podcasts: import to beets, read info from json, import cover from jpg

[prankousky]

Hi everybody,

previously, I exclusively used beets for music and audiobooks. This always worked great. Audiobooks usually needed to be tagged manually before important, but that was no big issue.

Now I am trying to build a perfectly tagged podcast library with beets, but don't know if this can be automated.

Scenario

I have

• <date-title>.mp3, for example 20190123-Können Namen peinlich sein_.mp3
• <date-title>.meta.json, for example 20190123-Können Namen peinlich sein_.meta.json
• <date-title>.jpg, for example 20190123-Können Namen peinlich sein_.jpg

In this case, the (...)sein_.jpg underscore is only there because the question mark in the title got renamed to an underscore; there won't be a _ in front of the file extension if the title just included letters and numbers.

mp3 is obviously the audio file itself.

json includes the following (contents of the actual file)

{
    "title": "Können Namen peinlich sein?",
    "contentSnippet": "Sabine Leutheusser-Schnarrenberger und Cheyenne Savannah Ochsenknecht, so heißen Menschen wirklich. Allerdings heißt Dietmar Wischmeyer eben Dietmar Wischmeyer. Er und seine Podcast-Partnerin Tina Voss wundern sich darüber – in Wischmeyers Stundenhotel von Bremen Zwei. Ihr Name geht schon eleganter über die Lippen.",
    "pubDate": "Wed, 23 Jan 2019 21:05:00 +0100"
}

and jpg is the cover file.

What I'd like to do

When beets imports the file(s), I'd need it to automatically do the following for each mp3 file using the appropriate json and jpg files (which it could identify by their name)

1. set the title in metadata to the title part from the json file
2. set the comment in metadata to the contentSnippet part from the json file
3. set the coverart to the jpg file that is named
4. either
   4.1 if possible set year in metadata to the year value from pubDate from the json file
   4.2 set copyright to pubDate from the json file
   4.3 or just do both

Is this possible? If so, I'd only have to manually change the artist and perhaps album metadata, instead of doing everything by hand.

I use podcast-dl to download my podcasts. Unfortunately, even though I thought it would, it doesn't create any metadata (tried the different command line options), nor does it write the artwork the audio file. It will download the metadata as json and artwork as jpg, but the mp3 file looks like this

➜ id3info 20190123-Können\ Namen\ peinlich\ sein_.mp3

*** Tag information for 20190123-Können Namen peinlich sein_.mp3
*** mp3 info
MPEG1/layer III
Bitrate: 128KBps
Frequency: 48KHz

There is nothing.

Is there a way to use beets to do this, then import it to my collection?

Thank you in advance for your help :)

[sampsyo]

We don't yet have any good way to read import matches from external files, like JSON. That sounds like an interesting plugin to write, however, and doesn't seem like it would be too hard! It could behave like any other "data source" plugin (like the ones for Beatport, Discogs, etc.) but actually just search within a JSON file. Feel free to give it a shot if you're interested!

[prankousky]

Thank you @sampsyo

Unfortunately, my skills aren't adequate; even if I were able to hack something together, I doubt it would be good enough to work for anybody else - that being said, I wouldn't even know were to begin at all, unfortunately.