fix: thread label-only row-group headers in real docling shapes (0.6.3)#14
Merged
Conversation
Untrack the committed .claude/settings.json and gitignore the whole .claude/ directory so personal Claude Code config (permissions, skills) is never pushed to this public repo. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…le shapes
Three shapes from real TableItem.export_to_html output still dropped the
line-item title from grouped values' row_headers; 0.6.x only handled the
idealized forms.
- Narrow title -> full-width description band -> values: the title's row-group
extent now extends THROUGH an immediately-following full-width description
band (absorbed as a nested header member, not a boundary), so the title
threads as the outer ancestor instead of terminating the extent and being
dropped. The absorbed band is bounded by the same extent so it does not leak
past the next title.
- Multi-cell title rows (leading item number/key + textual title, "10 | Travel
delay" / "3. | Permanent loss of:"): a label-only row is a group header when
at most ONE of its label cells is numeric-only. This admits number+title while
still rejecting a data row whose value columns merely happen to be empty
("Average: | 80.2 | 10.7 | 3.3", >=2 numeric) -- replaces the old exactly-one-
cell guard. A repeating key column (same col+text on the group's value rows)
is excluded from the promoted title so it is not duplicated in the path.
- Two-column Label|Value schedules: Signal D promotes the left column to the
row-label/stub even under a single-row thead header, scoped to max_cols == 2
so multi-column property tables are untouched. Also yields proper
one-record-per-line output for plain 2-col relational tables
("North | Sales: 100") instead of two disconnected "Header: value" lines.
4 existing golds improve (2-col relational tables now bind row<->value); no
value or label text dropped. New fixtures label-only-title-then-description-band
and label-only-title-number-key-matrix. Full suite green; ruff + mypy clean.
Known limitation (pre-existing, separate gate interaction, tracked for
follow-up): a "<n>. | Group: | (empty)" sub-grouped header with a colspan title
over a promoted descriptor column still falls back to flat.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.8 (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.
Release 0.6.3 — fixes the comprehensive label-only-row-group bug report (real docling
export_to_htmlshapes).Bug
Grouped values were dropping their line-item title from
row_headers, collapsing every line-item's limits under the section header. Three real shapes failed:10 | Travel delay) was rejected by the exactly-one-cell guard → "Travel delay" dropped.Label | Valueschedules didn't promote the stub column under a single-row thead.Fixes
Average: | 80.2 | 10.7 | 3.3). Repeating key columns are excluded from the promoted title.Label | Valuetable to the row-label/stub even under a single-row thead (scoped tomax_cols == 2).Verification
matrix/label-only-title-then-description-band,matrix/label-only-title-number-key-matrix.Known limitation (pre-existing, separate — tracked for follow-up)
A
<n>. | Group: | (empty)sub-grouped header with acolspantitle over a promoted descriptor column still falls back to flat (the spanned cell trips the gate's "rules originate from<td>" invariant). This is a gate interaction present before this change, not the label-only-threading path.Also includes a small hygiene commit: stop tracking
.claude/local config in this public repo.🤖 Generated with Claude Code