v0.7.2

API changes

• read_alignment function introduced in v0.7.0 has been pluralized to read_alignments

Maintenance

• Skip read_alignments tests on big-endian architectures by @nvictus in #216

Full Changelog: v0.7.1...v0.7.2

v0.7.1

Maintenance

• Refactor join arrayops and intidx internals by @nvictus in #204
• NumPy 2.0 was released. Pin numpy < 2 until we migrate.

Full Changelog: v0.7.0...v0.7.1

v0.7.0

API changes

• Add to_bed function to validate and write standard BED files @gamazeps in #203
• read_bam deprecated in favor of read_alignments @gamazeps in #206

Documentation

• Add "bioframe for bedtools users" guide to docs by @gamazeps in #198

Bug fixes

• Fix contig name and JSON issues in read_bam implementation by @gamazeps in #206

New Contributors

• @gamazeps made their first contribution in #203

Full Changelog: v0.6.4...v0.7.0

v0.6.4

Maintenance

• Migrate from setuptools pkg_resources to importlib.resources by @nvictus in #194
• Use importlib.metadata for versioning by @nvictus in #195

Bug fixes

• Overlap point segment patch #183 by @smitkadvani in #184
• #167: Replaced np.int with int as the attribute is deprecated by numpy by @harshit148 in #192

New Contributors

• @harshit148 made a first contribution in #192

Full Changelog: v0.6.3...v0.6.4

v0.6.3

Bug fixes

• Prevent dropout from closest in some cases of left intervals with no neighbors by @agalitsyna in #185
• Fix overlap returning float indexes causing failing tests (numpy v1.22.4, pandas v1.5.2) by @agalitsyna in #185

Full Changelog: v0.6.2...v0.6.3

v0.6.2

What's Changed

• cols and df_view_col passed to downstream functions by @smitkadvani in #182

Fixes

• Update to new UCSC hgdownload url by @golobor and @nvictus in #187

Full Changelog: v0.6.1...v0.6.2

v0.6.1

Default behavior of ensure_nullable option in overlap was modified to minimize the possibility of regressions in libraries that depend on legacy behavior.

• The new option was renamed ensure_int and is True by default. It ensures that output coordinate columns are always returned with an integer dtype, as was the case in prior versions. This is achieved by converting columns having non-nullable NumPy dtypes to Pandas nullable ones in the specific case where the result of an outer join generates missing values; otherwise, column dtypes are preserved unchanged in the output.
• Unlike previous minor versions of bioframe, the nullable dtype chosen will have the same underlying type as the corresponding column from the input (i.e, an input dataframe using np.uint32 start coordinates may yield a pd.UInt32 start column in the output).
• This behavior can be turned off by setting ensure_int to False, in which case outer joins on dataframes using NumPy dtypes may produce floating point output columns when missing values are introduced (stored as NaN), following the native casting behavior of such columns.

v0.6.0

API changes:

• overlap: In previous versions, output coordinate columns were always converted to Pandas "nullable" Int64DType before returning. In the interest of flexibility, memory efficiency, and least surprise, the coordinate columns returned in the output dataframe now preserve dtype from the input dataframes, following native type casting rules if missing data are introduced. We introduce the ensure_nullable argument to force Pandas nullable dtypes in the output coordinates. See the docs for more details. (#178)

Bug fixes:

• Fixed coverage with custom cols1 (#170)

Documentation:

• Added contributing guidelines and NumFOCUS affiliation.
• Updated README and added CITATION.cff file.
• Updated performance benchmarks.

v0.5.1

Bug fixes

• Series are treated like dict in make_chromarms

v0.5.0

API changes:

• New builtin curated genome assembly database (metadata, chromsizes, cytobands):
  • bioframe.list_assemblies()
  • bioframe.assembly_info()
• New UCSC RGB color converter utility #158
• Options added to pair_by_distance

Bug fixes:

• Make expand throw an error if both pad and scale are passed (#148)
• Fixes to bioframe.select query interval semantics (#147)

Maintenance:

• Migrate to hatch build system and pyproject.toml
• Various refactorings