Skip to content

Latest commit

 

History

History
66 lines (51 loc) · 4.02 KB

todo.org

File metadata and controls

66 lines (51 loc) · 4.02 KB

DONE: Check integrity of mp3 using

ffmpeg -v error -i not_DSL_macros_-_Christophe_Grand.mp3 -f null - from https://superuser.com/questions/100288/how-can-i-check-the-integrity-of-a-video-file-avi-mpeg-mp4 just do it until I find another faulty one

DONE: Don’t download webm encoded videos, my podcast app and the internal android music player cannot skip to a specific timecode with these files. Use mp4,m4a etc instead

DONE: try-except around downloading of youtube video

TODO: implement logging

DONE: use django’s builtin rss generator

DONE: try out domain-less url in enclosure. my podcatcher seems to be able to find the thumbnail using only the media… part

Does not work, probably because the enclosure url can be an arbitrary url that does not need to belong to the same url

DONE: Delete files when podcast/episode is deleted

Using pre_delete signal to delete everything

DONE: cache-control headers in rss response so that firefox doesn’t cache them. Content-type can be controller by setting

the content_type field in my feed generator but for the cache-control head I probably need to write a new view that instantiates that generator, generates a feed and then sends it using HttpResponse where I can set the headers. harmontown uses: Cache-control: no-cache, no-store, must-revalidate, private, max-age=0 https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Cache-Control?utm_source=mozilla&utm_medium=devtools-netmonitor&utm_campaign=default

DONE: add ignored-episodes list so that episodes that are in a playlist but are no longer available/I deleted them in the admin panel do not get put into the episode list again during an update

Not easily possible to have list fields with sqlite, instead just add an invalid flag to episode. But not a great solution either because now I have to filter them out every time. I should just delete them since invalid videos in a playlist are filtered out by pafy automatically

DONE: Make it so I can add extra urls to a playlist podcast. Then the YT_PLAYLIST choice becomes moot and together with the ignored-episodes I should just use a urls list as the central object

Podcast is now unified. It wasn’t a good abstraction to begin with

DONE: deploy on my raspberrypi using lighttpd or apache

apache config in /etc/apache2/sites-available/000-default.conf In the end I did it like https://www.digitalocean.com/community/tutorials/how-to-serve-django-applications-with-apache-and-mod_wsgi-on-ubuntu-14-04#configure-apache but apparently the difficulties I had earlier when using the django documentation were a result relative paths when reading the api key and had nothing per se to do with apache

DONE: generate new secret key and put it in separate file like api key

DONE: add form to create podcasts with

DONE: folder with mp3s

DONE: playlist url

DONE: extract description, images and more from the mp3s

DONE: make invalid a read/write field (OR save youtube-id upon deletion so that they do not get readded in the next update)

But I don’t hve a listField field for a model. The only way I know would be with another invalidEpisode class that has podcast as a foreign key and that just would not be nice Decided to just make invalid field writable for now

DONE: write tests

  • add podcast
  • update podcast
  • download podcast (preferably small files)
  • delete podcast
  • check generated rss

update tests, they probably don’t work anymore

fix headers for media file responses. atm Firefox and my podcast app cannot stream them

because it does not send the mime type I think. At least that’s what firefox was complaining about

add sync scheduled task by default (as a migration)

move mp3 upload to the admin site

add rss link to admin site

add more rss feed tags

fix urls in rss feed

put individual episode updates also in queue then wait for all to finish before download, alternatively chain all the updates and downloads