feat(chunks): pipeline unificata Stage 1+2 con md_optimizer

chunker.py ora esegue in sequenza:
  - Stage 1 (md_optimizer.py): _content_list_v2.json + _model.json → _clean.md
    con pulizia TOC, frontespizio, sommari interni, merge titoli capitolo
  - Stage 2: _clean.md → chunks.json (paragraph-overlap, atomici tabelle/liste)

config.py esteso con CHAPTER_PREFIX_PATTERNS, SOMMARIO_PATTERNS,
MODEL_SKIP_LABELS, MODEL_ABSTRACT_LABELS, MIN_CONTENT_CHARS.

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

chunks/fix_chunks.py

@@ -34,7 +34,7 @@ if str(_HERE) not in sys.path:
 import config as cfg
 from verify_chunks import verify_stem as _verify_stem
 
-MAX_CHARS = int(cfg.TARGET_CHARS * (1 + cfg.CHUNK_TOLERANCE))
+MAX_CHARS = cfg.MAX_CHARS
 
 
 def _load_thresholds(stem_dir: Path) -> int:
@@ -314,8 +314,8 @@ def fix_stem(stem: str, project_root: Path, max_chars: int, dry_run: bool,
     # Fase 1: risolvi blockers a convergenza (solo merge incomplete)
     chunks = _fix_blockers(chunks, report)
 
-    _min = int(cfg.TARGET_CHARS * (1 - cfg.CHUNK_TOLERANCE))
-    _max = int(cfg.TARGET_CHARS * (1 + cfg.CHUNK_TOLERANCE))
+    _min = cfg.MIN_CHARS
+    _max = cfg.MAX_CHARS
     prev_blockers = sum(len(v) for v in report.get("blockers", {}).values())
 
     for iteration in range(1, max_iter + 1):
@@ -378,7 +378,7 @@ if __name__ == "__main__":
 
     parser = argparse.ArgumentParser(description="Fix chunk")
     parser.add_argument("--stem", required=True, help="Nome del documento (sottocartella di chunks/)")
-    _max_def = int(cfg.TARGET_CHARS * (1 + cfg.CHUNK_TOLERANCE))
+    _max_def = cfg.MAX_CHARS
     parser.add_argument(
         "--max", type=int, default=_max_def,
         help=f"Soglia massima caratteri per lo split (default: TARGET×(1+TOL) = {_max_def})"