@@ -218,8 +218,17 @@ def update(self, doc, annotators=None, properties=None):
218218 parseFromDelimitedString (doc , r .content )
219219 return doc
220220
221- def tokensregex (self , text , pattern , filter = False , to_words = False ):
222- matches = self .__regex ('/tokensregex' , text , pattern , filter )
221+ def tokensregex (self , text , pattern , filter = False , to_words = False , annotators = None , properties = None ):
222+ # Error occurs unless put properties in params
223+ if properties is None :
224+ properties = self .default_properties
225+ properties .update ({
226+ 'annotators' : ',' .join (annotators or self .default_annotators ),
227+ 'inputFormat' : 'text' ,
228+ 'outputFormat' : 'serialized' ,
229+ 'serializer' : 'edu.stanford.nlp.pipeline.ProtobufAnnotationSerializer'
230+ })
231+ matches = self .__regex ('/tokensregex' , text , pattern , filter , properties )
223232 if to_words :
224233 matches = regex_matches_to_indexed_words (matches )
225234 return matches
@@ -233,7 +242,7 @@ def semgrex(self, text, pattern, filter=False, to_words=False):
233242 def tregrex (self , text , pattern , filter = False ):
234243 return self .__regex ('/tregex' , text , pattern , filter )
235244
236- def __regex (self , path , text , pattern , filter ):
245+ def __regex (self , path , text , pattern , filter , properties ):
237246 """Send a regex-related request to the CoreNLP server.
238247 :param (str | unicode) path: the path for the regex endpoint
239248 :param text: raw text for the CoreNLPServer to apply the regex
@@ -246,14 +255,25 @@ def __regex(self, path, text, pattern, filter):
246255 # HACK: For some stupid reason, CoreNLPServer will timeout if we
247256 # need to annotate something from scratch. So, we need to call
248257 # this to ensure that the _regex call doesn't timeout.
249- self .annotate (text )
258+ # self.annotate(text)
250259
251260 try :
252- r = requests .get (
261+ # Error occurs unless put properties in params
262+ input_format = properties .get ("inputFormat" , "text" )
263+ if input_format == "text" :
264+ ctype = "text/plain; charset=utf-8"
265+ elif input_format == "serialized" :
266+ ctype = "application/x-protobuf"
267+ else :
268+ raise ValueError ("Unrecognized inputFormat " + input_format )
269+ # change request method from `get` to `post` as required by CoreNLP
270+ r = requests .post (
253271 self .endpoint + path , params = {
254272 'pattern' : pattern ,
255273 'filter' : filter ,
256- }, data = text )
274+ 'properties' : str (properties )
275+ }, data = text ,
276+ headers = {'content-type' : ctype })
257277 r .raise_for_status ()
258278 return json .loads (r .text )
259279 except requests .HTTPError as e :
0 commit comments