Releases: scverse/mudata
Releases · scverse/mudata
v0.3.1
v0.3.0
New features
- New interface for pulling metadata (
.obs
/.var
) from modalities and pushing it to modalities (#57)- Use
mudata.set_options(pull_on_update=False)
to adopt the new.update()
behaviour - See documentation for examples
- Use
mudata.concat()
analogous toanndata.concat()
(#58)- Experimental
mudata.to_anndata(mdata)
(merging) andmudata.to_mudata(adata, key)
(splitting) (#54)
as well as
.mod_names
attribute to match.obs_names
and.var_names
- Pretty-print
MuData.mod
including axis information (#51) fsspec
support for reading HDF5 files
Improvements
- Improve handling of intersecting
.var_names
across modalities when modalities are added to an existing MuData object - Whenever possible,
.update()
usesbool
dtype — andpd.BooleanArray
when there are null values
Fixes
- Fix nested
MuData
dimensions in__repr__()
- Fix
.uns
contents for views - Fix
read_zarr()
to use modality order - Compatibility with
anndata
v0.11,pandas
v2.0,numpy
v2.0
Development
- Adopt the
src
layout - Use
hatch
for the build process
v0.2.4
Compatibility improvements
- Compatible with the numpy 2.0.0 release — requires anndata 0.10.8 or newer
- Compatible with the dtype argument deprecation in the anndata 0.11
v0.2.3
v0.2.2
v0.2.1
v0.2.0
- Compatibility with the new I/O specification from AnnData v0.8
- New axes interface allows to use MuData object as containers with different shared dimensions
.update()
behaviour and performance improvements
v0.1.2
- Enhanced
zarr
support (#7). - There's a new page in the documentation describing the sharp bits.
- There's a new mod-order attribute.