Bioconductor 3.20

Major changes

• Add environment for anndata v0.10.9. This is now the default environment for the Python reader/writer

Minor changes

• Avoid deprecation warning due to setting dtype when creating Python AnnData objects
• Standardise code styling using {styler}
• Add test using the more complete example H5AD file

Bug fixes

• Correctly assign levels to factors in R reader with anndata v0.7 files (Fixes #122)
• Correctly set filepath in the R reader with reading adata.raw with use_hdf5 = TRUE (PR #124 @GabrielHoffman, Fixes #123)

Full Changelog: v1.14.0...v1.16.0

Bioconductor 3.19

Major changes}{

• Add environment for anndata v0.10.6. This is now the default environment for the Python reader/writer.

Minor changes

• Improve warnings when converting matrices fails
• Minor change to writing DelayedArray matrices for compatibility with {HDF5Array} >= v1.31.1

Bug fixes

• Correctly handle use_backed = TRUE with newer anndata versions (Fixes #114)
• Correctly instantiate the anndata v0.10.2 environment (Fixes #103)
• Minor fixes for typos etc. (Fixes #100)

Full Changelog: v1.12.0...v1.14.0

Bioconductor 3.18

Major changes}

• Add environments for anndata v0.9.2 and v0.10.2. Version 0.10.20 is now the default environment for the Python reader/writer.

Minor changes}

• Changes for compatibility with {rhdf5} v2.45.1 including enum types that simplifies reading of nullable types in the native R reader
• Dimensions are now passed correctly when converting the raw slot (Fixes #96)
• Backed sparse matrices are now converted AnnData2SCE() (Fixes #96)

Full Changelog: v1.10.0...v1.12.0

Bioconductor 3.17

Major changes

• Add compatibility with the anndata v0.8 H5AD format to the the native R writer (By @jackkamm PR #86, fixes #78 and @mtmorgan, PR #93)
• Add functions for converting pandas arrays used by anndata when arrays have missing values (@lazappi, fixes #87)

Minor changes

• Add @rcannood (PR #90 @rcannood, fixes #88) and @jackkamm as contributors!
• Minor adjustments to tests to match reader changes (@lazappi)

Full Changelog: v1.8.0...v1.10.0

Bioconductor 3.16

Major changes

• Improve compatibility with the R {anndata} package. This required modifying conversion functions so that Python objects are explicitly converted rather than relying on automatic conversion. (PR #76 @rcannood, fixes #75)
• Added support for numpy recarrays. This solves a long-standing issue and allows results from scanpy's rank_genes_groups() function to be read. (PR #81, fixes #45, #28)

Minor changes

• The Python version is now pinned in the anndata v0.7.6 environment for compatibility with changes in {basilisk}
• Instantiate Python environments so they can be properly picked up by basilisk::configureBasiliskEnv() (Fixes #66)
• Allow missing obs/var names when use_hdf5 = TRUE (Fixes #65)
• Minor changes to the UI functions for compatibility with {cli} v3.4.0
• Minor changes for compatibility with {Matrix} v1.4-2
• Improvements to the UI for warnings
• Updates and improvements to tests

Full Changelog: v1.7.0...v1.8.0

Bioconductor 3.15

Major changes

• Added support for multiple {basilisk} environments with different anndata versions. Users can now specify the environment to use with options in readH5AD() and writeH5AD(). To faciliate this some exported objects were converted to functions but this should only affect developers.
• Updated the default environment to use anndata v0.8.0. This is a major update and files written with v0.8.0 cannot be read by previous anndata versions. This was the motivation for supporting multiple environments and users can select the previous environment with anndata v0.7.6 if compatibility is required.
• Standardise naming in AnnData2SCE(). Column names of data frames and names of list items will now be modified to match R conventions(according to make.names()). When this happens a warning will be issued listing the modifications. This makes sure than everything in the SingleCellExperiment is accessible.

Minor changes

• Allow data.frame's stored in varm to be converted in SCE2AnnData()
• Minor updates to the vignette and other documentation.
• Updates to tests to match the changes above.

Full Changelog: v1.4.0...v1.6.0

Bioconductor 3.14

• Add arguments to control how slots are converted in AnnData2SCE() and SCE2AnnData(). Each slot can now be fully converted, skipped entirely or only selected items converted. (#47)
• Add support for converting the raw slot to an altExp in AnnData2SCE() (#53, #57)
• Add recursive conversion of lists in AnnData2SCE()
• Add progress messages to various functions. These can be controlled by function arguments or a global variable.
• Add long tests for various public datasets. This should help to make the package more robust.
• Fix bug in converting dgRMatrix sparse matrices (#55)
• Correctly handle DataFrame objects stored in adata.obsm

Full Changelog: v1.2.0...v1.4.0

Bioconductor 3.13

• Update anndata and other Python dependencies, now using anndata v0.7.6
• Improved conversion checks for all slots in AnnData2SCE() (See #45)
• Enable return conversion of the varm slot in AnnData2SCE() (Fixes #43)
• Avoid converting obsp and varp to dense matrices in AnnData2SCE()
• AnnData2SCE() should now always return dgCMatrixmatrices when assays are sparse (Fixes #34)
• More consistent conversion of metadata to uns in SCE2AnnData() (Fixes #40)
• Handle conversion of list columns in colData and rowData in SCE2AnnData() (Fixes #26)
• Better support for converting anndata SparseDataset arrays (PR #41, Fixes #37, Fixes #42)
• Improved support for conversion of HDF5 backed AnnData objects
• Better support for writing DelayedArray assays in writeH5AD() (PR #35, Fixes #32)
• Store X_name in AnnData2SCE() for use by SCE2AnnData() and add an X_name argument to AnnData2SCE() and readH5AD() (Fixes #7)
• Add a compression argument to writeH5AD() (Fixes #49)
• Add an experimental native R reader to readH5AD()
• Export zellkonverterAnnDataEnv for use by other packages (Fixes #38)

Bioconductor 3.12

First Bioconductor release of {zellkonverter}

{zellkonverter} provides methods to convert between Python AnnData objects and SingleCellExperiment objects. These are primarily intended for use by downstream Bioconductor packages that wrap Python methods for single-cell data analysis. It also includes functions to read and write H5AD files used for saving AnnData objects to disk.