Previously, Trash Updater would crawl & parse the Trash Guide's markdown
files to obtain information about release profiles. This is complex and
error prone. Thanks to work done by Nitsua, we now have JSON files
available that describe release profiles in a more concise way. These
files are located at `docs/json/sonarr` in the [Trash Guide repo][1].
All of the markdown parsing code has been removed from Trash Updater.
Now, it shares the same git clone of the Trash Guide repository
originally used for Radarr custom formats to access those release
profile JSON files.
BREAKING CHANGE: The old `type:` property for release profiles is
removed in favor of `trash_id:`, which identifies a specific JSON file
to pull data from. Users are required to update their `trash.yml` and
other configuration files to use the new schema. Until changes are made,
users will see errors when they run `trash sonarr` commands.
[1]: https://github.com/TRaSH-/Guides/tree/master/docs/json/sonarr
I observed that when the user specifies an invalid URL to their Radarr
or Sonarr instance (such as forgetting to use the Base URL), it attempts
to redirect by returning HTTP 300. This is now prevented and a warning
is printed to output.
When `IncludeOptionals` was set to false, release profiles with only
optional terms in them would fail to sync to Sonarr because they were
empty. The filter profile logic now considers this config setting and if
set to false, it will ensure that optionals-only release profiles get
filtered out and not pushed to Sonarr.
The Trash sonarr guide was restructured so that optionals were in a
dedicated release profile that had no non-optional terms in it. Logic
was only checking if a profile had non-optional terms in it, and if not,
it got tossed out. Logic now also checks to make sure there are no
optional terms as well.
GitRepository now lazily-constructs its internal Repository object. This
makes the class testable since it no longer does a lot of work in its
constructor. Previously, the string passed to it had to be valid for the
Repository object to accept it.
Code for cloning the repository refactored to handle failures better.
Namely this means deleting the repository and cloning it again if there
is a failure.
Robert Dailey