feat(stats): add media-detail navigation from Sessions rows; fix(tokenizer): exclude そうだ auxiliary-stem from annotations

- Added hover-revealed ↗ button on SessionRow that navigates to the
  anime media-detail view for the session's videoId
- Added `sessions` origin type to MediaDetailOrigin and
  openSessionsMediaDetail() / closeMediaDetail() handling so the
  back button returns correctly to the Sessions tab ("Back to Sessions")
- Wired onNavigateToMediaDetail down through SessionsTab → SessionRow
- Excluded tokens with MeCab POS3 = 助動詞語幹 (e.g. そうだ grammar tails)
  from subtitle annotation metadata so frequency, JLPT, and N+1 styling
  no longer apply to grammar-tail tokens
- Added annotation-stage unit test and end-to-end tokenizeSubtitle test
  for the そうだ exclusion path
- Updated docs-site changelog, immersion-tracking, and
  subtitle-annotations pages to reflect both changes

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

docs-site/subtitle-annotations.md

@@ -4,7 +4,7 @@ SubMiner annotates subtitle tokens in real time as they appear in the overlay. F
 
 All four are opt-in and configured under `subtitleStyle`, `ankiConnect.knownWords`, and `ankiConnect.nPlusOne` in your config. They apply independently — you can enable any combination.
 
-Before any of those layers render, SubMiner strips annotation metadata from tokens that are usually just subtitle glue or annotation noise. Standalone particles, auxiliaries, adnominals, common explanatory endings like `んです` / `のだ`, merged trailing quote-particle forms like `...って`, repeated kana interjections, and similar non-lexical helper tokens remain hoverable in the subtitle text, but they render as plain tokens without known-word, N+1, frequency, JLPT, or name-match annotation styling.
+Before any of those layers render, SubMiner strips annotation metadata from tokens that are usually just subtitle glue or annotation noise. Standalone particles, auxiliaries, adnominals, common explanatory endings like `んです` / `のだ`, merged trailing quote-particle forms like `...って`, auxiliary-stem grammar tails like `そうだ` (MeCab POS3 `助動詞語幹`), repeated kana interjections, and similar non-lexical helper tokens remain hoverable in the subtitle text, but they render as plain tokens without known-word, N+1, frequency, JLPT, or name-match annotation styling.
 
 ## N+1 Word Highlighting