Mercurial > repos > stevecassidy > nltktools
diff g_collocation.py @ 1:fb617586f4b2 draft
planemo upload commit a81826fe44f09a3710a35c183aa88b745aeec064-dirty
author | stevecassidy |
---|---|
date | Mon, 05 Dec 2016 05:22:05 -0500 |
parents | e991d4e60c17 |
children | a47980ef2b96 |
line wrap: on
line diff
--- a/g_collocation.py Wed Oct 12 22:17:53 2016 -0400 +++ b/g_collocation.py Mon Dec 05 05:22:05 2016 -0500 @@ -18,8 +18,9 @@ def collocation(inp, outp, freq_filter, results, coll_type, pos): pos = bool(pos == 'true') - i = str(unicode(open(inp, 'r').read(), errors='ignore')) - o = open(outp, 'w') + with open(inp, 'r') as fd: + i = fd.read() + all_words = [] if pos: text = i.split(' ')[:-1] @@ -37,10 +38,10 @@ finder = TrigramCollocationFinder.from_words(all_words) finder.apply_freq_filter(int(freq_filter)) colls = finder.nbest(measures.pmi, int(results)) - for coll in colls: - o.write("%s\t%s" % coll) - o.write('\n') - o.close() + with open(outp, 'w') as output: + for coll in colls: + output.write("%s\t%s" % coll) + output.write('\n') if __name__ == '__main__': args = Parser()