cleanit

CleanIt

Subtitles extremely clean.

• Project page
  https://github.com/ratoaq2/cleanit

CleanIt is a command line tool that helps you to keep your subtitles clean. You can specify your own rules to detect entries to be removed or patterns to be replaced. Simple text matching or complex regex can be used. It comes with standard rules out of the box:

• ocr: Fix common OCR errors
• tidy: Fix common formatting issues (e.g.: extra/missing spaces after punctuation)
• no-sdh: Remove SDH descriptions
• no-lyrics: Remove lyrics
• no-spam: Remove ads and spams
• no-style: Remove font style tags like <i> and <b>
• minimal: includes only ocr and tidy rules
• default: includes all rules except no-style

Usage

CLI

Clean subtitles:

$ cat mysubtitle.srt
1
00:00:46,464 --> 00:00:48,549
-And then what?
-| don't know.

2
00:49:07,278 --> 00:49:09,363
- If you cross the sea
with an army you bought ...


$ cleanit -t default mysubtitle.en.srt
1 subtitle collected / 0 subtitle filtered out / 0 path ignored
1 subtitle saved / 0 subtitle unchanged

$ cat mysubtitle.srt
1
00:00:46,464 --> 00:00:48,549
- And then what?
- I don't know.

2
00:49:07,278 --> 00:49:09,363
If you cross the sea
with an army you bought...


$ cleanit -t ocr -t no-sdh -t tidy -l en -l pt-BR ~/subtitles/
423 subtitles collected / 107 subtitles filtered out / 0 path ignored
Cleaning subtitles  [####################################]  100%
268 subtitles saved / 155 subtitles unchanged

Using docker:

$ docker run -it --rm -v /medias:/medias -u $(id -u username):$(id -g username) ratoaq2/cleanit -t default /medias
1072 subtitles collected / 0 subtitle filtered out / 0 path ignored
Cleaning subtitles  [####################################]  100%
980 subtitle saved / 92 subtitles unchanged

API

from cleanit import Config, Subtitle

sub = Subtitle('/subtitle/path/subtitle.en.srt')
cfg = Config.from_path('/config/path')
rules = cfg.select_rules(tags={'ocr'})
if sub.clean(rules):
    sub.save()

YAML Configuration file

templates:
  - &ocr
    tags:
      - ocr
      - minimal
      - default
    priority: 10000
    languages: en

rules:
  replace-l-to-I-character[ocr:en]:
    <<: *ocr
    patterns: '\bl\b'
    replacement: 'I'
    examples:
      ? |
        And if l refuse?
      : |
        And if I refuse?