chilledcow-scrobbler

LastFM scrobbler for chilledcow youtube stream

Requirements

• Tesseract-OCR

Python Requirements

• requests
• opencv
• pafy
• youtube-dl
• numpy
• pillow image
• pytesseract
• pylast

Usage

python ccs.py data_json

Data File

• You need a LastFM API. You can apply from here.
• You also need a Youtube Data Api (no OAuth is necessary) here
• The file must be JSON in this format;

  {
  "LASTFM_API_KEY" : "sample",
  "LASTFM_SHARED_SECRET" : "sample",
  "username" : "sample",
  "password_hash" : "sample",
  "Youtube_Data_API_Key" : "sample",
  "chilledcow_youtube_channel_id" : "UCSJ4gkVC6NrvII8umztf0Ow",
  "song_list_google_doc_id" : "1P0p0SfP_BxiG3I4i8MTYakibsdg_O4753EA-I9h3lL8"
  }

• You can generate your password_hash by

  pylast.md5(password)

How does it work

• It gets the live youtube video id using youtube api.
• It gets the list of the songs from the google docs file
• It works in the background and scrobbles every song written in the stream until you kill the script.
• To do that it takes youtube link as argument and extracts the stream url.
• OpenCV opens the stream and captures one single frame every 30 seconds
• Image gets cropped
• Remove the background
• Tesseract-OCR reads the image
• Possible mistakes are fixed by comparing the song with the list of songs and finding the best one.
• Send to Last.FM (if it didn't send it already)