diff options
Diffstat (limited to 'python/scripts/enrich_scored_matches.py')
-rwxr-xr-x | python/scripts/enrich_scored_matches.py | 19 |
1 files changed, 11 insertions, 8 deletions
diff --git a/python/scripts/enrich_scored_matches.py b/python/scripts/enrich_scored_matches.py index 9fe1499..44c091c 100755 --- a/python/scripts/enrich_scored_matches.py +++ b/python/scripts/enrich_scored_matches.py @@ -17,29 +17,32 @@ And outputs JSON objects that are can be imported into fatcat with the No dependencies (only python3 stdlib) """ -import sys -import json import base64 +import json +import sys + def run(): for line in sys.stdin: - line = line.split('\t') + line = line.split("\t") assert len(line) == 5 - raw_sha1 = line[0].replace('sha1:', '') + raw_sha1 = line[0].replace("sha1:", "") dois = json.loads(line[1]) cdx = json.loads(line[2]) mimetype = line[3] size = int(line[4]) - sha1 = base64.b16encode(base64.b32decode(raw_sha1)).decode('ascii').lower() + sha1 = base64.b16encode(base64.b32decode(raw_sha1)).decode("ascii").lower() obj = dict( sha1=sha1, dois=dois, - cdx=[dict(url=cdx['url'], dt=cdx['dt'])], + cdx=[dict(url=cdx["url"], dt=cdx["dt"])], size=size, - mimetype=mimetype) + mimetype=mimetype, + ) print(json.dumps(obj)) -if __name__=='__main__': + +if __name__ == "__main__": run() |