Version 0.6.2

Version 0.6.2

• Fixed an issue where some feeds would fail to merge files with an apostrophe in their paths.

• Fixed an issue where chapters taken from the beginning of a file accidentally cause the entire file to be merged into a segment.

0.6.1

Version 0.6.1

• Fixed an issue on Python 3.9 where the app would crash because audio_feeder.data.site.fonts did not have an __init__.py.
• Version information is now stored in the git tag metadata (though hard-coded versions will be written to a file in all build artifacts).
• Adds a new /version endpoint, which displays the current version of audio-feeder that is running.

Version 0.6.0

• Now available: rendered feeds! This adds alternate feeds for files
  generated from the files as they exist on disk using ffmpeg (the
  generated files go into a media cache on disk and are generated when
  the RSS feed is downloaded). The available types of feed are:

  • Single file: In single file mode, all input files are merged
    into one big file, with chapter information if it's available
    (defaulting to considering each separate file a chapter).
  • Chapters: If chapter information is available, each chapter is
    a separate entry in the feed.
  • Segmented: This assumes that you want files broken up into
    duration ~60 minutes, and tries to accommodate that as best
    as possible. The segmenting algorithm recombines the existing
    files along chapter or file boundaries in such a way as to
    minimize the overall deviation from "60 minutes per file".
    It is slightly biased towards longer files, so it will prefer
    to create 1 90 minute file rather than 2 45 minute files, etc.

  A side-effect of this change is that file metadata is stored in the
  database now, which will take some time to add when first loading a
  large number of audiobooks. This also enables us to have chapter
  information in the RSS feeds.

• Added a test server script for easy manual debugging and testing.

• Changed config directory specification. You can now set the
  environment variable AF_CONFIG_DIR to specify exactly where your
  configuration comes from. Whether or not the current working
  directory is in the search path is also now context dependent.

• Removed schema.yml in favor of defining the schema types in
  object_handler.py

• Updated the audio-feeder install script to use
  importlib.resources and made sure that it can be run a second time
  to update the install base.

• "Updating database" status now cleared if the database update
  fails.

• Updated books pagination to consistently use a zero-based index.

Version 0.5.0

This version includes updates modernizing the default template, including:

• Responsive design
• Added a settings dialog exposing the "per page" and "order by" settings.
• Changes the default view to "date added", "descending"
• Audio-feeder now falls back to the last page if you provide and invalid (> max number of pages) page.

Version 0.4.0

This version includes a bunch of modernization of the codebase and the following major changes:

• Ability to use a SQLite database as the backing database instead of my YAML monstrosity.
• The /update target, to trigger an update of the audiobook database without restarting the server.
• Modernization of the main list template, which should now have a more responsive design.
• Dropped ruamel.yaml in favor of pyyaml.
• schema.yml is now loaded directly from the package, not from the database itself.
• The current working directory is now the first place audio-feeder looks for config data.