GitHub/gf-core
1
0
Fork 1
mirror of https://github.com/GrammaticalFramework/gf-core.git synced 2026-04-25 12:32:50 -06:00
Code Issues Packages Projects Releases Wiki Activity

change the API for literals in Java and Python. The input sentence is no longer a parameter to the callbacks.

Browse Source
This commit is contained in:
krasimir
2016-05-20 15:20:02 +00:00
parent f6b67e1550
commit 694cdffa72
8 changed files with 26 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/runtime/python/examples/translation_pipeline.py
View File

@@ -129,8 +129,8 @@ def clean_gfstrings(sentence):
sentence = sentence.replace(entry, ' '.join(entry[1:-1].split('_')[:-1]) if entry.find('_') != -1 else '');
return ' '.join( sentence.split() );
def parseNames(grammar, language):
def callback(lin_idx, sentence, start):
def parseNames(grammar, language, sentence):
def callback(lin_idx, start):
moving_start, end, eot = start, len(sentence), True;
if moving_start < end and (not sentence[moving_start].isupper()):
return None;
@@ -175,8 +175,8 @@ def parseNames(grammar, language):
return None;
return callback;
def parseUnknown(grammar, language):
def callback(lin_idx, sentence, start):
def parseUnknown(grammar, language, sentence):
def callback(lin_idx, start):
moving_start, end, eot = start, len(sentence), True;
isNewToken = (moving_start == 0) or (moving_start > 1 and sentence[moving_start-1].isspace()) # -- added to deal with segmentation errors like may => ma_N + Symb y
if moving_start < end and (not sentence[moving_start].isupper()):
@@ -271,7 +271,7 @@ def pipelineParsing(grammar, language, sentences, K=20):
#buf = [sent for sent in sentences];
buf, sentences = itertools.tee(sentences, 2);
sentences = itertools.imap(gf_utils.lexer(lang=language), sentences);
parser = gf_utils.getKBestParses(grammar, language, K, callbacks=[("PN", parseNames(grammar, language)), ("Symb", parseUnknown(grammar, language))]);
parser = gf_utils.getKBestParses(grammar, language, K);
for sent, (time, parsesBlock) in itertools.izip(buf, itertools.imap(parser, sentences)):
yield (sent, parsesBlock);