lttkgp/youtube_title_parse

Parse song & artist names from YouTube video titles

youtube title parse

Parse the title of a YouTube video to try and get artist & song name.

Description

Video titles on YouTube follow no strict format, and so passing the titles directly to music APIs to fetch metadata hardly works. This module attempts to recognize common patterns (using regex) and extract artist and song name.

Installation

To install youtube_title_parse, simply run:

pip install youtube_title_parse

Usage

CLI

youtube_title_parse comes with a CLI that you can use directly:

$ youtube_title_parse "Seoul - Stay With Us (Official Video)"
Seoul - Stay With Us

Module

You can also import youtube_title_parse as a module. If the module can successfully parse the input, get_artist_title will return a tuple of the format [artist, title] which you can use as below. If not found, [None, None] is returned.

from youtube_title_parse import get_artist_title

artist, title = get_artist_title("Seoul - Stay With Us (Official Video)")
assert artist == "Seoul"
assert title == "Stay With Us"

Credits

This module is originally a Python3 rewrite of the equivalent npm library, get-artist-title, but adds some extra functionality to catch more patterns.

Contributing

Pull requests and stars are always welcome. For bugs and feature requests, please create an issue.

License

youtube-title-parse is made available under the MIT license.

youtube_title_parse's Issues

Make unit tests more readable

Unit tests currently look like:

{
    "input": "Boats & BIrds - Gregory & the Hawk",
    "expected": ["Boats & BIrds", "Gregory & the Hawk"],
}

but it not entirely clear what the expected array is. Considering changing it to:

{
    "input": "Boats & BIrds - Gregory & the Hawk",
    "expected": {
        "artist": "Boats & BIrds",
        "title": "Gregory & the Hawk"
    },
}

Might also be a good time to consider if the return type of the module is best suited as a tuple or dict.

Return tuple instead of string

Is it possible to have get_artist_title return a tuple (artist, title) instead of a string, like how get-artist-title works?

Right now I want to get the parsed artist and title individually, but right now it returns a combined string which I have to split on -.

Enhance the return type to include full string, artist/song name if detected

Add an exhaustive test suite

Add a GitHub action to publish version upgrades (via setup.py) to PyPI

Releases are currently being done manually from local env when there is a new version. Would be nice to have a GitHub action to publish the package to PyPI when there is a commit to master (merged via PR or direct push) that contains a newer version of the package.