def insert_corpora(source_lang, source_text, target_lang, target_text, confidence):
#
# FIXME: Any better idea?
#
PUNCTUATION = ".,:;-_+={}[]()<>|'\"`~!@#$%^&*?"
if source_text == "" or source_text in PUNCTUATION:
return
if target_text == "" or target_text in PUNCTUATION:
return
if source_text == target_text:
return
corpus = Corpus.query.filter_by(
source_lang=source_lang, target_lang=target_lang, source_text=source_text, target_text=target_text
).first()
if corpus == None:
corpus = Corpus.insert(
source_lang=source_lang,
target_lang=target_lang,
source_text=source_text,
target_text=target_text,
confidence=confidence,
frequency=1,
commit=False,
)
else:
corpus.confidence += confidence
corpus.frequency += 1
def corpus_match():
query = request.args.get('q', '')
source_lang = request.args.get('sl', None)
target_lang = request.args.get('tl', None)
matches = Corpus.match(query, source_lang, target_lang)
return json.dumps(map(lambda x: x.serialize(), matches))
