Metadata-Version: 2.1
Name: scvae
Version: 2.1.1
Summary: Model single-cell transcript counts using deep learning.
Home-page: http://github.com/scvae/scvae
Author: Christopher Heje Grønbech, Maximillian Fornitz Vording
Author-email: christopher.groenbech@qlucore.com
License: Apache 2.0
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6, <3.8
Description-Content-Type: text/markdown
Requires-Dist: importlib-resources (>=1.0)
Requires-Dist: loompy (>=2.0)
Requires-Dist: numpy (>=1.16)
Requires-Dist: matplotlib (>=2.0)
Requires-Dist: pandas (>=0.24)
Requires-Dist: pillow (>=5.4)
Requires-Dist: scikit-learn (>=0.20)
Requires-Dist: scipy (>=1.2)
Requires-Dist: seaborn (>=0.9)
Requires-Dist: tables (>=3.5)
Requires-Dist: tensorflow (<2,>=1.15.2)
Requires-Dist: tensorflow-probability (==0.7)
Provides-Extra: docs
Requires-Dist: pygments (>=2.4) ; extra == 'docs'
Requires-Dist: sphinx (>=2.2) ; extra == 'docs'

# scVAE: Single-cell variational auto-encoders #

scVAE is a command-line tool for modelling single-cell transcript counts using variational auto-encoders.

Install scVAE using pip for Python 3.6 and 3.7:

	$ python3 -m pip install scvae

scVAE can then be used to train a variational auto-encoder on a data set of single-cell transcript counts:

	$ scvae train transcript_counts.tsv

And the resulting model can be evaluated on the same data set:

	$ scvae evaluate transcript_counts.tsv

For more details, see the [documentation][], which include a user guide and a short tutorial.

[documentation]: https://scvae.readthedocs.io


# Release History

## 2.1.1 (2020-02-24) ##

* Requires TensorFlow 1.15.2 because of a security vulnerability.
* Export of latent values as compressed TSV files.
* Make folder names and filenames more safe on Windows.
* Regrouped analyses, so fewer analyses are performed by default. All available analyses can be performed using ``--included-analyses all``.
* Fix loading of KL divergences when evaluating VAE models.
* Fix crash during model analyses, if the model did not exist.

## 2.1.0 (2019-11-12) ##

* Requires Python 3.6 or 3.7 as well as TensorFlow 1.15.
* Documentation with user guide and tutorial.
* Support for sparse matrices in HDF5 format.
* Improved support for Loom files by following conventions.
* Scatter plots of classes against the primary latent feature as well as the two primary latent features against each other when evaluating a model.
* Fix crash related to `argparse` when using Python 3.6.

## 2.0.0 (2019-05-18) ##

* Complete refactor and clean-up including structuring as Python package.
* Easier loading of custom data sets.
* Batch correction included in models for data sets with batch indices.
* Learnable mixture coefficients for the GMVAE model.
* Full covariance matrix for the GMVAE model.
* Sampling from models.

## 1.0 (2018-04-25) ##

Initial release.


