ibm watson cognitive - Tone analyser only returns analysis for 1 sentence -
when using tone analyser, able retrieve 1 result. example, if use following input text.
string m_stringtoanalyse = "the world rocks ! love !! bananas awesome! old king cole merry old soul!";
the results return analysis document level , sentence_id = 0, ie. "the world rocks !". analysis next 3 sentences not returned.
any idea doing wrong or missing out anything? case when running provided sample code well.
string m_stringtoanalyse = "this service enables people discover , understand, , revise impact of tone in content. uses linguistic analysis detect , interpret emotional, social, , language cues found in text.";
running tone analysis using sample code on sample sentence provided above return results document , first sentence only.
i have tried versions "2016-02-19" "2017-03-15" same results.
i believe if want sentence sentence analysis need send every separate sentence json object. return analysis in array id=sentence_num.
here example of 1 did using multiple youtube comments (using python):
def get_comments(video): #get comments youtube api using requests url = 'https://www.googleapis.com/youtube/v3/commentthreads?part=snippet&maxresults=100&videoid='+ video +'&key=' + youtube_credentials['api_key'] r = requests.get(url) comment_dict = list() # item in comments, add object list text of comment item in r.json()['items']: the_comment = {"text": item['snippet']['toplevelcomment']['snippet']['textoriginal']} comment_dict.append(the_comment) # return list json sentiment_analysis function return json.dumps(comment_dict) def sentiment_analysis(words): # load watson credentials using python sdk tone_analyzer = toneanalyzerv3( username=watson_credentials['username'], password=watson_credentials['password'], version='2016-02-11') # tone, based on json object passed sentiment_analysis return_sentiment = json.dumps(tone_analyzer.tone(text=words), indent=2) return_sentiment = json.loads(return_sentiment)
afterwards can whatever want json object. note else looking @ if want analysis of many objects, can add sentences=false in tone_analyzer.tone function.
Comments
Post a Comment