summaryrefslogtreecommitdiff
path: root/python/cdec/sa/extractor.py
diff options
context:
space:
mode:
authorMichael Denkowski <mdenkows@cs.cmu.edu>2013-09-23 08:21:58 -0700
committerMichael Denkowski <mdenkows@cs.cmu.edu>2013-09-23 08:21:58 -0700
commit6efdb6758dc2d204d203bf3dcdb0ea14b6e2bbd5 (patch)
tree41a5223781397a7c9bcab70054be0495c160101f /python/cdec/sa/extractor.py
parentee384acf34de0d7613a70b81c674d607a3bd782a (diff)
One extractor, multiple online contexts.
Diffstat (limited to 'python/cdec/sa/extractor.py')
-rw-r--r--python/cdec/sa/extractor.py14
1 files changed, 4 insertions, 10 deletions
diff --git a/python/cdec/sa/extractor.py b/python/cdec/sa/extractor.py
index acc13cbc..23534712 100644
--- a/python/cdec/sa/extractor.py
+++ b/python/cdec/sa/extractor.py
@@ -84,23 +84,17 @@ class GrammarExtractor:
for fn in cdec.sa._SA_CONFIGURE:
fn(config)
- def grammar(self, sentence):
+ def grammar(self, sentence, ctx_name=None):
if isinstance(sentence, unicode):
sentence = sentence.encode('utf8')
words = tuple(chain(('<s>',), sentence.split(), ('</s>',)))
meta = cdec.sa.annotate(words)
cnet = cdec.sa.make_lattice(words)
- return self.factory.input(cnet, meta)
+ return self.factory.input(cnet, meta, ctx_name)
# Add training instance to data
- def add_instance(self, sentence, reference, alignment):
+ def add_instance(self, sentence, reference, alignment, ctx_name=None):
f_words = cdec.sa.encode_words(sentence.split())
e_words = cdec.sa.encode_words(reference.split())
al = sorted(tuple(int(i) for i in pair.split('-')) for pair in alignment.split())
- self.factory.add_instance(f_words, e_words, al)
-
- # Debugging
- def dump_online_stats(self):
- self.factory.dump_online_stats()
- def dump_online_rules(self):
- self.factory.dump_online_rules() \ No newline at end of file
+ self.factory.add_instance(f_words, e_words, al, ctx_name)