This pull request addresses issue #13 where the parse_speech_fragments function incorrectly parses timestamps on certain locales on certain systems, such as sv_SE.UTF-8 on Tuxedo OS. The fix enforces a consistent numeric locale by setting LC_NUMERIC=C for the awk command within the function.
Changes
Modified the parse_speech_fragments function to include LC_NUMERIC=C for the awk command.
Issue Link
This pull request resolves #13.
Environment Tested
OS: Tuxedo OS (Ubuntu-based distribution)
Numeric Locale: sv_SE.UTF-8 (Swedish)
Testing Instructions
Ensure the numeric locale is set to a problematic one (e.g., LC_NUMERIC=sv_SE.UTF-8) on Tuxedo OS or possibly another Ubuntu-based distribution.
Run impd add or impd condense to condense audio from video with subtitles.
Verify that the timestamps are parsed correctly and the process completes without errors, confirming that the override works as intended.
Description
This pull request addresses issue #13 where the
parse_speech_fragmentsfunction incorrectly parses timestamps on certain locales on certain systems, such assv_SE.UTF-8on Tuxedo OS. The fix enforces a consistent numeric locale by settingLC_NUMERIC=Cfor theawkcommand within the function.Changes
parse_speech_fragmentsfunction to includeLC_NUMERIC=Cfor theawkcommand.Issue Link
This pull request resolves #13.
Environment Tested
sv_SE.UTF-8(Swedish)Testing Instructions
LC_NUMERIC=sv_SE.UTF-8) on Tuxedo OS or possibly another Ubuntu-based distribution.impd addorimpd condenseto condense audio from video with subtitles.