import requests
import pandas as pd

# Google API key
key = 'your-key'

# ID of the video you want to analyze (get it from YouTube url)
video_id = 'wSY-KRMpzhM'

# initialize empty list of users
user_list = []

# get the first set of comments
first_response = requests.get('https://www.googleapis.com/youtube/v3/commentThreads?key=' + key + 
'&textFormat=plainText&part=snippet&videoId=' + video_id + '&maxResults=100')

data1 = first_response.json()

d1 = data1.get('items')

# get the commenter's channel url for the first set of comments
for snippet in d1:
	url = snippet.get('snippet').get('topLevelComment').get('snippet').get('authorChannelUrl')
	user_list.append(url)

# get next page token if there are more than 100 comments
nextPageToken = data1.get('nextPageToken')

# get the rest of the comments
while nextPageToken:

	next_response = requests.get('https://www.googleapis.com/youtube/v3/commentThreads?key=' + key + 
	'&textFormat=plainText&part=snippet&videoId=' + video_id + '&maxResults=100&pageToken=' + nextPageToken)

	data2 = next_response.json()
	
	#print data2
	
	d2 = data2.get('items')

	nextPageToken = data2.get('nextPageToken')
	
	for snippet in d2:
		url = snippet.get('snippet').get('topLevelComment').get('snippet').get('authorChannelUrl')
		if url: user_list.append(url)
		
import pprint
pprint.pprint(user_list)
print len(user_list),' users commented.'

# initialize empty list for the data
user_list_subs = []

#loop through all channels who commented on the video and get their statistics
for channel in user_list:
	
	#get the channel id
	id = channel[-24:]
	print id
	
	#get channel statistics
	user_response_stats = requests.get('https://www.googleapis.com/youtube/v3/channels?part=statistics&key=' 
	+ key + '&id=' + id)
	
	user_data = user_response_stats.json()
	
	# get subscriber count
	sub_count = user_data.get('items')[0].get('statistics').get('subscriberCount')
	sub_count = int(sub_count)
	print sub_count
	
	# get the channel's view count
	view_count = user_data.get('items')[0].get('statistics').get('viewCount')
	view_count = int(view_count)
	print view_count
	
	# get the channel's video count (how many videos it uploaded)
	video_count = user_data.get('items')[0].get('statistics').get('videoCount')
	video_count = int(video_count)
	print video_count
	
	# get channel metadata, e.g. title and custom url
	user_response_snippet = requests.get('https://www.googleapis.com/youtube/v3/channels?part=snippet&key=' 
	+ key + '&id=' + id)
	
	user_data = user_response_snippet.json()
	
	title = user_data.get('items')[0].get('snippet').get('title')
	custom_url = user_data.get('items')[0].get('snippet').get('customUrl')
	print title

	
	
	#add all data to the list
	user_list_subs.append([channel,title,sub_count,view_count,video_count,custom_url])
	
	

pprint.pprint(user_list_subs)

#Convert user_list_subs to dataframe
import pandas as pd
labels = ['channel','title','subscriber_count','view_count','video_count','custom_url']
df = pd.DataFrame.from_records(user_list_subs, columns = labels)

# sort the data by number of subscribers
df = df.sort_values(by=['subscriber_count'],ascending=False)

print df

# save the sorted data frame to a csv
df.to_csv('parachute-young-march28.csv',index=False, encoding='utf-8')