# Copyright 2016 IBM All Rights Reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ * Thresholds for identifying meaningful tones returned by the Watson Tone Analyzer. Current values are * based on the recommendations made by the Watson Tone Analyzer at * https://console.bluemix.net/docs/services/tone-analyzer/using-tone.html * These thresholds can be adjusted to client/domain requirements. """ PRIMARY_EMOTION_SCORE_THRESHOLD = 0.5 WRITING_HIGH_SCORE_THRESHOLD = 0.75 WRITING_NO_SCORE_THRESHOLD = 0.0 SOCIAL_HIGH_SCORE_THRESHOLD = 0.75 SOCIAL_LOW_SCORE_THRESHOLD = 0.25 # Labels for the tone categories returned by the Watson Tone Analyzer EMOTION_TONE_LABEL = 'emotion_tone' WRITING_TONE_LABEL = 'writing_tone' SOCIAL_TONE_LABEL = 'social_tone' def updateUserTone(conversationPayload, toneAnalyzerPayload, maintainHistory): """ updateUserTone processes the Tone Analyzer payload to pull out the emotion, writing and social tones, and identify the meaningful tones (i.e., those tones that meet the specified thresholds). The conversationPayload json object is updated to include these tones. @param conversationPayload json object returned by the Watson Conversation Service @param toneAnalyzerPayload json object returned by the Watson Tone Analyzer Service @returns conversationPayload where the user object has been updated with tone information from the toneAnalyzerPayload """ emotionTone = None writingTone = None socialTone = None # if there is no context in a if 'context' not in conversationPayload: conversationPayload['context'] = {} if 'user' not in conversationPayload['context']: conversationPayload['context'] = initUser() # For convenience sake, define a variable for the user object user = conversationPayload['context']['user'] # Extract the tones - emotion, writing and social if toneAnalyzerPayload and toneAnalyzerPayload['document_tone']: for toneCategory in toneAnalyzerPayload['document_tone'][ 'tone_categories']: if toneCategory['category_id'] == EMOTION_TONE_LABEL: emotionTone = toneCategory if toneCategory['category_id'] == WRITING_TONE_LABEL: writingTone = toneCategory if toneCategory['category_id'] == SOCIAL_TONE_LABEL: socialTone = toneCategory updateEmotionTone(user, emotionTone, maintainHistory) updateWritingTone(user, writingTone, maintainHistory) updateSocialTone(user, socialTone, maintainHistory) conversationPayload['context']['user'] = user return conversationPayload def initUser(): """ initUser initializes a user object containing tone data (from the Watson Tone Analyzer) @returns user json object with the emotion, writing and social tones. The current tone identifies the tone for a specific conversation turn, and the history provides the conversation for all tones up to the current tone for a conversation instance with a user. """ return { 'user': { 'tone': { 'emotion': { 'current': None }, 'writing': { 'current': None }, 'social': { 'current': None } } } } def updateEmotionTone(user, emotionTone, maintainHistory): """ updateEmotionTone updates the user emotion tone with the primary emotion - the emotion tone that has a score greater than or equal to the EMOTION_SCORE_THRESHOLD; otherwise primary emotion will be 'neutral' @param user a json object representing user information (tone) to be used in conversing with the Conversation Service @param emotionTone a json object containing the emotion tones in the payload returned by the Tone Analyzer """ maxScore = 0.0 primaryEmotion = None primaryEmotionScore = None for tone in emotionTone['tones']: if tone['score'] > maxScore: maxScore = tone['score'] primaryEmotion = tone['tone_name'].lower() primaryEmotionScore = tone['score'] if maxScore <= PRIMARY_EMOTION_SCORE_THRESHOLD: primaryEmotion = 'neutral' primaryEmotionScore = None # update user emotion tone user['tone']['emotion']['current'] = primaryEmotion if maintainHistory: if 'history' not in user['tone']['emotion']: user['tone']['emotion']['history'] = [] user['tone']['emotion']['history'].append({ 'tone_name': primaryEmotion, 'score': primaryEmotionScore }) def updateWritingTone(user, writingTone, maintainHistory): """ updateWritingTone updates the user with the writing tones interpreted based on the specified thresholds @param: user a json object representing user information (tone) to be used in conversing with the Conversation Service @param: writingTone a json object containing the writing tones in the payload returned by the Tone Analyzer """ currentWriting = [] currentWritingObject = [] # Process each writing tone and determine if it is high or low for tone in writingTone['tones']: if tone['score'] >= WRITING_HIGH_SCORE_THRESHOLD: currentWriting.append(tone['tone_name'].lower() + '_high') currentWritingObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'likely high' }) elif tone['score'] <= WRITING_NO_SCORE_THRESHOLD: currentWritingObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'no evidence' }) else: currentWritingObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'likely medium' }) # update user writing tone user['tone']['writing']['current'] = currentWriting if maintainHistory: if 'history' not in user['tone']['writing']: user['tone']['writing']['history'] = [] user['tone']['writing']['history'].append(currentWritingObject) def updateSocialTone(user, socialTone, maintainHistory): """ updateSocialTone updates the user with the social tones interpreted based on the specified thresholds @param user a json object representing user information (tone) to be used in conversing with the Conversation Service @param socialTone a json object containing the social tones in the payload returned by the Tone Analyzer """ currentSocial = [] currentSocialObject = [] # Process each social tone and determine if it is high or low for tone in socialTone['tones']: if tone['score'] >= SOCIAL_HIGH_SCORE_THRESHOLD: currentSocial.append(tone['tone_name'].lower() + '_high') currentSocialObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'likely high' }) elif tone['score'] <= SOCIAL_LOW_SCORE_THRESHOLD: currentSocial.append(tone['tone_name'].lower() + '_low') currentSocialObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'likely low' }) else: currentSocialObject.append({ 'tone_name': tone['tone_name'].lower(), 'score': tone['score'], 'interpretation': 'likely medium' }) # update user social tone user['tone']['social']['current'] = currentSocial if maintainHistory: if not user['tone']['social']['current']: user['tone']['social']['current'] = [] user['tone']['social']['current'].append(currentSocialObject)