return json in camel case

handler.py

@@ -8,7 +8,7 @@ nltk.download('stopwords')
 
 from utils import clean_str, clean_str_nopunct
 import torch
-from utils import MultiHeadModel, BertInputBuilder, get_num_words
+from utils import MultiHeadModel, BertInputBuilder, get_num_words, MATH_PREFIXES, MATH_WORDS
 
 import transformers
 from transformers import BertTokenizer, BertForSequenceClassification
@@ -360,16 +360,16 @@ class EndpointHandler():
         question_model.run_inference(transcript)
         transcript.update_utterance_roles(uptake_speaker)
         transcript.calculate_aggregate_word_count()
-        return_dict = {'talk_distribution': None, 'talk_length': None, 'talk_moments': None, 'common_top_words': None, 'uptake_top_words': None}
+        return_dict = {'talkDistribution': None, 'talkLength': None, 'talkMoments': None, 'commonTopWords': None, 'uptakeTopWords': None}
         talk_dist, talk_len = transcript.get_talk_distribution_and_length(uptake_speaker)
-        return_dict['talk_distribution'] = talk_dist
-        return_dict['talk_length'] = talk_len
+        return_dict['talkDistribution'] = talk_dist
+        return_dict['talkLength'] = talk_len
         talk_timeline = transcript.get_talk_timeline()
         talk_moments = talk_timeline
-        return_dict['talk_moments'] = talk_moments
+        return_dict['talkMoments'] = talk_moments
         word_cloud, uptake_word_cloud = transcript.get_word_cloud_dicts()
-        return_dict['common_top_words'] = word_cloud
-        return_dict['uptake_top_words'] = uptake_word_cloud
+        return_dict['commonTopWords'] = word_cloud
+        return_dict['uptakeTopwords'] = uptake_word_cloud
 
 
         return return_dict