from googleapiclient import discovery
import json
import csv
from pathlib import Path

from time import sleep

from itertools import islice

API_KEY = open('perspective_api_key').read()

client = discovery.build("commentanalyzer","v1alpha",developerKey=API_KEY,discoveryServiceUrl="https://commentanalyzer.googleapis.com/$discovery/rest?version=v1alpha1",static_discovery=False,)

csvreader = csv.DictReader(open("all_data.csv"), dialect='unix')

outfile = Path("perspective_results.json")
already_scored = set()
if outfile.exists():
    already_scored = set([json.loads(l)['id'] for l in open(str(outfile),'r')])

print(f"loaded {len(already_scored)} scored comments")
with open("perspective_results.json",'a') as of:
    for line in csvreader:
        if line['id'] not in already_scored and len(line.get('comment_text','')) > 0:
            analyze_request = {'comment':{'text':line['comment_text']},
                               'languages':['en'],
                               'requestedAttributes':{'TOXICITY':{},
                                                      "SEVERE_TOXICITY":{},
                                                      "IDENTITY_ATTACK":{},
                                                      "INSULT":{},
                                                      "PROFANITY":{},
                                                      "THREAT":{}}}
            response = client.comments().analyze(body=analyze_request).execute()
            response['id'] = line['id']
            result = json.dumps(response)
            of.write(result + '\n')
            of.flush()

            sleep(0.10)