Pymusiclooper

Changed

- When successful, the `extend` command now displays the file path of the extended audio file to the user instead of the output directory

Fixed

- Fixed initial loops' output from the `extend` command being affected by the fade out parameter
- Fixed minutes/seconds rounding issues with the `extend` command file naming scheme

Added

- New command: `extend`. Export a longer, extended version of an audio track by looping it seamlessly to the desired length.
- Additional option in `export-points`: `--alt-export-top N`. Can be used to return the N top loop points; provides the same output as in interactive mode but without the table formatting. `--alt-export-top -1` to export all.
- Added CLI_README.md for a basic CLI overview/reference

Changed

- Playback progress bar now updates faster for smoother transitions
- Skip beat analysis if `--brute-force` is enabled, for slightly faster runtime
- Simplified many internal interfaces and added type hints and documentation to most core functions
- Lazy load yt_dlp and sounddevice so that they are only imported when needed
- Updated many core dependencies, including yt-dlp to their latest version
- Improve `--debug` option with better exception debug output
- Lowered the minimum supported Python version to 3.8, since it is already compatible
- Enabled initial Python 3.12 support

Fixed

- Fixed zero crossing algorithm offset potentially being out-of-bounds in very rare cases

Added

- Playback progress bar showing current playback progress and time

Changed

- Candidate pruning now activates if there are 100 or more initial pairs, instead of 250 as in the previous version

Fixed

- Loudness and note difference threshold incorrectly swapped in candidate pair pruning

Added

- New functionality: `tag`. Export metadata tags of loop points to a copy of the input audio file(s) - (Credit: some of the implementation code was based on/inspired by RemedyTwo's fork)
- New functionality: `play-tagged`. Reads the metadata tags of loop points from an audio file and plays it looping
- New audio source option: `--url`. Can now load and process audio from a youtube link (or any stream supported by [yt-dlp](https://github.com/yt-dlp/yt-dlp))
- New loop point search options
- `--min-loop-duration` , `--max-loop-duration` : added min loop duration and max loop duration (in seconds) as optional constraints to the CLI
- `--approx-loop-position` : specify the approximate desired loop start and loop end in seconds, searching around those points only +/- 2 seconds
- `--brute-force` : enables an alternative loop discovery mode that checks the entire audio track instead of the detected beats; useful in case the main algorithm does not yield the desired results.
- `--disable-pruning` : disables the internal filtering of potential loop points

- New export option for split-audio command: `--format`, to change the format of the exported split audio files (currently supported formats: WAV, FLAC, OGG, MP3)
- Official Python 3.10 and 3.11 support

Changed

- Complete re-write of the CLI with much better interface and usability, use `pymuisclooper --help` for the new commands and options or consult the README
- Reimplemented playback using the python sounddevice library instead of mpg123 for better cross-platform compatibility
- Significant runtime improvement to the core loop search algorithm (now runs 10x faster).
- Better loop point alignment thanks to an internal implementation of Audacity's "At Zero Crossings" functionality (less cases of audio popping/clicking due to misaligned loop points)
- Much nicer formatting and interface in interactive mode thanks to the `rich` python package
- Increased the minimum Python requirement to Python 64-bit >=3.9

Removed

- Multiprocessing option (`--n-jobs`). Batch mode operations are otherwise unaffected and work as if `--n-jobs` was fixed to 1.

- Completely removed defunct preserve tags function and its associated errors

- Hotfix for v2.5.1's redundant mpg123-related error message