microsoft-word 0.0.0-stage → 0.1.4
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/AGENTS.md +10 -0
- package/CHANGELOG.md +129 -0
- package/LICENSE +21 -0
- package/README.md +66 -2
- package/SKILL.md +177 -0
- package/package.json +62 -4
- package/references/automation.md +74 -0
- package/references/conversion.md +53 -0
- package/references/extraction.md +63 -0
- package/references/fallbacks.md +61 -0
- package/references/libraries.md +125 -0
- package/references/ooxml.md +111 -0
- package/references/recipes.md +144 -0
- package/references/security.md +88 -0
- package/references/troubleshooting.md +23 -0
- package/references/word-ui.md +40 -0
- package/scripts/docx_refresh.py +423 -0
- package/scripts/docx_replace.py +321 -0
- package/scripts/docx_sanitize.py +507 -0
- package/scripts/inspect_docx.py +723 -0
- package/scripts/requirements.txt +1 -0
- package/scripts/wordpkg.py +328 -0
- package/skill.json +55 -0
|
@@ -0,0 +1,507 @@
|
|
|
1
|
+
#!/usr/bin/env python3
|
|
2
|
+
"""Make a clean copy of a Word document for sharing: accept tracked changes, remove comments,
|
|
3
|
+
hidden text, glossary, personal metadata and active content. Then re-inspect the copy and
|
|
4
|
+
report what remains.
|
|
5
|
+
|
|
6
|
+
Usage:
|
|
7
|
+
python docx_sanitize.py IN OUT [--accept-changes] [--remove-comments] [--strip-metadata]
|
|
8
|
+
[--remove-hidden] [--remove-glossary] [--remove-active]
|
|
9
|
+
[--all | --share] [--json]
|
|
10
|
+
|
|
11
|
+
What each option does (only the parts it names are rewritten; other parts keep their bytes):
|
|
12
|
+
--accept-changes w:ins and w:moveTo are unwrapped; w:del, w:moveFrom and move range marks are
|
|
13
|
+
removed; formatting changes (rPrChange, pPrChange, sectPrChange, tbl/tr/tc
|
|
14
|
+
PrChange, numberingChange) are dropped so the current formatting stays; a
|
|
15
|
+
deleted paragraph mark merges the paragraph into the next one; deleted table
|
|
16
|
+
rows are removed. Cell-level insert/delete/merge marks are not handled and
|
|
17
|
+
are reported as residue: finish those in Word.
|
|
18
|
+
--remove-comments Comment ranges, references and the comments, commentsExtended, commentsIds,
|
|
19
|
+
commentsExtensible and people parts with their relationships.
|
|
20
|
+
--strip-metadata docProps/core.xml creator, lastModifiedBy, lastPrinted, revision, created,
|
|
21
|
+
modified; app.xml Company, Manager, Template, TotalTime, HyperlinkBase,
|
|
22
|
+
Application, AppVersion; all custom properties; the thumbnail
|
|
23
|
+
(docProps/thumbnail.*, a picture of page 1); w:rsids and every w:rsid*
|
|
24
|
+
attribute; w:docVars; the attached template reference; customXml parts
|
|
25
|
+
(SharePoint and data-binding properties) and w:dataBinding links to them.
|
|
26
|
+
Title, subject, keywords, description, category and contentStatus are kept
|
|
27
|
+
and listed: check them yourself.
|
|
28
|
+
--remove-hidden Runs hidden by w:vanish, w:specVanish or w:webHidden, set directly or through
|
|
29
|
+
character styles, paragraph styles (basedOn chains) or docDefaults, in the
|
|
30
|
+
body, headers, footers, notes and the glossary. Hidden runs that hold field
|
|
31
|
+
codes are kept (removing them breaks the field) and reported; exit 1.
|
|
32
|
+
--remove-glossary The glossary document (building blocks, AutoText): its text travels with the
|
|
33
|
+
file but is never shown in the body.
|
|
34
|
+
--remove-active vbaProject.bin and vbaData.xml (a .docm becomes a .docx: name OUT .docx),
|
|
35
|
+
ActiveX controls, OLE objects (embedded and linked, with their pictures),
|
|
36
|
+
altChunk imports, every external relationship except hyperlinks (remote
|
|
37
|
+
template, linked images and beacons, frames, subdocuments) and the elements
|
|
38
|
+
that use them, the mail merge data source and updateFields. DDE, DDEAUTO,
|
|
39
|
+
INCLUDETEXT, INCLUDEPICTURE, INCLUDE, IMPORT, LINK, DATABASE, MACROBUTTON,
|
|
40
|
+
AUTOOPEN, AUTOEXEC fields, and fields built from nested fields or QUOTE
|
|
41
|
+
character codes, are replaced by their stored result text.
|
|
42
|
+
--all All of the above. Exits 1 if any risk remains after the re-inspection.
|
|
43
|
+
--share --all, and exits 1 if any residue remains too (for example visually masked
|
|
44
|
+
text, which only a person can decide to delete).
|
|
45
|
+
|
|
46
|
+
This is not full redaction. It does not touch images (EXIF, cropped areas), charts, alt text,
|
|
47
|
+
text the user covered with black highlight, white font or shapes, or the visible text itself.
|
|
48
|
+
The final inspection lists what remains; read it before sharing.
|
|
49
|
+
|
|
50
|
+
Safety: OUT must not exist and must differ from IN. A macro-enabled input with a .docx OUT
|
|
51
|
+
is refused unless --remove-active is given, and --remove-active refuses a .docm or .dotm OUT
|
|
52
|
+
(Word rejects a file whose extension and content type disagree).
|
|
53
|
+
Exit codes: 0 clean for the requested categories (and no risks with --remove-active, no
|
|
54
|
+
residue with --share), 1 something requested remains (see "remaining", "risks_after"),
|
|
55
|
+
2 unsupported or unreadable input, 3 safety refusal (output exists, size cap, DOCTYPE,
|
|
56
|
+
duplicate zip entries, extension mismatch). Requires: pip install lxml
|
|
57
|
+
"""
|
|
58
|
+
import argparse
|
|
59
|
+
import json
|
|
60
|
+
import os
|
|
61
|
+
import sys
|
|
62
|
+
|
|
63
|
+
import inspect_docx
|
|
64
|
+
import wordpkg as wp
|
|
65
|
+
from wordpkg import NS, clean, w
|
|
66
|
+
|
|
67
|
+
PROP_CHANGES = ["rPrChange", "pPrChange", "sectPrChange", "tblPrChange", "trPrChange",
|
|
68
|
+
"tcPrChange", "tblGridChange", "tblPrExChange", "numberingChange"]
|
|
69
|
+
MOVE_MARKS = ["moveFromRangeStart", "moveFromRangeEnd", "moveToRangeStart", "moveToRangeEnd"]
|
|
70
|
+
COMMENT_TYPES = {"comments", "commentsExtended", "commentsIds", "commentsExtensible", "people"}
|
|
71
|
+
OFF = {"0", "false", "off"}
|
|
72
|
+
|
|
73
|
+
|
|
74
|
+
def unwrap(el):
|
|
75
|
+
parent = el.getparent()
|
|
76
|
+
idx = parent.index(el)
|
|
77
|
+
for i, c in enumerate(list(el)):
|
|
78
|
+
parent.insert(idx + i, c)
|
|
79
|
+
# keep tail text (whitespace) attached
|
|
80
|
+
parent.remove(el)
|
|
81
|
+
|
|
82
|
+
|
|
83
|
+
def accept(root):
|
|
84
|
+
n = 0
|
|
85
|
+
for tag in PROP_CHANGES:
|
|
86
|
+
for el in list(root.iter(w(tag))):
|
|
87
|
+
el.getparent().remove(el)
|
|
88
|
+
n += 1
|
|
89
|
+
for tag in MOVE_MARKS:
|
|
90
|
+
for el in list(root.iter(w(tag))):
|
|
91
|
+
el.getparent().remove(el)
|
|
92
|
+
# paragraph marks and table rows first (they sit in pPr/rPr and trPr)
|
|
93
|
+
for mark in list(root.iter(w("ins"))):
|
|
94
|
+
par = mark.getparent()
|
|
95
|
+
if par is not None and par.tag in (w("rPr"), w("trPr")):
|
|
96
|
+
par.remove(mark)
|
|
97
|
+
n += 1
|
|
98
|
+
for mark in list(root.iter(w("del"))):
|
|
99
|
+
par = mark.getparent()
|
|
100
|
+
if par is None:
|
|
101
|
+
continue
|
|
102
|
+
if par.tag == w("trPr"):
|
|
103
|
+
row = par.getparent()
|
|
104
|
+
row.getparent().remove(row)
|
|
105
|
+
n += 1
|
|
106
|
+
elif par.tag == w("rPr") and par.getparent() is not None and par.getparent().tag == w("pPr"):
|
|
107
|
+
p = par.getparent().getparent()
|
|
108
|
+
par.remove(mark)
|
|
109
|
+
nxt = p.getnext()
|
|
110
|
+
if nxt is not None and nxt.tag == w("p"):
|
|
111
|
+
# the deleted mark joins this paragraph's content to the next paragraph,
|
|
112
|
+
# which keeps its own paragraph properties (its mark survives)
|
|
113
|
+
kids = [c for c in p if c.tag != w("pPr")]
|
|
114
|
+
npr = nxt.find("w:pPr", NS)
|
|
115
|
+
pos = 0 if npr is None else nxt.index(npr) + 1
|
|
116
|
+
for i, c in enumerate(kids):
|
|
117
|
+
nxt.insert(pos + i, c)
|
|
118
|
+
p.getparent().remove(p)
|
|
119
|
+
n += 1
|
|
120
|
+
for el in list(root.iter(w("del"), w("moveFrom"))):
|
|
121
|
+
if el.getparent() is not None:
|
|
122
|
+
el.getparent().remove(el)
|
|
123
|
+
n += 1
|
|
124
|
+
for el in list(root.iter(w("ins"), w("moveTo"))):
|
|
125
|
+
if el.getparent() is not None:
|
|
126
|
+
unwrap(el)
|
|
127
|
+
n += 1
|
|
128
|
+
return n
|
|
129
|
+
|
|
130
|
+
|
|
131
|
+
def remove_comment_marks(root):
|
|
132
|
+
n = 0
|
|
133
|
+
for el in list(root.iter(w("commentRangeStart"), w("commentRangeEnd"))):
|
|
134
|
+
el.getparent().remove(el)
|
|
135
|
+
for ref in list(root.iter(w("commentReference"))):
|
|
136
|
+
r = ref.getparent()
|
|
137
|
+
r.remove(ref)
|
|
138
|
+
n += 1
|
|
139
|
+
if r.tag == w("r") and all(c.tag == w("rPr") for c in r):
|
|
140
|
+
r.getparent().remove(r)
|
|
141
|
+
return n
|
|
142
|
+
|
|
143
|
+
|
|
144
|
+
def remove_hidden(root, hid):
|
|
145
|
+
"""Remove hidden runs. Returns (removed, kept_field_runs)."""
|
|
146
|
+
n = kept = 0
|
|
147
|
+
for r in list(root.iter(w("r"))):
|
|
148
|
+
if r.getparent() is None or not hid.run_hidden(r):
|
|
149
|
+
continue
|
|
150
|
+
if r.find("w:fldChar", NS) is not None or r.find("w:instrText", NS) is not None:
|
|
151
|
+
kept += 1 # never break field structure; reported instead
|
|
152
|
+
continue
|
|
153
|
+
r.getparent().remove(r)
|
|
154
|
+
n += 1
|
|
155
|
+
return n, kept
|
|
156
|
+
|
|
157
|
+
|
|
158
|
+
def neutralize_fields(root):
|
|
159
|
+
"""Replace risky fields by their stored result: drop the begin..separate runs (the
|
|
160
|
+
instruction, including fields nested in it) and the end run; unwrap risky fldSimple."""
|
|
161
|
+
fields = [f for f in inspect_docx.fields_of(root) if inspect_docx.field_risk(f)]
|
|
162
|
+
if not fields:
|
|
163
|
+
return 0
|
|
164
|
+
runs = list(root.iter(w("r")))
|
|
165
|
+
idx = {id(r): i for i, r in enumerate(runs)}
|
|
166
|
+
drop = set()
|
|
167
|
+
n = 0
|
|
168
|
+
for f in fields:
|
|
169
|
+
if "simple" in f:
|
|
170
|
+
el = f["simple"]
|
|
171
|
+
if el.getparent() is not None:
|
|
172
|
+
unwrap(el)
|
|
173
|
+
n += 1
|
|
174
|
+
continue
|
|
175
|
+
if f["end"] is None:
|
|
176
|
+
continue
|
|
177
|
+
ib = idx.get(id(f["begin"].getparent()))
|
|
178
|
+
ie = idx.get(id(f["end"].getparent()))
|
|
179
|
+
isep = idx.get(id(f["sep"].getparent())) if f["sep"] is not None else ie
|
|
180
|
+
if ib is None or ie is None or isep is None:
|
|
181
|
+
continue
|
|
182
|
+
drop.update(range(ib, isep + 1))
|
|
183
|
+
drop.add(ie)
|
|
184
|
+
n += 1
|
|
185
|
+
for i in sorted(drop):
|
|
186
|
+
r = runs[i]
|
|
187
|
+
if r.getparent() is not None:
|
|
188
|
+
r.getparent().remove(r)
|
|
189
|
+
return n
|
|
190
|
+
|
|
191
|
+
|
|
192
|
+
def rels_source(rels_name):
|
|
193
|
+
"""word/_rels/document.xml.rels -> word/document.xml; _rels/.rels -> '' (the package)."""
|
|
194
|
+
return rels_name.replace("_rels/", "").rsplit(".rels", 1)[0]
|
|
195
|
+
|
|
196
|
+
|
|
197
|
+
CORE_DROP = ("creator", "lastModifiedBy", "lastPrinted", "revision", "created", "modified")
|
|
198
|
+
CORE_KEEP = ("title", "subject", "keywords", "description", "category", "contentStatus")
|
|
199
|
+
APP_DROP = ("Company", "Manager", "Template", "TotalTime", "HyperlinkBase", "Application", "AppVersion")
|
|
200
|
+
CATEGORIES = ("accept-changes", "remove-comments", "strip-metadata", "remove-hidden", "remove-glossary",
|
|
201
|
+
"remove-active")
|
|
202
|
+
PLAIN_TYPE = {"docm": "application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml",
|
|
203
|
+
"dotm": "application/vnd.openxmlformats-officedocument.wordprocessingml.template.main+xml"}
|
|
204
|
+
VML_O = "urn:schemas-microsoft-com:office:office"
|
|
205
|
+
R_ATTR = f"{{{wp.R}}}"
|
|
206
|
+
|
|
207
|
+
|
|
208
|
+
def main():
|
|
209
|
+
ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
|
|
210
|
+
ap.add_argument("input")
|
|
211
|
+
ap.add_argument("output")
|
|
212
|
+
for f in CATEGORIES + ("all", "share"):
|
|
213
|
+
ap.add_argument(f"--{f}", action="store_true")
|
|
214
|
+
ap.add_argument("--json", action="store_true")
|
|
215
|
+
ap.add_argument("--max-mb", type=int, default=200)
|
|
216
|
+
a = ap.parse_args()
|
|
217
|
+
if a.all or a.share:
|
|
218
|
+
for f in CATEGORIES:
|
|
219
|
+
setattr(a, f.replace("-", "_"), True)
|
|
220
|
+
|
|
221
|
+
def done(code, **kw):
|
|
222
|
+
if a.json:
|
|
223
|
+
print(json.dumps(kw, ensure_ascii=False, indent=1))
|
|
224
|
+
elif not kw.get("ok"):
|
|
225
|
+
sys.stderr.write(f"error: {kw.get('error')}\n")
|
|
226
|
+
if kw.get("remaining"):
|
|
227
|
+
sys.stderr.write(f"remaining residue: {kw['remaining']}\n")
|
|
228
|
+
if kw.get("risks_after"):
|
|
229
|
+
sys.stderr.write(f"remaining risks: {kw['risks_after']}\n")
|
|
230
|
+
else:
|
|
231
|
+
print(f"wrote {kw['output']}: {kw['actions']}")
|
|
232
|
+
print(f"residue still present (not requested or not handled): {kw['remaining'] or 'none'}")
|
|
233
|
+
print(f"risks still present: {kw['risks_after'] or 'none'}")
|
|
234
|
+
if kw.get("kept_metadata"):
|
|
235
|
+
print(f"kept metadata, review it: {kw['kept_metadata']}")
|
|
236
|
+
sys.exit(code)
|
|
237
|
+
|
|
238
|
+
if not any(getattr(a, f.replace("-", "_")) for f in CATEGORIES):
|
|
239
|
+
done(2, ok=False, error="choose at least one of " + " ".join("--" + f for f in CATEGORIES)
|
|
240
|
+
+ " --all --share")
|
|
241
|
+
if wp.same_file(a.input, a.output):
|
|
242
|
+
done(3, ok=False, error="OUT is the same file as IN; write to a new path")
|
|
243
|
+
out_ext = os.path.splitext(a.output)[1].lower()
|
|
244
|
+
try:
|
|
245
|
+
pkg = wp.Package(a.input, max_mb=a.max_mb)
|
|
246
|
+
if pkg.kind in ("docm", "dotm") and out_ext in (".docx", ".dotx") and not a.remove_active:
|
|
247
|
+
raise wp.Refusal(f"input is a macro-enabled {pkg.kind} but OUT is {out_ext}: add --remove-active "
|
|
248
|
+
f"to drop the macros, or name OUT .{pkg.kind}")
|
|
249
|
+
if a.remove_active and out_ext in (".docm", ".dotm"):
|
|
250
|
+
raise wp.Refusal("--remove-active writes a macro-free document: name OUT .docx (or .dotx)")
|
|
251
|
+
replaced, removed, actions = {}, set(), {}
|
|
252
|
+
main_rels_name = pkg.rels_name(pkg.main)
|
|
253
|
+
stories = pkg.story_parts(include_comments=False)
|
|
254
|
+
glossary = pkg.glossary_parts()
|
|
255
|
+
xml_parts = [n for n in pkg.names if n.startswith("word/") and n.endswith(".xml")]
|
|
256
|
+
touched = set()
|
|
257
|
+
hidden_field_runs = 0
|
|
258
|
+
if a.accept_changes:
|
|
259
|
+
c = 0
|
|
260
|
+
for part in xml_parts:
|
|
261
|
+
k = accept(pkg.xml(part))
|
|
262
|
+
if k:
|
|
263
|
+
touched.add(part)
|
|
264
|
+
c += k
|
|
265
|
+
actions["tracked_changes_accepted"] = c
|
|
266
|
+
if a.remove_comments:
|
|
267
|
+
c = 0
|
|
268
|
+
for part in stories + glossary:
|
|
269
|
+
k = remove_comment_marks(pkg.xml(part))
|
|
270
|
+
if k:
|
|
271
|
+
touched.add(part)
|
|
272
|
+
c += k
|
|
273
|
+
drop = {r["id"]: r["target"] for r in pkg.rels(pkg.main) if r["type"] in COMMENT_TYPES}
|
|
274
|
+
if drop:
|
|
275
|
+
rels = pkg.xml(main_rels_name)
|
|
276
|
+
for r in list(rels.findall("pr:Relationship", NS)):
|
|
277
|
+
if r.get("Id") in drop:
|
|
278
|
+
rels.remove(r)
|
|
279
|
+
touched.add(main_rels_name)
|
|
280
|
+
removed |= set(drop.values())
|
|
281
|
+
actions["comment_references_removed"] = c
|
|
282
|
+
actions["comment_parts_removed"] = sorted(set(drop.values()))
|
|
283
|
+
if a.remove_hidden:
|
|
284
|
+
hid = wp.Hidden(pkg)
|
|
285
|
+
c = 0
|
|
286
|
+
for part in stories + glossary:
|
|
287
|
+
k, kept_f = remove_hidden(pkg.xml(part), hid)
|
|
288
|
+
hidden_field_runs += kept_f
|
|
289
|
+
if k:
|
|
290
|
+
touched.add(part)
|
|
291
|
+
c += k
|
|
292
|
+
actions["hidden_runs_removed"] = c
|
|
293
|
+
actions["hidden_field_code_runs_kept"] = hidden_field_runs
|
|
294
|
+
if a.remove_glossary and glossary:
|
|
295
|
+
removed |= set(glossary)
|
|
296
|
+
removed |= {n for n in pkg.names if n.startswith("word/glossary/")}
|
|
297
|
+
actions["glossary_parts_removed"] = sorted(removed & set(glossary))
|
|
298
|
+
kept = {}
|
|
299
|
+
if a.strip_metadata:
|
|
300
|
+
stripped = []
|
|
301
|
+
for r in pkg.rels(""):
|
|
302
|
+
if r["external"] or r["target"] not in pkg.names:
|
|
303
|
+
continue
|
|
304
|
+
if r["type"] == "core-properties":
|
|
305
|
+
x = pkg.xml(r["target"])
|
|
306
|
+
for el in list(x):
|
|
307
|
+
k = el.tag.rsplit("}", 1)[-1]
|
|
308
|
+
if k in CORE_DROP:
|
|
309
|
+
x.remove(el)
|
|
310
|
+
stripped.append(k)
|
|
311
|
+
elif k in CORE_KEEP and (el.text or "").strip():
|
|
312
|
+
kept[k] = clean(el.text, 120)
|
|
313
|
+
touched.add(r["target"])
|
|
314
|
+
elif r["type"] == "extended-properties":
|
|
315
|
+
x = pkg.xml(r["target"])
|
|
316
|
+
for el in list(x):
|
|
317
|
+
k = el.tag.rsplit("}", 1)[-1]
|
|
318
|
+
if k in APP_DROP:
|
|
319
|
+
x.remove(el)
|
|
320
|
+
stripped.append(k)
|
|
321
|
+
touched.add(r["target"])
|
|
322
|
+
elif r["type"] == "custom-properties":
|
|
323
|
+
x = pkg.xml(r["target"])
|
|
324
|
+
for el in list(x):
|
|
325
|
+
x.remove(el)
|
|
326
|
+
stripped.append("custom properties")
|
|
327
|
+
touched.add(r["target"])
|
|
328
|
+
elif r["type"] == "thumbnail":
|
|
329
|
+
removed.add(r["target"])
|
|
330
|
+
removed |= {n for n in pkg.names if n.lower().startswith("docprops/thumbnail")}
|
|
331
|
+
removed |= {n for n in pkg.names if n.startswith("customXml/")}
|
|
332
|
+
rsid_n = 0
|
|
333
|
+
for part in xml_parts:
|
|
334
|
+
root = pkg.xml(part)
|
|
335
|
+
hit = False
|
|
336
|
+
for el in root.iter():
|
|
337
|
+
for k in [k for k in el.attrib if k.startswith(f"{{{wp.W}}}rsid")]:
|
|
338
|
+
del el.attrib[k]
|
|
339
|
+
rsid_n += 1
|
|
340
|
+
hit = True
|
|
341
|
+
for db in list(root.iter(w("dataBinding"))):
|
|
342
|
+
db.getparent().remove(db)
|
|
343
|
+
hit = True
|
|
344
|
+
if hit:
|
|
345
|
+
touched.add(part)
|
|
346
|
+
st = [r for r in pkg.rels(pkg.main) if r["type"] == "settings"]
|
|
347
|
+
if st and st[0]["target"] in pkg.names:
|
|
348
|
+
sx = pkg.xml(st[0]["target"])
|
|
349
|
+
for tag in ("rsids", "attachedTemplate", "docVars"):
|
|
350
|
+
for el in sx.findall(f"w:{tag}", NS):
|
|
351
|
+
sx.remove(el)
|
|
352
|
+
stripped.append(tag)
|
|
353
|
+
srels_name = pkg.rels_name(st[0]["target"])
|
|
354
|
+
if srels_name in pkg.names:
|
|
355
|
+
sr = pkg.xml(srels_name)
|
|
356
|
+
for r in list(sr.findall("pr:Relationship", NS)):
|
|
357
|
+
if r.get("Type", "").endswith("/attachedTemplate"):
|
|
358
|
+
sr.remove(r)
|
|
359
|
+
touched.add(srels_name)
|
|
360
|
+
touched.add(st[0]["target"])
|
|
361
|
+
actions["metadata_stripped"] = sorted(set(stripped))
|
|
362
|
+
actions["rsid_attributes_removed"] = rsid_n
|
|
363
|
+
if a.remove_active:
|
|
364
|
+
act = {}
|
|
365
|
+
macro = {n for n in pkg.names if n.lower().endswith(("vbaproject.bin", "vbadata.xml"))}
|
|
366
|
+
ax = {n for n in pkg.names if n.startswith(("word/activeX/", "word/embeddings/"))}
|
|
367
|
+
removed |= macro | ax
|
|
368
|
+
act["macro_parts_removed"] = sorted(macro)
|
|
369
|
+
act["activex_and_ole_parts_removed"] = len(ax)
|
|
370
|
+
objs = chunks = fields_n = 0
|
|
371
|
+
for part in stories + glossary + [n for n in xml_parts if n.endswith("comments.xml")]:
|
|
372
|
+
if part not in pkg.names or part in removed:
|
|
373
|
+
continue
|
|
374
|
+
root = pkg.xml(part)
|
|
375
|
+
hit = False
|
|
376
|
+
for el in list(root.iter(w("object"), w("control"))):
|
|
377
|
+
if el.getparent() is not None:
|
|
378
|
+
el.getparent().remove(el)
|
|
379
|
+
objs += 1
|
|
380
|
+
hit = True
|
|
381
|
+
for el in list(root.iter(w("altChunk"))):
|
|
382
|
+
rid = el.get(f"{R_ATTR}id")
|
|
383
|
+
tgt = next((r["target"] for r in pkg.rels(part) if r["id"] == rid and not r["external"]), None)
|
|
384
|
+
if tgt:
|
|
385
|
+
removed.add(tgt)
|
|
386
|
+
el.getparent().remove(el)
|
|
387
|
+
chunks += 1
|
|
388
|
+
hit = True
|
|
389
|
+
k = neutralize_fields(root)
|
|
390
|
+
if k:
|
|
391
|
+
fields_n += k
|
|
392
|
+
hit = True
|
|
393
|
+
if hit:
|
|
394
|
+
touched.add(part)
|
|
395
|
+
act["objects_and_controls_removed"] = objs
|
|
396
|
+
act["alt_chunks_removed"] = chunks
|
|
397
|
+
act["risky_fields_replaced_by_result"] = fields_n
|
|
398
|
+
# external relationships other than hyperlinks, and the elements that use them
|
|
399
|
+
ext_n = 0
|
|
400
|
+
for rn in [n for n in pkg.names if n.endswith(".rels")]:
|
|
401
|
+
src = rels_source(rn)
|
|
402
|
+
gone = set()
|
|
403
|
+
rx = pkg.xml(rn)
|
|
404
|
+
for r in list(rx.findall("pr:Relationship", NS)):
|
|
405
|
+
if r.get("TargetMode") == "External" and not r.get("Type", "").endswith("/hyperlink"):
|
|
406
|
+
gone.add(r.get("Id"))
|
|
407
|
+
rx.remove(r)
|
|
408
|
+
if not gone:
|
|
409
|
+
continue
|
|
410
|
+
ext_n += len(gone)
|
|
411
|
+
touched.add(rn)
|
|
412
|
+
if src and src in pkg.names and src.endswith(".xml") and src not in removed:
|
|
413
|
+
root = pkg.xml(src)
|
|
414
|
+
for el in list(root.iter()):
|
|
415
|
+
hits = [k for k, v in el.attrib.items() if k.startswith(R_ATTR) and v in gone]
|
|
416
|
+
if not hits or el.getparent() is None:
|
|
417
|
+
continue
|
|
418
|
+
if el.tag.rsplit("}", 1)[-1] in ("blip", "imagedata"):
|
|
419
|
+
for k in hits:
|
|
420
|
+
del el.attrib[k]
|
|
421
|
+
else:
|
|
422
|
+
el.getparent().remove(el)
|
|
423
|
+
touched.add(src)
|
|
424
|
+
for part in stories + glossary:
|
|
425
|
+
if part in pkg.names and part not in removed:
|
|
426
|
+
for el in pkg.xml(part).iter("{urn:schemas-microsoft-com:vml}imagedata"):
|
|
427
|
+
if el.attrib.pop(f"{{{VML_O}}}href", None) is not None:
|
|
428
|
+
ext_n += 1
|
|
429
|
+
touched.add(part)
|
|
430
|
+
act["external_relationships_removed"] = ext_n
|
|
431
|
+
st = [r for r in pkg.rels(pkg.main) if r["type"] == "settings"]
|
|
432
|
+
if st and st[0]["target"] in pkg.names:
|
|
433
|
+
sx = pkg.xml(st[0]["target"])
|
|
434
|
+
for tag in ("mailMerge", "updateFields"):
|
|
435
|
+
for el in sx.findall(f"w:{tag}", NS):
|
|
436
|
+
sx.remove(el)
|
|
437
|
+
act.setdefault("settings_removed", []).append(tag)
|
|
438
|
+
touched.add(st[0]["target"])
|
|
439
|
+
if pkg.kind in PLAIN_TYPE:
|
|
440
|
+
ct = pkg.xml("[Content_Types].xml")
|
|
441
|
+
for o in ct.findall("ct:Override", NS):
|
|
442
|
+
if o.get("PartName", "").lstrip("/") == pkg.main:
|
|
443
|
+
o.set("ContentType", PLAIN_TYPE[pkg.kind])
|
|
444
|
+
touched.add("[Content_Types].xml")
|
|
445
|
+
act["content_type"] = f"{pkg.kind} -> {'docx' if pkg.kind == 'docm' else 'dotx'}"
|
|
446
|
+
actions["active_content"] = act
|
|
447
|
+
if removed:
|
|
448
|
+
# relationships that point at removed parts, from every remaining .rels file
|
|
449
|
+
for rn in [n for n in pkg.names if n.endswith(".rels")]:
|
|
450
|
+
src = rels_source(rn)
|
|
451
|
+
if src in removed or rn in removed:
|
|
452
|
+
removed.add(rn)
|
|
453
|
+
continue
|
|
454
|
+
targets = {r["id"]: r["target"] for r in pkg.rels(src) if not r["external"]}
|
|
455
|
+
rx = pkg.xml(rn)
|
|
456
|
+
for r in list(rx.findall("pr:Relationship", NS)):
|
|
457
|
+
if targets.get(r.get("Id")) in removed:
|
|
458
|
+
rx.remove(r)
|
|
459
|
+
touched.add(rn)
|
|
460
|
+
ct = pkg.xml("[Content_Types].xml")
|
|
461
|
+
for o in list(ct.findall("ct:Override", NS)):
|
|
462
|
+
if o.get("PartName", "").lstrip("/") in removed:
|
|
463
|
+
ct.remove(o)
|
|
464
|
+
touched.add("[Content_Types].xml")
|
|
465
|
+
for part in touched:
|
|
466
|
+
if part not in removed:
|
|
467
|
+
replaced[part] = wp.serialize(pkg.xml(part))
|
|
468
|
+
wp.write_package(pkg.zip, a.output, replaced, removed)
|
|
469
|
+
except (wp.Unsupported, wp.Refusal, OSError) as e:
|
|
470
|
+
done(3 if isinstance(e, wp.Refusal) else 2, ok=False, error=clean(str(e), 400))
|
|
471
|
+
|
|
472
|
+
after = inspect_docx.inspect(a.output)
|
|
473
|
+
remaining = after["residue"]
|
|
474
|
+
wanted = set()
|
|
475
|
+
if a.accept_changes:
|
|
476
|
+
wanted.add("tracked_changes")
|
|
477
|
+
if a.remove_comments:
|
|
478
|
+
wanted.add("comments")
|
|
479
|
+
if a.remove_hidden:
|
|
480
|
+
wanted.add("hidden_text")
|
|
481
|
+
if a.remove_glossary:
|
|
482
|
+
wanted.add("glossary_document")
|
|
483
|
+
if a.strip_metadata:
|
|
484
|
+
wanted |= {"personal_metadata", "custom_properties", "thumbnail", "edit_session_ids",
|
|
485
|
+
"document_variables", "template_or_base_path", "custom_xml"}
|
|
486
|
+
if a.remove_active:
|
|
487
|
+
wanted |= {"macros_or_activex", "embedded_objects", "external_relationships"}
|
|
488
|
+
bad = sorted(set(remaining) & wanted)
|
|
489
|
+
errors = []
|
|
490
|
+
if bad:
|
|
491
|
+
errors.append(f"requested categories still present: {bad}")
|
|
492
|
+
if a.remove_hidden and hidden_field_runs:
|
|
493
|
+
errors.append(f"{hidden_field_runs} hidden runs hold field codes and were kept; fix them in Word")
|
|
494
|
+
if a.remove_active and after["risks"]:
|
|
495
|
+
errors.append("risks remain after --remove-active")
|
|
496
|
+
if a.share and remaining:
|
|
497
|
+
errors.append(f"--share: residue remains: {remaining}")
|
|
498
|
+
res = {"ok": not errors, "output": a.output, "actions": actions, "remaining": remaining,
|
|
499
|
+
"kept_metadata": kept, "risks_after": after["risks"],
|
|
500
|
+
"not_checked": after["clean_scope"]["not_checked"]}
|
|
501
|
+
if errors:
|
|
502
|
+
res["error"] = "; ".join(errors)
|
|
503
|
+
done(1 if errors else 0, **res)
|
|
504
|
+
|
|
505
|
+
|
|
506
|
+
if __name__ == "__main__":
|
|
507
|
+
main()
|