diff options
author | Michael Denkowski <michael.j.denkowski@gmail.com> | 2012-12-23 21:58:24 -0500 |
---|---|---|
committer | Michael Denkowski <michael.j.denkowski@gmail.com> | 2012-12-23 21:58:24 -0500 |
commit | cfb1ea5f6b77e87dc28a20116a58f04823501f53 (patch) | |
tree | f9c96dc143ac7dc0337157d1a7a2c2b6bba35413 /python/src | |
parent | ce0e5fd27ba4ca992ddaaee9bf5a8f0667386be1 (diff) |
debugging
Diffstat (limited to 'python/src')
-rwxr-xr-x | python/src/sa/online_extractor.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/python/src/sa/online_extractor.py b/python/src/sa/online_extractor.py index 90087f30..8aae3959 100755 --- a/python/src/sa/online_extractor.py +++ b/python/src/sa/online_extractor.py @@ -111,7 +111,7 @@ class OnlineGrammarExtractor: # f_ i and j are current, e_ i and j are previous def extract(f_i, f_j, e_i, e_j, wc, links, nt, nt_open): # Phrase extraction limits - if wc > self.max_length or (f_j + 1) > f_len or \ + if wc + len(nt) > self.max_length or (f_j + 1) > f_len or \ (f_j - f_i) + 1 > self.max_size: return # Unaligned word @@ -327,6 +327,7 @@ def main(argv): extractor = OnlineGrammarExtractor() for line in sys.stdin: + print >> sys.stderr, line.strip() f_words, e_words, a_str = (x.split() for x in line.split('|||')) alignment = sorted(tuple(int(y) for y in x.split('-')) for x in a_str) extractor.add_instance(f_words, e_words, alignment) |