v0.7.0

perf: support batch denoising
perf: improve GPU performance

v0.6.0

v0.6.0 (2024-05-25)

Feature

• feat(scar): support MPS, but not recommend due to lower speed than CPU on Mac (9c3145c)

Unknown

• Develop (#75)

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg

• chore: update workflow file

• refactor codes (#73) (#74)

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg (4d4ea48)

• refactor codes (#73)

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg

• fix: refactor pyproject.toml file, delete setup.cfg (830677b)

v0.5.5

Full Changelog: v0.5.4...v0.5.5

• Update release notes
• Require setuptools>=68.1.2 instead of >=42 to be compatible with pytorch-cuda12

v0.5.4

Skip the previous version v0.5.3 due to dependency conflict

v0.5.3

What's Changed

• Develop by @CaibinSh in #67
• fix: update python=3.8.6 -> python=3.10 to fix numpy & pandas conflict by @CaibinSh in #69

Full Changelog: v0.5.2...v0.5.3

v0.5.2: add an option to denoise scATACseq

What's Changed

• clip the predicted ambient counts by @CaibinSh in #60
• new release by @CaibinSh in #61
• Default clip_to_obs to False by @CaibinSh in #63
• feat: add a feature/tutorial for denoising scATACseq by @CaibinSh in #65

Full Changelog: v0.5.0...v0.5.2

v0.5.1: Default clip_to_obs to False (#63)

Feature

• Update dependencies (40e906e)
• Default clip_to_obs to False, as it leads to overall over-correction (944f09a)
• Support AnnData object as the input (24a1ab6)
• Estimate the ambient profile for all features together, add ambient_profile_all to adata.uns (e1913fd)

Fix

• setup: Change sample and n_batch parameters to allow using all droplets by default (4f5e05f)
• setup: Fix setupdata (63afe9d)

Documentation

• tutorials: Update tutorials (a21e2ec)
• Print message (7659672)
• Print message (203687f)

v0.5.0

Feature

• *: Remove tensorboard parameter (98717e7)
• model: Add an option to clip the prediction (b250ebf)

v0.4.4

Documentation

• Update dependency (03cf19e)
• Update dependencies (9bd7f1c)
• Update documentations (418996c)
• Update dependencies (1bde351)
• main: Add link to anndata and scanpy (8436e05)
• main: Update dependencies (984df35)
• main: Update documentation for .h5 file (2a309e0)
• Add a link of binary installers (2faed3e)
• Update documentations (e26a6e9)
• Add competing methods (8564b2b)
• scar: Add versionadded directives for parameter sparsity and round_to_int (33e35ca)
• Update docs (a4da539)
• Update introduction (a036b24)
• Change readthedocs template (421e52f)
• data_generator: Update docs (1f8f668)
• data_generator: Re-style docs (afef9fb)
• *: Re-style docs (2d550fa)

Performance

• main: Command line tool supports a new input: filtered_feature_bc_matrix.h5 (73bc13e)
• setup: Add an error raise statement (f4fb1a8)

v0.4.3

Fix

• setup: Fix a bug to allow sample reasonable numbers of droplets (ef6f7e4)
• main: Fix a bug in main to set default NN number (794ff17)

Documentation

• main: Add scanpy as dependency (252a492)

Performance

• main: Set a separate batchsize_infer parameter for inference (8727f04)
• setup: Add an option of random sampling droplets to speed up calculation (ce042dd)
• setup: Enable manupulate large-scale emptydroplets (15f1840)