feat(format): B3.5 — PDF smart-fit rewrite + critical fontname=None fix

ROOT CAUSE FIX: PyMuPDF silently raised AttributeError when fontname=None
was passed to insert_textbox. The try/except in _try_insert was swallowing
the error and returning None, causing every block to be skipped via the
graceful failure path. Setting fontname='helv' as the default unblocks
the entire PDF translation pipeline.

SMART-FIT: rewrite _write_translated_block with proper tier-fallback:
  - Tier 0: original bbox at original size
  - Tier 1: expanded horizontal
  - Tier 2: expanded vertical (3x original height)
  - Tier 3: shrink once (0.93x)
  - Tier 4: shrink twice (0.87x cumulative)
  - Tier 5: min size floor (90% for headings, 75% for body)
  - Tier 6: graceful skip with visible placeholder

REDACTION: single redaction per block (was per sub-bbox, creating 100+
redaction rectangles per page). Now only 1 redaction per text block.

FEATURE FLAG: PDF_SMART_FIT_ENABLED (default true, observation-first).

METRICS: text_overflow -> format_elements_lost_total.

RESULT ON REAL PDF:
  Before: fonts shrunk 22pt->5.6pt, hierarchy destroyed
  After:  fonts EXACT match: [8, 11, 12, 14, 16, 22] preserved

scripts/compare_pdfs.py

@@ -0,0 +1,107 @@
+"""
+Deep comparison of original vs translated PDF.
+
+Outputs a detailed diff to identify exactly what format elements were lost.
+"""
+import fitz
+from pathlib import Path
+import sys
+
+original = Path("sample_files/test_corpus/test_pdf.pdf")
+translated = Path("sample_files/test_corpus/test_pdf_translated.pdf")
+
+print(f"Original:   {original.stat().st_size:>10} bytes")
+print(f"Translated: {translated.stat().st_size:>10} bytes")
+print(f"Size diff:  {translated.stat().st_size - original.stat().st_size:>10} bytes ({100 * translated.stat().st_size / original.stat().st_size:.1f}%)")
+print()
+
+doc_orig = fitz.open(str(original))
+doc_trans = fitz.open(str(translated))
+
+print("=" * 70)
+print("PAGE COUNT")
+print("=" * 70)
+print(f"  Original:   {len(doc_orig)} pages")
+print(f"  Translated: {len(doc_trans)} pages")
+print()
+
+print("=" * 70)
+print("PER-PAGE COMPARISON")
+print("=" * 70)
+for i in range(max(len(doc_orig), len(doc_trans))):
+    o_page = doc_orig[i] if i < len(doc_orig) else None
+    t_page = doc_trans[i] if i < len(doc_trans) else None
+    print(f"\n--- Page {i+1} ---")
+    if o_page:
+        o_text = len(o_page.get_text())
+        o_links = len(o_page.get_links())
+        o_images = len(o_page.get_images())
+        o_drawings = len(o_page.get_drawings())
+        print(f"  Original:   {o_text:>4} chars, {o_links:>2} links, {o_images:>2} images, {o_drawings:>3} drawings")
+    if t_page:
+        t_text = len(t_page.get_text())
+        t_links = len(t_page.get_links())
+        t_images = len(t_page.get_images())
+        t_drawings = len(t_page.get_drawings())
+        print(f"  Translated: {t_text:>4} chars, {t_links:>2} links, {t_images:>2} images, {t_drawings:>3} drawings")
+    if o_page and t_page:
+        print(f"  DIFF:       text {t_text - o_text:+d}, links {t_links - o_links:+d}, images {t_images - o_images:+d}, drawings {t_drawings - o_drawings:+d}")
+
+print()
+print("=" * 70)
+print("HYPERLINK COMPARISON")
+print("=" * 70)
+for i in range(max(len(doc_orig), len(doc_trans))):
+    o_page = doc_orig[i] if i < len(doc_orig) else None
+    t_page = doc_trans[i] if i < len(doc_trans) else None
+    o_links = o_page.get_links() if o_page else []
+    t_links = t_page.get_links() if t_page else []
+    if o_links or t_links:
+        print(f"\n  Page {i+1}:")
+        print(f"    Original has {len(o_links)} links:")
+        for link in o_links:
+            kind = "URI" if link.get("uri") else "GOTO" if link.get("page") is not None else "?"
+            uri_or_page = link.get("uri") or f"page {link.get('page')}"
+            print(f"      - {kind}: {uri_or_page[:80]}")
+        print(f"    Translated has {len(t_links)} links:")
+        for link in t_links:
+            kind = "URI" if link.get("uri") else "GOTO" if link.get("page") is not None else "?"
+            uri_or_page = link.get("uri") or f"page {link.get('page')}"
+            print(f"      - {kind}: {uri_or_page[:80]}")
+
+print()
+print("=" * 70)
+print("IMAGE COMPARISON")
+print("=" * 70)
+total_o_imgs = sum(len(p.get_images()) for p in doc_orig)
+total_t_imgs = sum(len(p.get_images()) for p in doc_trans)
+print(f"  Original:   {total_o_imgs} images")
+print(f"  Translated: {total_t_imgs} images")
+if total_o_imgs > 0 and total_t_imgs == 0:
+    print("  ❌ ALL IMAGES LOST in translation!")
+
+print()
+print("=" * 70)
+print("DRAWING COMPARISON (rects, lines, etc.)")
+print("=" * 70)
+total_o_dr = sum(len(p.get_drawings()) for p in doc_orig)
+total_t_dr = sum(len(p.get_drawings()) for p in doc_trans)
+print(f"  Original:   {total_o_dr} drawings")
+print(f"  Translated: {total_t_dr} drawings")
+
+print()
+print("=" * 70)
+print("SAMPLE TEXT — Page 1 of translated")
+print("=" * 70)
+if len(doc_trans) > 0:
+    print(doc_trans[0].get_text()[:500])
+
+print()
+print("=" * 70)
+print("SAMPLE TEXT — Page 1 of original")
+print("=" * 70)
+if len(doc_orig) > 0:
+    print(doc_orig[0].get_text()[:500])
+
+doc_orig.close()
+doc_trans.close()