perf(rust/sedona-raster-gdal): RS_Clip follow-ups from review — zero-copy band output, rotated-raster tests, tester CRS engine - #1055
Merged
Conversation
james-willis
force-pushed
the
jw/rs-clip-followups
branch
2 times, most recently
from
July 17, 2026 06:07
385600c to
fe1610d
Compare
… and appends it zero-copy The mask/crop helpers write their plane's bytes into the caller-owned per-band buffer instead of returning a fresh Vec per plane, and the finished band bytes move into the output BinaryViewArray as a shared view block (append_band_data_buffer) instead of being copied through the builder — one allocation per band, no redundant copies, pinned by a data_buffers() assertion so a refactor can't quietly reintroduce the copy. Criterion (local, --save-baseline): Clip(1024x1024, Polygon(large)) 704.8µs -> 635.9µs (-9.8%); the micro cases are bounded by per-call GDAL setup (~17µs) and move less than 0.5µs.
…sterio reference The crop window maps the geometry envelope's corners through the inverted affine and the output origin shifts by the full affine (both skew terms), so skewed and rotated rasters clip correctly — but every existing clip fixture was north-up. Two parity cases (sheared, and ~30-degree rotated, crop on and off) pin it against rasterio, whose geometry_window agrees with the envelope math on rotated grids.
…jecting UDF tests The tester's ConfigOptions already carry SedonaOptions, but the default CRS engine errors on use, so a tester-invoked UDF could never exercise a reprojection path (only config-less direct calls reach the global-PROJ fallback). with_crs_engine swaps in a real engine; rs_clip demonstrates the pattern with sedona-proj's LazyProjEngine as a dev-dependency, pinning both halves: the default engine errors, and the reprojected clip matches the same-CRS clip through the tester.
james-willis
force-pushed
the
jw/rs-clip-followups
branch
from
July 17, 2026 06:14
fe1610d to
9670b6e
Compare
…ec helpers The result-side navigation in the pre-existing clip tests — ScalarValue destructures into RasterStructArray band-byte chains and raw column_by_name/downcast walks — becomes assert_raster_scalar_equals / assert_rasters_equal against declarative RasterSpec expectations. Besides being readable, the specs assert strictly more: the crs-mismatch test now pins the exact clip output instead of only 4326==3857 equality, the band-column regression pins both rows' contents instead of mere inequality, and band-0 pins both bands' values rather than a band count.
james-willis
marked this pull request as ready for review
July 17, 2026 06:58
paleolimbot
approved these changes
Jul 17, 2026
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
2 participants
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.
clean up comments from dewey in #1000
also drive by clean up some tests to use the raster test harness methods