fix(subtitle): handle VTT positioning directives in subtitle parsing

[abdelaziz-mahdy]

This pull request improves the handling of VTT subtitle parsing in the SubtitleParser class to better support VTT-specific directives. The main change is to ensure that positioning and styling directives in VTT subtitle lines are skipped, so only the actual subtitle text is extracted.

Subtitle parsing improvements:

• Updated SubtitleParser logic to detect VTT directives (e.g., line, align, position, etc.) in group 10 and, when present, extract subtitle text from group 11 instead. This prevents directives from being misinterpreted as subtitle text.

Test case update:

• Modified the VTT subtitle test input to include a line: 20% directive, ensuring the parser correctly skips the directive and extracts only the subtitle text.