@@ -163,13 +163,16 @@ def _request(self, buf, properties):
163163 else :
164164 raise AnnotationException (r .text )
165165
166- def annotate (self , text , annotators = None , properties = None ):
166+ def annotate (self , text , annotators = None , output_format = None , properties = None ):
167167 """Send a request to the CoreNLP server.
168168
169169 :param (str | unicode) text: raw text for the CoreNLPServer to parse
170+ :param (list | string) annotators: list of annotators to use
171+ :param (str) output_format: output type from server: serialized, json, text, conll, conllu, or xml
170172 :param (dict) properties: properties that the server expects
171173 :return: request result
172174 """
175+ # set properties for server call
173176 if properties is None :
174177 properties = self .default_properties
175178 properties .update ({
@@ -178,10 +181,24 @@ def annotate(self, text, annotators=None, properties=None):
178181 'outputFormat' : 'serialized' ,
179182 'serializer' : 'edu.stanford.nlp.pipeline.ProtobufAnnotationSerializer'
180183 })
184+ elif "annotators" not in properties :
185+ properties .update ({'annotators' : ',' .join (annotators or self .default_annotators )})
186+ # if an output_format is specified, use that to override
187+ if output_format is not None :
188+ properties ["outputFormat" ] = output_format
189+ # make the request
181190 r = self ._request (text .encode ('utf-8' ), properties )
182- doc = Document ()
183- parseFromDelimitedString (doc , r .content )
184- return doc
191+ # customize what is returned based outputFormat
192+ if properties ["outputFormat" ] == "serialized" :
193+ doc = Document ()
194+ parseFromDelimitedString (doc , r .content )
195+ return doc
196+ elif properties ["outputFormat" ] == "json" :
197+ return r .json ()
198+ elif properties ["outputFormat" ] in ["text" , "conllu" , "conll" , "xml" ]:
199+ return r .text
200+ else :
201+ return r
185202
186203 def update (self , doc , annotators = None , properties = None ):
187204 if properties is None :
0 commit comments