Refactor: improve enclosed tessellation fallback handling#184
Closed
yu-ta-sato wants to merge 8 commits into
Closed
Refactor: improve enclosed tessellation fallback handling#184yu-ta-sato wants to merge 8 commits into
yu-ta-sato wants to merge 8 commits into
Conversation
The retry ladder in _run_tessellation_with_retries handled terminal
failures asymmetrically: a GEOSException after the coarser-grid retry
degraded the unit to an empty tessellation, but the same exception
raised during the TypeError-initiated retry, a ValueError ("No objects
to concatenate") raised on any retry call, or any known failure with
caller-pinned simplify/grid_size options propagated and aborted the
whole run.
Every known failure (coverage_simplify TypeError, GEOSException, momepy
concat ValueError) now ends in a warning plus an empty tessellation at
whichever rung it occurs; only unknown errors propagate. Caller-pinned
options are still never overridden — a pinned grid_size or simplify now
degrades to empty instead of raising once the remaining options are
exhausted. The overlap-repair retry likewise keeps the original cells
and drops the broken enclosures when its own retry fails, instead of
raising away the good ones.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Empty enclosed-tessellation outputs previously carried a tess_id column on some paths (momepy concat failure) but not others (retry-ladder degradation, empty enclosures, empty input), while non-empty outputs always have both tess_id and enclosure_index. Every empty enclosed result now uses the same [geometry, enclosure_index, tess_id] schema via _create_empty_tessellation, and the include_tess_id plumbing is removed. The empty-input early return in create_tessellation reuses the helper instead of building the frame inline; the no-barrier (morphological) empty schema is unchanged. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Replace the nested try/except ladder in _run_tessellation_with_retries with a single retry loop: _classify_tessellation_failure names the known failure kinds (concat ValueError, coverage_simplify TypeError, GEOSException) and _next_retry_overrides picks the next rung (coarser grid first, then simplify=False) while never overriding caller-pinned options. The rung order and all log substrings are unchanged; the full tessellation test suite passes without any expectation changes. _repair_or_drop_degenerate_enclosures now reuses the same runner (seeded with the coarser grid_size and a repair-specific exhaustion warning) instead of duplicating the TypeError handling, and the interim helpers _run_final_tessellation_attempt, _retry_at_coarser_grid and _is_known_tessellation_error are absorbed. The coarse precision 1e-3 moves into the _COARSE_GRID_SIZE constant. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The whole-tessellation fallback (_fallback_tessellation_without_ enclosures) and the unenclosed-building fallback (_include_unenclosed_ building_tessellation) built their cells with separate inline code, and only the latter recorded the source building index. Whole-tessellation fallback cells therefore relied on the representative-point spatial join in _add_building_info, which duplicates rows and mis-assigns buildings when one footprint contains several representative points — the same failure mode already fixed for the unenclosed path. Both paths now build their cells through a shared _building_fallback_cells helper that always records the source index, so the exact-match logic applies uniformly. The whole-tessellation fallback's empty returns also gain the proper place schema instead of echoing raw building columns, and the internal source-index column no longer leaks into the place nodes when keep_buildings=False. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Pure extraction of the fallback deduplication and exact-match block into _match_fallback_cells_to_source_buildings, keeping the positional masking that tolerates sjoin-duplicated index labels. No behavior change; no test changes. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
With barriers present, create_tessellation degrades the momepy concat failure to an empty tessellation internally, so the ValueError conversion in _create_enclosed_tessellation only fires on the barrier-free morphological path; the enclosed path reaches the footprint fallback through the empty-tessellation check instead. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
yu-ta-sato
changed the title
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.
Description
This PR improves enclosed tessellation robustness around known
momepy/ Shapely failure modes and makes fallback behavior more consistent.coverage_simplifygeometry-type errors, andNo objects to concatenatefailures.geometry,enclosure_index, andtess_id.keep_buildings=True, avoiding duplicate or incorrect spatial-join matches.suppress_empty_errorapplies.Related Issues
N/A
Checklist