app.py

import streamlit as st
import os
import googleapiclient.discovery
import pandas as pd
import re

pd.set_option("display.max_colwidth", -1)
def extract_all_comments(video_id,youtube,page_token='',comments_list =[]):
    request = youtube.commentThreads().list(part = ['id','snippet'],
                                      maxResults = 100,videoId = video_id ,pageToken= page_token)
    response = request.execute()
    for comment_details in response['items']:
        text_dsiplay = comment_details.get('snippet').get('topLevelComment').get('snippet').get('textDisplay')
        text_original = comment_details.get('snippet').get('topLevelComment').get('snippet').get('textOriginal')
        likes = comment_details.get('snippet').get('topLevelComment').get('snippet').get('likeCount')
        published_at = comment_details.get('snippet').get('topLevelComment').get('snippet').get('publishedAt')
        updated_at = comment_details.get('snippet').get('topLevelComment').get('snippet').get('updatedAt')
        reply_count = comment_details.get('snippet').get('totalReplyCount')
        comments_list.append({'text_dsiplay':text_dsiplay,'text_original':text_original,
               'likes':likes,'published_at':published_at,'updated_at':updated_at,
               'reply_count':reply_count})
    
    if 'nextPageToken' in response.keys():
        if len(comments_list) < 500:
            print("getting next batch of comments")
            extract_all_comments(video_id = video_id,youtube = youtube,page_token= response['nextPageToken'],comments_list= comments_list)
        else:
            print("Limiting results for speed up")
            st.info('Limiting results for speed up')

    return comments_list
  
def extract_comments_from_video(video_id,youtube_api_key):
    try:
    	youtube = googleapiclient.discovery.build(
    		api_service_name, api_version, developerKey = youtube_api_key)
    	
    	found_comments = extract_all_comments(video_id = video_id,youtube = youtube,page_token='')
    	print(len(found_comments))
    	comments_df = pd.DataFrame(found_comments)
    except:
        print("An exception occurred")
        return pd.DataFrame()
    
    return comments_df


#os.environ["OAUTHLIB_INSECURE_TRANSPORT"] = "1"
api_service_name = "youtube"
api_version = "v3"

def find_video_id(youtube_video_id):
	if 'v=' in youtube_video_id:
		video_filter = youtube_video_id.split('v=')[1]
		video_filter = str(video_filter.split('&')[0])
		return video_filter
	return youtube_video_id

st.header('Youtube Comments Extractor')

link='How to create FREE YouTube API for yourself [link](https://blog.hubspot.com/website/how-to-get-youtube-api-key)'
st.markdown(link,unsafe_allow_html=True)

with st.form("my_form"):
	st.write('Enter Youtube API key( Will not be stored )')
	youtube_api_key = st.text_input('API key',placeholder='For less frequent use,Keep this field empty')
	st.write('Enter Youtube Video ID/ Video Link')
	youtube_video_id = st.text_input('Video')
	submitted = st.form_submit_button("Submit")


if submitted:
	youtube_video_id = find_video_id(youtube_video_id)

	if len(youtube_api_key)==0:
		st.warning('Using Default API,Please create your own for frequent use')
		youtube_api_key= os.environ["youtube_api"]
	#st.write("API here",os.environ["youtube_api"],len(os.environ["youtube_api"]))
	df = extract_comments_from_video(youtube_video_id,youtube_api_key)
	
	if len(df) > 0:
		st.dataframe(df)
	else:
		st.info('This video comments are not found')