podchamp

Crates.iopodchamp
lib.rspodchamp
version0.5.1
sourcesrc
created_at2020-10-28 16:43:46.952094
updated_at2022-01-11 15:25:01.775616
descriptionFetches your podcasts
homepage
repositoryhttps://github.com/rkanati/podchamp
max_upload_size
id306408
size64,233
Rachel Knight (rkanati)

documentation

README

Podchamp

what

A no-frills tool for fetching your podcasts.

Work in progress, but I've been using it exclusively for six months without any major issues.

why

I've found very few podcast tools that aren't either horribly bloated (I don't want my downloader to also be a half-assed media player), decent-but-fragile (shell scripts; tries too hard to make sense of hopelessly broken feeds), or inflexible (insists on performing the download itself; wants to know filename patterns and directories).

I've used greg for years now, but that project is more-or-less dead, and I don't like a few of its design choices, so this is a replacement of sorts.

who

It's for me. This will not get lots of extraneous features like playback or tagging; don't @ me.

how

Podchamp keeps a record of feeds and previously-downloaded episodes in a small database. When checking for new episodes, podchamp downloads the feed xml, parses it, and launches downloads for any episodes it doesn't remember downloading previously.

You add a feed like this:

$ podchamp add mbmbam 'https://feeds.simplecast.com/wjQvYtdl'

and then fetch new episodes like so:

$ podchamp fetch
Fetching mbmbam

If there are new episodes of MBMBaM, podchamp launches PODCHAMP_DOWNLOADER with the download link as an argument. By default this is just wget, which is not super-useful; it's intended that you write your own script that does whatever you feel is appropriate. For example, here's a simplified version of mine:

#!/usr/bin/fish
set dir "$HOME/podcasts/$PODCHAMP_FEED"
mkdir -p "$dir"
cd "$dir"
wget -q "$PODCAST_URI" -O - | \
    nice ffmpeg -y \
    -i pipe:0 -c:a libopus -b:a 64k \
    -metadata title="$PODCHAMP_DATE - $PODCHAMP_TITLE" \
    -metadata artist="$PODCHAMP_FEED" \
    "$PODCHAMP_DATE - $PODCHAMP_TITLE.opus"

Normally, when you add a new feed, it has a backlog of 1. This means it will download only the most recent episode the first time you fetch, and every episode newer than it subsequently. If you want more to be going on with, you can set a larger backlog:

$ podchamp add -n 10 streetfight 'http://feeds.feedburner.com/streetfightradio'

or, if you added it already (or reset the feed) you can change the backlog:

$ podchamp mod streetfight backlog 10

This will download the 10 most recent episodes the first time you fetch, and every episode newer than the oldest of those subequently.

Currently there's no way to download particular episodes, but I'll implement it eventually.

If you decide you don't like a podcast and want podchamp to stop fetching it, you can remove its feed:

# to hell with bean dad
$ podchamp rm roderickontheline

In case you want to re-download previous episodes of a podcast (say, you lost the files, or want fresh copies to transcode them differently), you can reset the feed:

$ podchamp reset guaranteedaudio

when

This will be considered done (i.e. 1.0) when I'm happy with it.

where

On linux, definitely. On windows or other platforms, possibly, but you're on your own; please don't ask for support.

todo

yes

  • Config file - pretty minimal, probably.
  • More metadata - only a few pieces of feed and episode metadata are available to download scripts, and it would be trivial to export more.
  • Tests - duh.
  • Better docs - the above is incomplete; I need to explain environment variables, command-line options, and fetch-since dates.

maybe

  • Self-downloading - this could be added without too much hassle or bloat, but really it's out-of-scope.
  • More robust ordering and episode tracking - currently, podchamp ignores any feed item that has no pubDate or guid; this might be improvable without undue work.
  • Feeds listed in a text file - every time I think about this, it seems inherently brittle and error-prone; an "import" feature might be useful, though.

nope

  • Playback - is what media players are for.
  • Tagging - an absolute tar-pit of complexity; can already be accomplished more reliably by existing tools
  • Support for broken feeds - we may be stuck with the distaster that is RSS, but I have no interest in making podchamp try to work around totally mangled feeds.

whence

Rachel Knight.

whither

All code here is MIT licensed.

Commit count: 41

cargo fmt