\n",
    "    return study_types\n",
    "\n",
    "pmidlist = list(pmidmesh['pmid'].drop_duplicates())\n",
    "genepmids_str = list(map(str, pmidlist))\n",
    "study_type = get_study_type(genepmids_str)\n",
    "pmids_studytype = pd.DataFrame(list(zip(genepmids_str, study_type)), columns=[gene + '_PMID', 'study type'])\n",
    "request_counter += 1\n",
    "\n",
    "#study type count:\n",
    "pmids_studytype_count = pmids_studytype.groupby('study type').describe().reset_index()\n",
    "pmids_studytype_count.columns = pmids_studytype_count.columns.to_flat_index()\n",
    "new_column_names = ['study_type', 'pmid-count', 'pmid-unique', 'pmid-top', 'pmid-freq']\n",
    "pmids_studytype_count.columns = new_column_names\n",
    "pmids_studytype_countsort = pmids_studytype_count.sort_values(by='pmid-count', ascending=False)"
   ]
  }
 ],
 "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}