Skip to content

Walk one DOCX in place: zdanie, akapit, rozdział, całość - #371

Closed
mikolaj92 wants to merge 2 commits into
mainfrom
cursor/docx-stay-or-go-walk-9f83
Closed

mikolaj92 wants to merge 2 commits into
mainfrom
cursor/docx-stay-or-go-walk-9f83

Conversation

@mikolaj92

@mikolaj92 mikolaj92 commented Oct 4, 2026 •

Copy link
Copy Markdown
Owner

A DOCX review walks that same file. It does not review an abstract copy and write a different file at the end.

Order

  1. zdanie by zdanie. On a zdanie: go to the next one, or stay and iterate this same zdanie again. A side effect on that same file may add, update, or delete a comment, or change the text. Later stays can add another comment while earlier ones remain.
  2. After every zdanie is done, the same stay-or-go loop over every akapit.
  3. Then the same loop over every rozdział.
  4. Then całość, walked n times (once, or as many times as stay says).

That walk is the review. README describes it with one mermaid diagram using those four words throughout.

review_docx(path, reviewer) is the entry point. Pack tree name-then-judge (review_tree / TaktReviewer) is unchanged and is not this walk. Temida is not touched.

Test

Start from one DOCX; stay on the first zdanie and add two comments before moving to the second; then an akapit pass, a rozdział pass, and more than one pass over całość. Comments and a text change land on that same file during the walk. An existing comment is updated and then deleted.

The 0.24.1 two-scan nested post-order plant (sequential_scan) still exists for Pack trees; it is not this review.

Checks

pytest (except native Takt/Mojo) 511 passed. ruff and mypy clean.

Open in Web Open in Cursor 

cursoragent and others added 2 commits October 4, 2026 08:23
ReviewKit now reviews a DOCX by staying or going on that same file:
zdanie, then akapit, then rozdział, then całość n times. Comments and
text changes land during the walk, including two comments on one zdanie
and update/delete of existing comments.

Co-authored-by: Patryk Mikołajczyk <mikolaj92@users.noreply.github.com>
The całość text change is a tracked replace on the same file. Assert the
effective post-change paragraph text, not the raw projection that still
shows deleted plus inserted runs.

Co-authored-by: Patryk Mikołajczyk <mikolaj92@users.noreply.github.com>

@mikolaj92 mikolaj92 left a comment

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Alfred (Grok Bot) — review COMMENT

Co robi: Walk jednego DOCX in-place: zdanie→akapit→rozdział→całość ze stay-or-go; entry review_docx(path, reviewer). Pack/review_tree nietknięte.

Zgodność: to jest właściwy tor produktowy (ReviewKit review_docx = jedyna ścieżka review; Temida nie chodzi pliku).

Ryzyka:

  1. Duplikat vs #372 (review_one_docx / inne moduły docx_live/stay_or_go) — dwa równoległe API na to samo. Wybierz jeden kontrakt publiczny zanim undraft.
  2. Mergeable: CONFLICTING.
  3. Docxtor ops: efekty komentarzy/tekstu w docx_effects — upewnij się, że Word ops idą przez Docxtor pin, nie ad-hoc.

Blocker: konflikt + decyzja #371 vs #372. Draft OK; nie merge.

@PSyron

PSyron commented Oct 6, 2026

Copy link
Copy Markdown
Collaborator

The one-DOCX sentence → paragraph → chapter → whole-body walk is already merged through #373/#374. Merged #375 supplies source-hash-bound physical comment regions and complete chapter/whole-body coverage; this older draft uses obsolete scalar/first-paragraph addressing. Closing the superseded alternative without merging its obsolete implementation. Branch preserved.

Disposition is based on independent exact-head source review against main fc1e5340a69e122e51720a36cd8b0764a06e904c; no local tests, application imports or runtime scenarios were run.

@PSyron PSyron closed this Oct 6, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants