Round 2 Word bisection: isolate the field-run co-factor

bisect_03 (fields removed) was the only round-1 file Word opened, and
document.xml becomes byte-identical between the full book and bisect_03
once field runs are removed — but the mini book opens with identical
PAGEREF fields, so the fields interact with a second factor. Add six
candidates that keep the 28 PAGEREF fields while peeling other layers:

- 07 drop deep-copied chapter prose (keep titles/bookmarks/fields)
- 08 drop numbering properties
- 09 inline hyperlink wrappers
- 10 remove the whole TOC section (cleaner no-fields control)
- 11 rewrite all PAGEREF targets to a single bookmark
- 12 keep fields only for the first 3 TOC entries

tests/make_bisect_files.py

@@ -41,6 +41,12 @@ NAMES = [
     'bisect_04_no_styles.docx',
     'bisect_05_first8.docx',
     'bisect_06_no_images.docx',
+    'bisect_07_no_prose.docx',
+    'bisect_08_no_numbering.docx',
+    'bisect_09_no_hyperlinks.docx',
+    'bisect_10_no_toc_section.docx',
+    'bisect_11_same_target_fields.docx',
+    'bisect_12_first3_toc.docx',
 ]
 
 
@@ -116,6 +122,85 @@ def make_no_images(doc):
             p.getparent().remove(p)
 
 
+def _sections(doc):
+    """Return list of [start, end) kid-slices, one per body section,
+    using the in-paragraph sectPr holder paragraphs as boundaries."""
+    body = doc.element.body
+    holders = []
+    for s in body.iter(qn('w:sectPr')):
+        par = s.getparent().getparent()
+        if par is not None and par.getparent() is body:
+            holders.append(par)
+    kids = [c for c in body if c.tag == qn('w:p')]
+    secs, start = [], 0
+    for h in holders:
+        hi = kids.index(h)
+        secs.append((start, hi + 1))
+        start = hi + 1
+    if start < len(kids):
+        secs.append((start, len(kids)))
+    return kids, secs
+
+
+def make_no_prose(doc):
+    """Drop the deep-copied chapter prose (sections 5..32) while keeping each
+    chapter's title paragraph (with its bookmark) and section holder."""
+    kids, secs = _sections(doc)
+    body = doc.element.body
+    drop = []
+    for ci, (a, b) in enumerate(secs):
+        if 5 <= ci <= 32:  # chapters 1..28
+            drop.extend(kids[a + 1:b - 1])  # keep title [a], holder [b-1]
+    for k in drop:
+        body.remove(k)
+
+
+def make_no_numbering(doc):
+    """Remove all list-numbering properties (keeps fields)."""
+    for np_ in list(doc.element.body.iter(qn('w:numPr'))):
+        np_.getparent().remove(np_)
+
+
+def make_no_hyperlinks(doc):
+    """Remove w:hyperlink wrappers, inlining their runs (keeps fields)."""
+    for hl in list(doc.element.body.iter(qn('w:hyperlink'))):
+        par = hl.getparent()
+        anchor = hl
+        for r in list(hl):
+            par.addnext(r)
+        par.remove(hl)
+
+
+def make_no_toc_section(doc):
+    """Remove the whole TOC section (heading + entries + its section holder)."""
+    kids, secs = _sections(doc)
+    body = doc.element.body
+    a, b = secs[4]  # TOC section
+    for k in kids[a:b]:
+        body.remove(k)
+
+
+def make_same_target_fields(doc):
+    """Rewrite every PAGEREF target to _UF_ch_0 (uniform targets, fields kept)."""
+    for it in doc.element.body.iter(qn('w:instrText')):
+        if (it.text or '').strip().startswith('PAGEREF'):
+            it.text = ' PAGEREF _UF_ch_0 '
+
+
+def make_first3_toc(doc):
+    """Keep only the first 3 TOC entries' fields; plain-text the rest."""
+    seen = 0
+    for p in doc.element.body.iter(qn('w:p')):
+        if p.find(".//" + qn('w:instrText')) is None:
+            continue
+        if seen < 3:
+            seen += 1
+            continue
+        for r in list(p.findall(qn('w:r'))):
+            if r.find(qn('w:fldChar')) is not None or r.find(qn('w:instrText')) is not None:
+                p.remove(r)
+
+
 MAKERS = [
     make_no_running,
     make_no_oddpage,
@@ -123,6 +208,12 @@ MAKERS = [
     make_no_styles,
     make_first8,
     make_no_images,
+    make_no_prose,
+    make_no_numbering,
+    make_no_hyperlinks,
+    make_no_toc_section,
+    make_same_target_fields,
+    make_first3_toc,
 ]