TileDB-SOMA 0.5.0a2

This is for development on the main branch, in contrast to 0.1.x releases which track development on the main-old branch.

Ongoing development

• NdArray refinement - tranche 4 by @bkmartinjr in #367
• [python] Conform to spec by reading as pyarrow.Table not pyarrow.RecordBatch by @johnkerl in #355
• Correctly handle empty query results in SparseNdArray read operations by @bkmartinjr in #371
• Initial SOMADataFrame class by @aaronwolen in #377
• Fix for SOMAIndexedDataFrames with float dims by @aaronwolen in #379
• Initial SOMAIndexedDataFrame class by @aaronwolen in #378
• Prevent word splitting path in bld by @aaronwolen in #387
• [r] Enhance support for (partial) table extraction by @eddelbuettel in #389
• Python collections refinement by @bkmartinjr in #381
• [python] Use true ASCII attributes in dataframes by @johnkerl in #359
• [r] Support Query Condition objects, small other polishes by @eddelbuettel in #398
• Refine soma_ column handling in Python API by @bkmartinjr in #397
• [r] Add basic unit testing for SOMAReader interface by @eddelbuettel in #401
• [c++] Add fast nnz by @gspowley in #404
• fix index in pandas read all by @bkmartinjr in #408
• [python] Disambiguate bytes/ascii on tiledb schema -> arrow schema by @johnkerl in #406
• rename type to soma_type by @bkmartinjr in #409
• [python] Match SparseNdArray dim/attr names to spec by @johnkerl in #411
• [c++] Expose two functions from managed_query in soma_reader by @eddelbuettel in #412
• add unit tests for issues 415, 418, 419 by @bkmartinjr in #420
• [r] Provide nnz() accessor by @eddelbuettel in #422
• [c++,r] Simplify spdlog setup for use from R by @eddelbuettel in #421
• Refactor from_anndata support by @bkmartinjr in #416
• Use C++17 to ease build requirements by @eddelbuettel in #417
• [python] Disallow duplicates [main branch] by @johnkerl in #437
• [r] Add support for row selection by @eddelbuettel in #402
• [r] Add Continuous Integration for the R package by @eddelbuettel in #403
• [python/c++] Connect C++ nnz to Python by @johnkerl in #439
• [c++] Revert #439 by @johnkerl in #443
• [c++] Internal context-handling updates by @johnkerl in #444
• [python/ci] Run Python CI on 3.8 and 3.9 by @johnkerl in #445
• Correct configure by ensuring R_HOME is set early by @eddelbuettel in #449
• [python] Depend on TileDB Core 2.12.0 by @johnkerl in #448
• [c++] Do not require query condition columns in column_names by @gspowley in #393
• [python/c++] Connect C++ nnz to Python by @johnkerl in #446
• [c++] Add column name to the arrow schema by @gspowley in #452
• [c++] Support passing column_names as None by @gspowley in #451
• [r] Enhanced soma reader by @eddelbuettel in #454
• [c++] Add obs slice X test by @gspowley in #456
• [python/c++] Connect C++ reader for dataframes by @johnkerl in #400
• [python] Correctly handle zero-length dataframes with SOMAReader by @johnkerl in #459
• [python] More unit-test cases for #459 by @johnkerl in #460
• [c++] Expose is_complete predicate from managed_query in soma reader by @eddelbuettel in #461
• [python/c++] Connect C++ reader for SOMASparseNdArray by @johnkerl in #447
• [r] Initial support for iterated reads from SOMAReader by @eddelbuettel in #462
• [python] Unprefix SOMA names by @johnkerl in #464
• [r] Minor rearrangement demoting initial test function and renaming by @eddelbuettel in #468
• [python] Fix #471 by @johnkerl in #472
• [python] Update to TileDB-Py 0.18.0 [main] by @johnkerl in #476
• [python] Replace Exception with SOMA-specific exceptions by @johnkerl in #486
• [r] Expand iterated reader via refactored row/col selectors by @eddelbuettel in #480
• [r] Initial SOMADenseNDArray class by @aaronwolen in #488
• [c++] Accept dim_points as pyarrow.Array by @gspowley in #498
• [r] Add coverage reports in script and yaml by @eddelbuettel in #504
• [python] Depend on TileDB-Py 0.18.1 [main] by @johnkerl in #503
• [r] Arrow conversion utilities by @aaronwolen in #505
• [r] Use libtiledbsoma.a in build by @eddelbuettel in #510
• [c++] Fix macos build error by @gspowley in #513
• [c++] Enable internal TileDB stats collection by @gspowley in #515
• [python] Update slicing logic for SparseNdArray by @johnkerl in #485
• [r] Enable code coverage token use by @eddelbuettel in #521
• [c++] Improve build experience by @gspowley in #522
• [python] Update SOMAIndexedDataFrame indexing logic by @johnkerl in #483
• add temporary configuration to SOMAReader instances by @bkmartinjr in #527
• improving IndexedDataFrame test coverage by @mlin in #529
• [c++] Cmake bug fix and improvements by @gspowley in #525
• [python] Remove DataFrame in favor of IndexedDataFrame by @johnkerl in #538
• [r] Support logging via spdlog by @eddelbuettel in #530
• [python] Remove DataFrame::is_indexed by @johnkerl in #545
• [r] Remove DataFrame in favor of IndexedDataFrame by @johnkerl in #537
• [c++] Release Python GIL on reads by @gspowley in #550
• [r] Ignore directory with TileDB Core headers in coverage analysis by @eddelbuettel in #552
• [r] Update logging code utilizing templated interface by @eddelbuettel in #555
• [python] unit tests and error checking for #300 by @bkmartinjr in #559
• Add the platform_config parameter to create methods. by @thetorpedodog in #542
• [r] Extend iterated reader interface with platform_config by @eddelbuettel in #556
• DataFrame read result_order by @mlin in #528
• [c++] Map 8-bit TILEDB_BOOL to 1-bit Arrow boolean by @bekadavis9 in #561
• [r] Simplify logging via 'spdl' package by @eddelbuettel in #563
• Add documentation for platform_config to the Python README. by @thetorpedodog in #568
• [r] Make source build more robust by @eddelbuettel in #572
• [r] Tighten package by removing some initial exploratory code by @eddelbuettel in #577
• Fix #578 (tile-order/cell-order typo) by @johnkerl in #580
• [python] Support half-slices in indexed reads by @bekadavis9 in #582
• Fix DenseNdArray slicing by @johnkerl in #581
• Use C++ reader for SOMA.DenseNdArray by @johnkerl in #574
• Fix flaky unit test by @johnkerl in #585
• Update to TileDB 2.13 [main] by @johnkerl in #587
• Support value-filter on boolean columns [pending core 2.13 / tiledb-py 0.19] by @johnkerl in #579
• Un-xfail half-slice tests in DenseNdArray by @johnkerl in #589
• Extend indexing-related unit-test cases for DenseNdArray by @johnkerl in #588
• Use soma_reader in SOMADataFrame by @eddelbuettel in #590
• [r] Convert arrow::large_utf8() to ascii by @aaronwolen in #591
• Support string dims in soma.DataFrame by @johnkerl in #592
• Include Python 3.10 in CI matrix by @johnkerl in #597
• Use soma_reader in SOMADenseNdArray read method by @eddelbuettel in #598
• [python] Add tests for point indexing with ndarray by @bkmartinjr in #604
• [r] Tweak logging default value enabling easier use by @eddelbuettel in #607
• [python] Rename NdArray -> NDArray by @johnkerl in #608
• Support reading data that uses old NdArray name by @bkmartinjr in #610
• [r] Making package R CMD check clean by @eddelbuettel in #614

Continuous integration and build infrastructure

• [python] Pause PyPI publish on main by @johnkerl in #369
• [ci] Work around recent Mac GHA issue by @johnkerl in #370
• [ci] Fix scripts/test for MacOS by @johnkerl in #358
• [python] Depend on TileDB-Py 0.17.5 by @johnkerl in #394
• [r] Support build against external artifact and system library by @eddelbuettel in #388
• [ci] Use MacOS 12 for CI by @johnkerl in #413
• [ci] Update to checkout/actions@v3 to suppress nag about node.js 12 by @eddelbuettel in #455
• Build flow improvements by @gspowley in #496
• report pytest coverage to codecov by @mlin in #495
• Build python sdist & wheels by @mlin in #605
• py packaging: strike 'v' prefix from expected version numbers by @mlin in #613

New Contributors

• @mlin made their first contribution in #495
• @bekadavis9 made their first contribution in #561

Full Changelog: 0.5.0a1...0.5.0a2