feat(ocr): unify config, add manga translation pipeline and context lookahead

- Consolidate module configs into root config.example.json with ocr, vision, and epub sections

- Split LLM OCR workflows into novel_ocr.py (prose) and manga_ocr_llm.py (manga)

- Remove gemini_direct_ocr.py in favor of OpenAI-compatible API endpoints

- Support direct manga translation via --translate, --target-lang, and glossary.md

- Add bidirectional context support: past translations (--context-pages) and lookahead Japanese text (--context-pages-ahead)

- Add per-page JSON audit logging under logs/ and expose OpenAI sampling parameters

ocr/glossary.example.md

@@ -0,0 +1,25 @@
+# Glossary
+
+A glossary of established translations for names, recurring terms, and any
+standing style notes. Copy this file to `glossary.md` in this same `ocr/`
+folder (next to the scripts — not with your page images) and it's picked
+up automatically whenever you run any OCR script with `--translate` (or
+point at a different file explicitly with `--glossary`).
+
+The exact Markdown structure below isn't parsed or validated — it's sent to
+the model as-is, so feel free to add/remove sections or entries. Keep the
+`## Names` / `## Terms` / `## Notes` headings if you want a rough separation
+between "must stay consistent" name spellings, recurring in-universe terms,
+and general translation style notes.
+
+## Names
+
+- 澤村・スペンサー・英梨々: Савамура Спенсер Эрири
+
+## Terms
+
+- ビジュアルノベル: Визуальная новелла
+
+## Notes
+
+- Keep Japanese suffixes (e.g. -chan, -san).