Example #1
0
 def accepts(self):
     return (
         InputFormat(self, format_id='tei', extension='tei.xml'),
         InputFormat(self, format_id='docx', extension='docx'),
         InputFormat(self, format_id='rst', extension='rst'),
         InputFormat(self, format_id='alpinodocdir', extension='alpinodocdir',directory=True),
     )
Example #2
0
 def accepts(self):
     """Returns a tuple of all the initial inputs and other workflows this component accepts as input (a disjunction, only one will be selected)"""
     return (
         InputFormat(self, format_id='folia', extension='folia.xml'),
         InputFormat(self, format_id='txt', extension='txt'),
         InputComponent(self, ConvertToFoLiA)
     )
Example #3
0
 def accepts(self):
     return [(InputFormat(self,
                          format_id='events',
                          extension='.events.integrated',
                          inputparameter='events'),
              InputFormat(self,
                          format_id='archivedir',
                          extension='.archive',
                          inputparameter='archivedir'))]
Example #4
0
 def accepts(self):
     return InputFormat(self, format_id='events',
                        extension='.enhanced'), InputFormat(
                            self,
                            format_id='events',
                            extension='.integrated'), InputFormat(
                                self,
                                format_id='events',
                                extension='.types')
Example #5
0
 def accepts(self):
     #Note: tuple in a list, the outer list corresponds to options, while the inner tuples are conjunctions
     return [(InputFormat(self,
                          format_id='train',
                          extension='train',
                          inputparameter='trainfile'),
              InputFormat(self,
                          format_id='test',
                          extension='test',
                          inputparameter='testfile'))]
Example #6
0
 def accepts(self):
     """Returns a tuple of all the initial inputs and other workflows this component accepts as input (a disjunction, only one will be selected)"""
     return (InputFormat(self,
                         format_id='txtdir',
                         extension='txtdir',
                         directory=True),
             InputFormat(self,
                         format_id='foliadir',
                         extension='foliadir',
                         directory=True))
Example #7
0
 def accepts(self):
     return (InputFormat(self, format_id='tokenized', extension='tok.json'),
             InputComponent(self,
                            Tokenize,
                            config=self.config,
                            strip_punctuation=self.strip_punctuation,
                            lowercase=self.to_lowercase))
Example #8
0
 def accepts(self):
     return (InputFormat(self,
                         format_id='events',
                         extension='.deduplicated'),
             InputComponent(self,
                            DeduplicateEvents,
                            similarity_threshold=self.similarity_threshold))
Example #9
0
 def accepts(self):
     return (InputFormat(self,
                         format_id='enhanced_events',
                         extension='.enhanced'),
             InputComponent(self,
                            EnhanceEvents,
                            similarity_threshold=self.similarity_threshold))
Example #10
0
 def accepts(self):
     return (InputFormat(self,
                         format_id='filtered',
                         extension='.filtered.json'),
             InputComponent(self,
                            FilterTweets,
                            format_json=self.format_json))
Example #11
0
 def accepts(self):
     return (InputFormat(self,
                         format_id='merged_events',
                         extension='.merged'),
             InputComponent(self,
                            MergeEvents,
                            overlap_threshold=self.overlap_threshold,
                            similarity_threshold=self.similarity_threshold))
Example #12
0
 def accepts(self):
     return [(InputFormat(self,
                          format_id='entity_counts',
                          extension='.counts.npz',
                          inputparameter='entity_counts'),
              InputFormat(self,
                          format_id='dates',
                          extension='.counts_dates',
                          inputparameter='dates'),
              InputFormat(self,
                          format_id='vocabulary',
                          extension='.counts_vocabulary',
                          inputparameter='vocabulary'),
              InputFormat(self,
                          format_id='events',
                          extension='.events',
                          inputparameter='events'))]
Example #13
0
 def accepts(self):
     return (InputFormat(self, format_id='cityref', extension='.json'),
             InputComponent(self,
                            ExtractCityref,
                            config=self.config,
                            strip_punctuation=self.strip_punctuation,
                            to_lowercase=self.to_lowercase,
                            citylist=self.citylist,
                            skip_date=self.skip_date,
                            skip_month=self.skip_month,
                            skip_timeunit=self.skip_timeunit,
                            skip_day=self.skip_day))
Example #14
0
 def accepts(self):
     return (InputFormat(self,
                         format_id='dateref',
                         extension='.dateref.json'),
             InputComponent(self,
                            ExtractDateref,
                            config=self.config,
                            strip_punctuation=self.strip_punctuation,
                            to_lowercase=self.to_lowercase,
                            skip_datematch=self.skip_date,
                            skip_monthmatch=self.skip_month,
                            skip_timeunitmatch=self.skip_timeunit,
                            skip_daymatch=self.skip_day))
Example #15
0
 def accepts(self):
     return InputFormat(self,
                        format_id='txtdir',
                        extension='txtdir',
                        directory=True)
Example #16
0
 def accepts(self):
     return InputFormat(self, format_id='txt', extension='txt')
Example #17
0
 def accepts(self):
     return InputFormat(self,
                        format_id='tiff',
                        extension=self.tiff_extension,
                        directory=True),
Example #18
0
 def accepts(self):
     return [ ( InputFormat(self,format_id='tweetdir',extension='.tweets',inputparameter='tweetdir'), InputFormat(self,format_id='events',extension='.events',inputparameter='events'), InputFormat(self,format_id='entity_burstiness',extension='.burstiness.txt',inputparameter='entity_burstiness') ) ]
Example #19
0
 def accepts(self):
     return (InputFormat(self, format_id='tweetdir', extension='.tweets'), )
Example #20
0
 def accepts(self):
     return InputFormat(self,format_id='events',extension='.events.integrated')
Example #21
0
 def accepts(self):
     return (InputFormat(self, format_id='txt',
                         extension='txt'), InputComponent(self, Lowercaser))
Example #22
0
 def accepts(self):
     return InputFormat(self, format_id='eventdir', extension='.events')
Example #23
0
 def accepts(self):
     return [(InputFormat(self,
                          format_id='events',
                          extension='.events',
                          inputparameter='events'))]
Example #24
0
 def accepts(self):
     return [ ( InputFormat(self,format_id='predictiondir',extension='.instances',inputparameter='predictiondir'), InputFormat(self,format_id='events',extension='.events.integrated',inputparameter='events') ) ]
Example #25
0
 def accepts(self):
     return InputFormat(self, format_id='tweets', extension='.gz')
Example #26
0
 def accepts(self):
     return (
         InputFormat(self, format_id='folia', extension=self.folia_extension),
         InputFormat(self, format_id='foliadir', extension='foliadir'))
Example #27
0
 def accepts(self):
     return [ ( InputFormat(self,format_id='entity_counts_dir',extension='.timeseries',inputparameter='entity_counts_dir') ) ]