id-to-qid.py

"""
If query.sparql and PMCID_list.txt are downloaded, the script can simply be called as

python iq-to-qid.py
"""

import argparse
import os
import subprocess
import time
import pandas as pd

import requests
from requests.exceptions import HTTPError
from tqdm import tqdm

HEADERS = {"User-Agent": "ID-to-QID"}


def main(
    test: bool = False,
    inputFile: str = "PMCID_list.txt",
    batchSize: int = 100,
):
    """From a list of IDs, go through and sequentially output a list of their properties and values."""

    if not os.path.isfile(inputFile):
        print(inputFile + " does not exist")
        exit()

    with open("query.sparql") as r:
        query = r.readlines()
        query = "".join(query)

    with open(inputFile) as f:
        inputList = [line.strip() for line in f]

    data = []

    t = tqdm(
        total=round(len(inputList) / 100) + 1, bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt} {elapsed_s:.0f}s"
    )

    for i in range(0, len(inputList), batchSize):
        IDs = inputList[i:i+batchSize]

        t.update()
        t.set_description(IDs[0])

        IDstring = " ".join(["'" + q + "'" for q in IDs])

        data += getData(query, IDstring)

        if test:
            break

    df = pd.DataFrame(data, columns =['qid', 'pmc'])
    df.to_csv("qids-pmcs.csv", index=False)

    t.close()


def getData(query, IDstring):
    data = runQuery(
        query.format(
            values=IDstring,
        )
    )

    output = []

    for item in data["results"]["bindings"]:
        QID = item["item"]["value"][31:]
        id = item["id"]["value"]
        output.append([QID, id])

    return output


def runQuery(query):
    url = "https://query.wikidata.org/sparql"
    params = {"query": query, "format": "json"}
    try:
        response = requests.get(url, params=params, headers=HEADERS)
        return response.json()
    except HTTPError as e:
        print(response.text)
        print(e.response.text)
        exit()


def timer(tick, msg=""):
    print("--- %s %.3f seconds ---" % (msg, time.time() - tick))
    return time.time()


def defineArgParser():
    """Creates parser for command line arguments"""
    parser = argparse.ArgumentParser(
        description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter
    )

    parser.add_argument(
        "-t",
        "--test",
        help="",
        action="store_true",
    )

    return parser


if __name__ == "__main__":

    argParser = defineArgParser()
    clArgs = argParser.parse_args()

    tick = time.time()
    main(
        test=clArgs.test,
    )
    timer(tick)

PMCID_list.txt

6042825
6628208
3228863
4111116
7718173
4207244
1008230
6800166
3750528
3962008
5082607
3685583
3430699
3722524
4760270
5127930
6396093
5105880
6345265
4752027
4826505
4497953
2752617
3649846
5100142
4103489
5978649
2376761
3888149
5266326
2971582
2844990
3413384
5105883
5996309
4708103
6717989
4423409
5726607
2971575
6013758
4380025
5860209
4909101
2375448
3065683
3521214
2530886
4141640
5638226
3659198
6492420
6707396
2910028
3967109
7422831
4410647
5370611
2796819
4277354
5624584
3009535
4426829
5089131
4816032
4271147
4176421
6607905
2572702
4183485
4287948
4103495
3262844
6053696
403769
4133763
4168749
6306265
7584444
5472867
2982160
4236746
5333177
4264639
3654709
5167065
2935447
4642848
4168754
3835703
3599859
3582273
4653389
6880260
6039917
3646686
4174321
2693143
4224201
3651407
4670004
5588246
4535771
6039939
6821346
2638934
6392199
4743236
2427162
4288400
3083346
3855388
4670012
3307106
226690
2666812
4411368
5998007
3307116
6203375
6051191
3031041
7194485
4499222
6013757
6974772
5260057
3514294
4828368
2958747
4706054
3530913
4642660
3190406
4197741
4184317
5618529

qids-pmcs.csv

query.sparql

SELECT ?item ?id {{
  VALUES ?id {{ {values} }}
  
  ?item wdt:P932 ?id .
  
  SERVICE wikibase:label {{ bd:serviceParam wikibase:language "en" }}
}}