def addUserInputToConceptModel(self, user_input, cutoff=0.2): """ This method attempts to resolve user input (`input`) into a list of related concepts. At issue is the fact that user input has to be resolved somehow to the name of the nearest Wikipedia article, not always easy. In cases where this is not possible, this method will return a False flag. In cases where this happens as desired, this method will return a True flag. """ # Fetch the precise name of the node (article title) associated with the institution. concept_node = event_insight_lib.annotateText(user_input, backend.getToken()) # If the correction call is successful, keep going. if 'annotations' in concept_node.keys() and len(concept_node['annotations']) != 0: concept_node_title = concept_node['annotations'][0]['concept']['label'] related_concepts = event_insight_lib.fetchRelatedConcepts(concept_node_title, backend.getToken()) model = backend.parseRawConceptCall(related_concepts, cutoff) new_concept_model = ConceptModel(model=model, maturity=1) self.iterateModel(new_concept_model) return True # Otherwise, if the call was not successful, return a False flag. else: return None
def addEventToConceptModel(self, event_text, cutoff=0.2): """ In the case that we are resolving a known concept, we can skip the verification step present in the above method. We also use a different Watson API call entirely to retrieve what we want. """ merger_model = ConceptModel(model=backend.parseRawEventCall(event_insight_lib.annotateText(event_text, backend.getToken()), cutoff)) self.iterateModel(merger_model)
