SlipySlidy Playlist Scraper

slipy_slidy_playlist_scraper.py

import re
from pytube import Playlist
from pytube import YouTube
import jsonpickle
from dataclasses import dataclass

@dataclass
class Metadata:
    albumArtist: str
    artist: str
    title: str
    year: str
    
@dataclass
class Source:
    client: str
    date: str
    structure: str
    version: str

@dataclass
class Song:
    description: str
    filename: str
    mark: str
    metadata: Metadata
    source: Source
    youtube: str


url = "https://www.youtube.com/playlist?list=PLARr36qkoiWbhGZEZDvgv6F5vfRMR79_q"
playlist = Playlist(url)
playlist._video_regex = re.compile(r"\"url\":\"(/watch\?v=[\w-]*)")

f = open("Bgm_MapleStory2.json", "a")

for url in playlist.video_urls:
    youtube = url.split("https://www.youtube.com/watch?v=")[1]
    yt = YouTube(url)
    description = yt.title.split("[MapleStory 2 BGM] ")[1]
    filename = "".join(description.split())
    mark = filename
    date = yt.publish_date.date()
    metadata = Metadata("Wizet", "StudioEIM", description, f'{date.year}')
    source = Source("MS2", date, "Bgm_MapleStory2", "1.0") 
    song = Song(description, filename, mark, metadata, source, youtube)
    songJson = jsonpickle.encode(song, unpicklable=False)
    print(songJson)
    f.write(f'{songJson},')
f.close()

scrape https://www.youtube.com/user/SlipySlidy playlist and write to json format in https://github.com/maplestory-music/maplebgm-db

need to do some manual work to make it a json array and format nicely 😅