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/diagnose_pdf_layout.py

@@ -0,0 +1,71 @@
+"""
+Deeper investigation: where exactly is the text on each page?
+
+Compares bounding boxes of the original text vs translated text.
+If they're in totally different positions, the layout is broken.
+"""
+import fitz
+
+orig = fitz.open("sample_files/test_corpus/test_pdf.pdf")
+trans = fitz.open("sample_files/test_corpus/test_pdf_translated.pdf")
+
+for i in range(min(3, len(orig), len(trans))):
+    o = orig[i]
+    t = trans[i]
+    print(f"\n{'=' * 70}")
+    print(f"PAGE {i+1}")
+    print('=' * 70)
+
+    # Get text blocks with positions for original
+    print(f"\n--- Original (page {i+1}) text blocks ---")
+    o_blocks = o.get_text("blocks")
+    for j, b in enumerate(o_blocks[:10]):
+        x0, y0, x1, y1, text, *_ = b
+        text_preview = text[:50].replace("\n", " ")
+        print(f"  [{j}] ({x0:.0f},{y0:.0f})-({x1:.0f},{y1:.0f}) '{text_preview}'")
+
+    print(f"\n--- Translated (page {i+1}) text blocks ---")
+    t_blocks = t.get_text("blocks")
+    for j, b in enumerate(t_blocks[:10]):
+        x0, y0, x1, y1, text, *_ = b
+        text_preview = text[:50].replace("\n", " ")
+        print(f"  [{j}] ({x0:.0f},{y0:.0f})-({x1:.0f},{y1:.0f}) '{text_preview}'")
+
+    # Check fonts
+    print(f"\n--- Fonts (page {i+1}) ---")
+    o_fonts = set()
+    t_fonts = set()
+    for blk in o_blocks:
+        for line in blk[4].split("\n") if blk[4] else []:
+            pass
+    # Get fonts from the dict format
+    o_dict = o.get_text("dict")
+    t_dict = t.get_text("dict")
+    for block in o_dict.get("blocks", []):
+        for line in block.get("lines", []):
+            for span in line.get("spans", []):
+                o_fonts.add(f"{span.get('font', '?')}@{span.get('size', 0):.1f}")
+    for block in t_dict.get("blocks", []):
+        for line in block.get("lines", []):
+            for span in line.get("spans", []):
+                t_fonts.add(f"{span.get('font', '?')}@{span.get('size', 0):.1f}")
+    print(f"  Original fonts:   {sorted(o_fonts)[:5]}")
+    print(f"  Translated fonts: {sorted(t_fonts)[:5]}")
+
+    # Check colors used
+    print(f"\n--- Colors (page {i+1}) ---")
+    o_colors = set()
+    t_colors = set()
+    for block in o_dict.get("blocks", []):
+        for line in block.get("lines", []):
+            for span in line.get("spans", []):
+                o_colors.add(span.get("color", 0))
+    for block in t_dict.get("blocks", []):
+        for line in block.get("lines", []):
+            for span in line.get("spans", []):
+                t_colors.add(span.get("color", 0))
+    print(f"  Original colors:   {sorted(o_colors)[:5]}")
+    print(f"  Translated colors: {sorted(t_colors)[:5]}")
+
+orig.close()
+trans.close()