fix(documents): strip a page-header breadcrumb laid out as a one-row table - #1364
Merged
Merged
Conversation
…table 5.59.0 regressed ExxonMobil's 10-K Item 7 (0000034088-26-000045): it opened on "Table of Contents Financial Table of Contents" again, failing test_issue_rv86_incorporated_mda on the release PR's regression job. XOM's page header is a one-row table holding both labels. _strip_leading_nav matched one navigation label per line, which held while TOC section text walked table cells onto separate lines. Since tables render cell by cell (bf0441b, edgartools-wzgu) the row arrives as one line with its cells separated by two or more spaces, and matched nothing. A line now counts as a breadcrumb when every cell is a navigation label or a page number and at least one is a label, so a row of bare numbers ("2024 2023") and a label sharing a row with real text are still kept. The same regression job also failed test_issue_826 on AAPL's 10-K Item 8 length pin (60,874 -> 62,223). That one is benign: the letters and digits are identical before and after bf0441b and all 34 tables are present once; the difference is cell separation, so the pin is re-captured. Both network tests were missed because the wzgu PR ran only the fast suite. Verified here: the 56 section-related network regression files, 535 passed, 0 failed; -m fast 7726 passed. New offline test: 7 cases, 2 fail on main. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
dgunning
enabled auto-merge (squash)
September 26, 2026 11:25
Merged
dgunning
added a commit
that referenced
this pull request
Sep 26, 2026
Two fixes merged since 5.59.0 earlier today. ExxonMobil's 10-K Item 7 opened on a "Table of Contents" breadcrumb again, a 5.59.0 regression from rendering section tables cell by cell (#1364). And edgar_read's 13F holdings section, which returned None for every filing, now lists the top holdings through the row reader shared with edgar_ownership (#1365, GH #1337, thanks @wolfgang-aura). 2 fragments folded into the dated section. test-fast green: 7729 passed, 12 skipped, 1 xfailed. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What was wrong
5.59.0 regressed ExxonMobil's 10-K Item 7 (
0000034088-26-000045): it opened onTable of Contents Financial Table of Contentsagain. The release PR's regression job failedtest_issue_rv86_incorporated_mdaon it.XOM's page header is a one-row table holding both labels.
_strip_leading_navmatched one navigation label per line, which held while TOC section text walked table cells onto separate lines. Since #1361 renders tables cell by cell, the row arrives as one line with its cells separated by two or more spaces, and it matched nothing.Fix
A line counts as a breadcrumb when every cell is a navigation label or a page number, and at least one cell is a label. A row of bare numbers (
2024 2023) and a label sharing a row with real text are still kept.Also in this PR
The same regression job failed
test_issue_826on AAPL's 10-K Item 8 length pin, 60,874 → 62,223. That change is benign: the letters and digits are identical before and after #1361, and all 34 tables are present exactly once. The difference is cell separation, so the pin is re-captured.Why it was missed, and verification
#1361 ran only the fast suite; both failures are network tests. This PR was checked against:
test_wzgu_breadcrumb_table_row.py: 7 cases, 2 of which fail on main.-m fast -n 4: 7,726 passed, 0 failed.ruff: no new findings. The 4 reported are pre-existing on main, on lines this PR doesn't touch.Bead: edgartools-wzgu (follow-up)
🤖 Generated with Claude Code