Improve Java comment filtering

[Vladyslav-Kuksiuk]

This PR improves Java comment filtering.

Resolves this issue

[Copilot]

Pull request overview

This PR improves embedding/commentfilter so comment stripping is more language-accurate, notably by teaching the marker-based filter to treat Java """ text blocks as multi-line literals (so // and /* ... */ inside the text block are preserved when comments="none").

Changes:

• Extend MarkerCommentFilter to track multi-line “text block” literals (e.g., Java """) and skip comment scanning while inside them.
• Add dedicated comment filters for Kotlin (nested block comments + raw """ strings with ${...} interpolation) and Visual Basic (apostrophe / doc / REM comments).
• Add regression tests covering Java text blocks, Kotlin raw strings/interpolation, and Visual Basic comment forms.

Reviewed changes

Copilot reviewed 3 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
embedding/commentfilter/marker_comment_filter.go	Adds text-block state tracking and scanning helpers to avoid stripping comment markers inside configured multi-line literals.
embedding/commentfilter/config.go	Registers Java """ as a text-block delimiter; wires Kotlin/VB extensions to their new filters.
embedding/commentfilter/kotlin_filter.go	Implements Kotlin-aware comment filtering (nested block comments, raw strings, interpolation).
embedding/commentfilter/visual_basic_filter.go	Implements Visual Basic-aware comment filtering (doc ''', apostrophe, and REM).
embedding/commentfilter/filter_test.go	Adds/extends tests for Java text blocks, Kotlin raw strings/interpolation, and Visual Basic comment behavior.