@@ -201,8 +201,17 @@ def update(self, doc, annotators=None, properties=None):
201201 parseFromDelimitedString (doc , r .content )
202202 return doc
203203
204- def tokensregex (self , text , pattern , filter = False , to_words = False ):
205- matches = self .__regex ('/tokensregex' , text , pattern , filter )
204+ def tokensregex (self , text , pattern , filter = False , to_words = False , annotators = None , properties = None ):
205+ # Error occurs unless put properties in params
206+ if properties is None :
207+ properties = self .default_properties
208+ properties .update ({
209+ 'annotators' : ',' .join (annotators or self .default_annotators ),
210+ 'inputFormat' : 'text' ,
211+ 'outputFormat' : 'serialized' ,
212+ 'serializer' : 'edu.stanford.nlp.pipeline.ProtobufAnnotationSerializer'
213+ })
214+ matches = self .__regex ('/tokensregex' , text , pattern , filter , properties )
206215 if to_words :
207216 matches = regex_matches_to_indexed_words (matches )
208217 return matches
@@ -216,7 +225,7 @@ def semgrex(self, text, pattern, filter=False, to_words=False):
216225 def tregrex (self , text , pattern , filter = False ):
217226 return self .__regex ('/tregex' , text , pattern , filter )
218227
219- def __regex (self , path , text , pattern , filter ):
228+ def __regex (self , path , text , pattern , filter , properties ):
220229 """Send a regex-related request to the CoreNLP server.
221230 :param (str | unicode) path: the path for the regex endpoint
222231 :param text: raw text for the CoreNLPServer to apply the regex
@@ -229,14 +238,25 @@ def __regex(self, path, text, pattern, filter):
229238 # HACK: For some stupid reason, CoreNLPServer will timeout if we
230239 # need to annotate something from scratch. So, we need to call
231240 # this to ensure that the _regex call doesn't timeout.
232- self .annotate (text )
241+ # self.annotate(text)
233242
234243 try :
235- r = requests .get (
244+ # Error occurs unless put properties in params
245+ input_format = properties .get ("inputFormat" , "text" )
246+ if input_format == "text" :
247+ ctype = "text/plain; charset=utf-8"
248+ elif input_format == "serialized" :
249+ ctype = "application/x-protobuf"
250+ else :
251+ raise ValueError ("Unrecognized inputFormat " + input_format )
252+ # change request method from `get` to `post` as required by CoreNLP
253+ r = requests .post (
236254 self .endpoint + path , params = {
237255 'pattern' : pattern ,
238256 'filter' : filter ,
239- }, data = text )
257+ 'properties' : str (properties )
258+ }, data = text ,
259+ headers = {'content-type' : ctype })
240260 r .raise_for_status ()
241261 return json .loads (r .text )
242262 except requests .HTTPError as e :
0 commit comments