Mercurial > hg > cc > azure
view master/src/wecu/sac_mapper.py @ 66:b04870ab3035
don't over-count duplicate URIs in multiple properties, produce composite keys instead
author | Henry S. Thompson <ht@markup.co.uk> |
---|---|
date | Thu, 04 Jun 2020 16:10:55 +0000 |
parents | cfaf5223b071 |
children |
line wrap: on
line source
#!/usr/bin/python3 import os import sys import re is_regex = sys.argv[1] == 'true' search_terms = sys.argv[2:] search_terms_counters = dict() # Initialise for term in search_terms: search_terms_counters[term] = 0 if not is_regex: for line in sys.stdin: for term in search_terms_counters: search_terms_counters[term] += line.count(term) else: for line in sys.stdin: for term_regex in search_terms_counters: search_terms_counters[term_regex] += len(re.findall(term_regex, line)) for term in search_terms_counters: print("{}\t{}".format(term, search_terms_counters[term]))