Pull the ratings associated to a given MD5 checksum.

mriqc_webapi_ratings.py

from os import getenv
from pathlib import Path
from json import load
import urllib.request, json
import pandas as pd

def get_rating(checksum):
    """
    Grab the latest rating found for a given MD5 checksum.
    
    Example
    -------
    >>> get_rating('7e7e085a93dac55632cf9289650f6dd3')
    {'rating': '1', 'comment': ['head-motion', 'inu']}
    
    """
    url = 'https://mriqc.nimh.nih.gov/api/v1/rating?where={"md5sum":"%s"}'
    
    with urllib.request.urlopen(url % checksum) as conn:
        data = json.loads(conn.read().decode())
        
    try:
        element = data['_items'][-1]
    except IndexError:
        return {'rating': 'n/a', 'comment': 'n/a'}

    # Compose a pandas dataframe
    return {
        'rating': data['_items'][-1]['rating'],
        'comment': data['_items'][-1]['comment'].replace('"', '')[1:-1].split(',')
    }

def compose_table(path, modality, suffix):
    """Colate a dataframe of subject IDs, ratings and comments."""
    mriqc_files = path.glob('sub-*/%s/sub-*_%s.json' % (modality, suffix))
    qc_data = {'participant_id': [], 'rating': [], 'comment': []}
    
    for im in mriqc_files:
        sub = json.loads(im.read_text())
        rating_data = get_rating(sub['provenance']['md5sum'])

        qc_data['participant_id'].append(sub['bids_meta']['subject_id'])
        qc_data['rating'].append(rating_data['rating'])
        qc_data['comment'].append('+'.join(rating_data['comment']))
        
    return pd.DataFrame(qc_data)

if __name__ == '__main__':
    mriqc_dir = Path(getenv('OAK', '/oak/stanford/groups/russpold/data/P1000/derivatives/mriqc-0.15.0'))
    
    t1w_df = compose_table(mriqc_dir, 'anat', 'T1w')
    t1w_df.to_csv('qc_T1w.tsv', sep='\t', index=None)

    t2w_df = compose_table(mriqc_dir, 'anat', 'T2w')
    t2w_df.to_csv('qc_T2w.tsv', sep='\t', index=None)

    bold_df = compose_table(mriqc_dir, 'func', 'bold')
    bold_df.to_csv('qc_bold.tsv', sep='\t', index=None)