def create_simple_entity_entries()

in pca-server/src/pca/pca-aws-sf-process-turn-by-turn.py [0:0]

• 24 lines of code
• 12 McCabe index (conditional complexity)

    def create_simple_entity_entries(self, speech_segments):
        """
        Searches through the speech segments given and updates them with any of the simple entity mapping
        entries that we've found.  It also updates the line-level items.  Both methods simulate the same
        response that we'd generate if this was via Standard or Custom Comprehend Entities
        """

        # Need to check each of our speech segments for each of our entity blocks
        for nextTurn in speech_segments:
            # Now check this turn for each entity
            turnText = nextTurn.segmentText.lower()
            for nextEntity in self.simpleEntityMap:
                if nextEntity in turnText:
                    self.matchedSimpleEntities[nextEntity] = self.simpleEntityMap[nextEntity]

        # Loop through each segment looking for matches in our cut-down entity list
        for entity in self.matchedSimpleEntities:

            # Start by recording this in the header
            entityEntry = self.matchedSimpleEntities[entity]
            self.update_header_entity_count(entityEntry["Type"], entityEntry["Original"])

            # Work through each segment
            # TODO Need to check we don't highlight characters in the middle of transcribed word
            # TODO Need to try and handle simple plurals (e.g. type="log" should match "logs")
            for segment in speech_segments:
                # Check if the entity text appear somewhere
                turnText = segment.segmentText.lower()
                searchFrom = 0
                index = turnText.find(entity, searchFrom)
                entityTextLength = len(entity)

                # If found then add the data in the segment, and keep going until we don't find one
                while index != -1:
                    # Got a match - add this one on, then look for another
                    # TODO if entityText is capitalised then use it, otherwise use segment text
                    nextSearchFrom = index + entityTextLength
                    newLineEntity = {}
                    newLineEntity["Score"] = 1.0
                    newLineEntity["Type"] = entityEntry["Type"]
                    newLineEntity["Text"] = entityEntry["Original"]  # TODO fix as per the above
                    newLineEntity["BeginOffset"] = index
                    newLineEntity["EndOffset"] = nextSearchFrom
                    segment.segmentCustomEntities.append(newLineEntity)

                    # Now look to see if it's repeated in this segment
                    index = turnText.find(entity, nextSearchFrom)