* Add 'Language-equals' support
This feature will treat couples of languages as equal for list-subtitles
operations. It's optional; its methods won't do anything if an empy list
is set. See more info at docstrings from 'subliminal_patch.core'.
For example, let's say I only want to have "Spanish (es.srt)" subtitles
and I don't care about the differences between Spain and LATAM spanish.
This feature will allow me to always get European Spanish even from
LATAM Spanish providers like Argenteam and Subdivx.
Example for config.ini:
language_equals = ['spa-MX:spa']
(Which means all Latam Spanish subtitles from every provider will be
converted to European Spanish)
* Add PT and ZH language tests
* Add HI and Forced parsing for language pairs
Format example: ["en@HI:en", "es-MX@forced:es-MX"]
* Update languages.py
* Update API definition to reflect the previous change
* Add language equals table to the UI (test only)
* Add global language selector and get language from code3 utilities
* Add unit tests for language equal feature
* Add encode function to language equal feature
* Add CRUD methods to the language equals panel
* Add equals description
* Add parsing support for alpha3 custom languages
* no log: add more tests
* Add forced and hi support to the language equal target
---------
Co-authored-by: morpheus65535 <louis_vezina@hotmail.com>
Co-authored-by: LASER-Yi <liangyi0007@gmail.com>
Currently only configurable via manual `data/config/config.ini`
text edition.
New configurable values are `series_scores` and `movie_scores`.
For each config section, the sum of the config values (except hash)
must be equal to the hash value plus one (1), otherwise default
values will be used (notified via debug log).
Hash values are not meant to be modified; the value is shown in
`config.ini` for reference. Modifying hash values would imply
breaking Bazarr's score logic.
I was unable to reproduce the exception, but I found a strange pattern:
whenever the hints["expected_title"] value had the same value as the
title found in the filename, there was a list from the rebulk module
constantly growing. Once I remove the expected_title "hint", this
behavior is gone.
This change is transparent. If for some reason guessit returns a
different movie/series title, Bazarr will refine it with the database
anyway.
Use the whole subtitle file to detect encoding. Seems like if trying to detect encoding only by a part of the file, is not always working. This modification will lead to worse performance, but should be more reliable.
Vitiko