Make YouTube playlist from posts in Facebook group

add_urls_to_playlist.py

import requests
import json
import re
from more_itertools import unique_everseen

with open("urls.json", "r") as fp:
    urls = json.load(fp)

# lots of regex to extract video ID from different types of YouTube links
def get_id(url):
    if "youtube" in url:
        if "attribution" in url:
            try:
                video_id = re.search("%3D(.*)\%26", url).group(1)
            except Exception:
                print(url)
                video_id = ""
        else:
            try:
                video_id = re.search("(\?|\&)(v|ci)=(.*)", url).group(3)
            except Exception:
                print(url)
                video_id = ""
    else:
        try:
            video_id = re.search("\.be\/(.*)", url).group(1)
        except Exception:
                print(url)
                video_id = ""
    return video_id

# get video IDs from URLs, then deduplicate
ids = list(map(get_id, urls))
ids = list(filter(lambda x: len(x) > 0, ids))
ids = list(unique_everseen(ids))

# access token can be acquired from https://developers.google.com/oauthplayground/
access_token = "YOUR_ACCESS_TOKEN"

# your YouTube playlist ID can be found in the playlist URL
playlist_id = "YOUR_PLAYLIST_ID"

# correct API path
path = "https://www.googleapis.com/youtube/v3/playlistItems?part=snippet&access_token={access_token}".format(access_token=access_token)
headers={'Content-Type': 'application/json'}

# finally insert the video into your playlist
for video_id in ids:
    payload = {
        "snippet": {
            "playlistId": playlist_id,
            "resourceId": {
                "videoId": video_id,
                "kind": "youtube#video"
            }
        }
    }

    request = requests.post(path, data=json.dumps(payload), headers=headers)
    # for confirmation of success
    print(request.content)

filter_posts_for_likes.py

import pandas as pd
import json

# set required likes here
required_likes = 5

all_post_data = []


filename = "all_group_post_data.json"
with open(filename, "r", encoding="utf-8") as fp:
    all_post_data += json.load(fp)

one_dimension_data = []

# reshape json into 1D for use with Pandas
for post in all_post_data:
    post_data = {}
    try:
        post_data["poster"] = post["from"]["name"]
    except KeyError:
        post_data["poster"] = ""
    try:
        post_data["message"] = post["message"]
    except KeyError:
        post_data["message"] = ""
    try:
        post_data["name"] = post["name"]
    except KeyError:
        post_data["name"] = ""
    try:
        post_data["link"] = post["link"]
    except KeyError:
        post_data["link"] = ""
    try:
        post_data["likes"] = len(post["likes"]["data"])
    except KeyError:
        post_data["likes"] = 0
    try:
        post_data["url"] = post["permalink_url"]
    except KeyError:
        post_data["url"] = ""
    try:
        post_data["time"] = post["created_time"]
    except KeyError:
        post_data["time"] = ""

    one_dimension_data.append(post_data)

# create dataframe for easy filtering and manipulation
df = pd.DataFrame(one_dimension_data)
# all link fields must contain 'youtu'
youtube_df = df[df["link"].str.contains("youtu")]
# all posts must have required_likes or more
likes_youtube_df = youtube_df[youtube_df["likes"] >= required_likes]

urls_series = likes_youtube_df["link"]
urls = list(urls_series)

#serialise YouTube URLs
with open("urls.json", "w") as fp:
    json.dump(urls, fp)

get_group_post_data.py

import requests
import json

# Facebook API path, group_id appears in group URL
# Personal access token can be acquired here: https://developers.facebook.com/tools/accesstoken/
path = "https://graph.facebook.com/v2.9/"
group_id = "YOUR_GROUP_ID"
group_posts_path = "{group_id}/feed/?fields=id".format(group_id=group_id)
access_token = "YOUR_ACCESS_TOKEN"

group_url = "{path}{group_posts_path}&access_token={access_token}".format(
    path=path,
    group_posts_path=group_posts_path,
    access_token=access_token)

url_array = []

# go through all group pagination to get all URLs to scrape
def recurse_pages(url):
    global url_array
    response = requests.get(url)
    json_content = response.json()
    print("Getting next url after {url}".format(url=url))
    if len(json_content["data"]) > 0:
        next_url = json_content["paging"]["next"]
        url_array.append(next_url)
        recurse_pages(next_url)


recurse_pages(group_url)

# serialise URLs to scrape
with open("all_group_urls.json", "w") as fp:
    json.dump(url_array, fp)

all_post_ids = []

# scrape all urls for post IDs
def get_ids(url):
    print("\nGetting IDs for {}".format(url))
    post_ids = []
    response = requests.get(url)
    json_content = response.json()
    data = json_content["data"]
    if len(data) > 0:
        for post in data:
            post_id = post["id"]
            post_ids.append(post_id)
    return post_ids


for url in url_array:
    post_ids = get_ids(url)
    all_post_ids += post_ids

# serialise post IDs
with open("all_group_post_ids.json", "w") as fp:
    json.dump(all_post_ids, fp)

all_post_data = []

# define fields to get from posts
fields = "link,story,message,created_time,from,id,likes,name,permalink_url"

# get all post data from all above IDs
def get_post_data(post_id):
    url = "https://graph.facebook.com/{post_id}?fields={fields}&access_token={access_token}".format(
        post_id=post_id,
        fields=fields,
        access_token=access_token)
    print("\nGetting data from {url}".format(url=url))
    response = requests.get(url)
    data = response.json()
    return data


for post_id in all_post_ids:
    all_post_data.append(get_post_data(post_id))

# serialise post data
with open("all_group_post_data.json", "w") as fp:
    json.dump(all_post_data, fp)