Merge pull request #834 from YosefLab/staging

Staging for official release

README.rst

@@ -39,7 +39,7 @@ If you find a model useful for your research, please consider citing the corresp
 Package transition
 ------------------
 
-``scvi`` is transitioning to ``scvi-tools``. If you're looking for ``scvi`` source code, please use the branch tags. ``scvi`` is still available via pypi and bioconda, but there will be no future releases past ``0.6.8``. A beta release of ``scvi-tools`` is currently available. We expect an official release by October 14th.
+``scvi`` is now ``scvi-tools``. If you're looking for ``scvi`` source code, please use the branch tags. ``scvi`` is still available via pypi and bioconda, but there will be no future releases past ``0.6.8``. Documentation for ``scvi`` is available at the same documentation website.
 
 .. _documentation: https://scvi-tools.org/
 .. _`contributing guide`: https://scvi.readthedocs.io/en/stable/contributing.html

docs/authors.rst

@@ -44,4 +44,4 @@ Many have diverse contributions to the code. Here we highlight some specific pro
 .. _Valentine Svensson: http://www.nxn.se/
 .. _Achille Nazaret: https://nazaret.me/
 .. _Galen Xing: https://galenxing.com/
-.. _Yosef Lab: https://niryosef.wordpress.com/
+.. _Yosef Lab: https://yoseflab.github.io/

docs/history.rst

@@ -2,18 +2,18 @@
 Release notes
 =============
 
-0.7.0 (2020-X-XX)
+0.7.0 (2020-10-14)
 -----------------
 scvi is now scvi-tools. Version 0.7 introduces many breaking changes. The best way to learn how to use scvi-tools is with our documentation and tutorials.
 
 * New high-level API and data loading, please see tutorials and examples for usage.
-* ``GeneExpressionDataset`` and associated classes have been removed.
-* Built-in dataests now return ``AnnData`` objects.
-* ``scvi`` now relies entirely on the AnnData_ format.
-* ``scvi.models`` has been moved to ``scvi.core.modules``.
+* `GeneExpressionDataset` and associated classes have been removed.
+* Built-in datasets now return `AnnData` objects.
+* `scvi-tools` now relies entirely on the AnnData_ format.
+* `scvi.models` has been moved to `scvi.core.modules`.
 * `Posterior` classes have been reduced to wrappers on `DataLoaders`
-* ``scvi.inference`` has been split to ``scvi.core.data_loaders`` for `ScviDataLoader` classes and ``scvi.core.trainers`` for trainer classes.
-* Usage of classes like ``Trainer`` and ``ScviDataLoader`` now require the ``AnnData`` data object as input.
+* `scvi.inference` has been split to `scvi.core.data_loaders` for `ScviDataLoader` classes and `scvi.core.trainers` for trainer classes.
+* Usage of classes like `Trainer` and `ScviDataLoader` now require the `AnnData` data object as input.
 
 
 0.6.7 (2020-8-05)

docs/index.rst

@@ -17,7 +17,7 @@ These models are able to simultaneously perform many downstream tasks such as le
 * If you find a model useful for your research, please consider citing the corresponding publication.
 
 .. important:: ``scvi`` is now ``scvi-tools``.
-   If you'd like to view documentation for ``scvi``, please change the documentation version using the menu at the bottom right.
+   If you'd like to view documentation for ``scvi``, please change the documentation version using the menu at the bottom right (versions <= 0.6.8).
 
 .. raw:: html
 

docs/user_guide/index.rst

@@ -124,7 +124,7 @@ This is a walkthrough of a totalVI-based analysis pipeline, from dimension reduc
 
 Here we describe how to use scVI and scANVI for integrating data from Tabula Muris.
 
-.. image:: ../_static/tutorials/scanvi.svg
+.. image:: ../_static/tutorials/scanvi.png
    :align: center
    :height: 300px
 

pyproject.toml

@@ -20,7 +20,7 @@ packages = [
   {include = "scvi"},
 ]
 readme = "README.rst"
-version = "0.7.0-beta.0"
+version = "0.7.0"
 
 [tool.poetry.dependencies]
 anndata = ">=0.7.4"

scvi/data/_anndata.py

@@ -296,9 +296,9 @@ def _verify_and_correct_data_format(adata, data_registry):
     for k in keys:
         data = get_from_registry(adata, k)
         if isspmatrix(data) and (data.getformat() != "csr"):
-            logger.debug("{} is csc_matrix. Overwriting to csr_matrix.".format(k))
-            data = data.tocsr()
-            _set_data_in_registry(adata, data, k)
+            logger.warning(
+                "Training will be faster when sparse matrix is formatted as CSR. It is safe to cast before model initialization."
+            )
         elif isinstance(data, np.ndarray) and (data.flags["C_CONTIGUOUS"] is False):
             logger.debug(
                 "{} is not C_CONTIGUOUS. Overwriting to C_CONTIGUOUS.".format(k)

tests/dataset/test_anndata.py

@@ -18,7 +18,6 @@
 )
 from scvi import _CONSTANTS
 from scvi.data._anndata import get_from_registry
-from scipy.sparse import csc_matrix
 
 
 def test_transfer_anndata_setup():
@@ -124,28 +123,6 @@ def test_data_format():
         get_from_registry(adata, _CONSTANTS.PROTEIN_EXP_KEY),
     )
 
-    # if data is sparse, check that after setup_anndata, data is csr_matrix
-    adata = synthetic_iid(run_setup_anndata=False)
-    adata.X = csc_matrix(adata.X)
-    adata.obsm["protein_expression"] = csc_matrix(adata.obsm["protein_expression"])
-    old_x = adata.X
-    old_pro = adata.obsm["protein_expression"]
-    setup_anndata(adata, protein_expression_obsm_key="protein_expression")
-
-    assert adata.X.getformat() == "csr"
-    assert adata.obsm["protein_expression"].getformat() == "csr"
-
-    assert np.array_equal(old_x.toarray(), adata.X.toarray())
-    assert np.array_equal(old_pro.toarray(), adata.obsm["protein_expression"].toarray())
-
-    assert np.array_equal(
-        adata.X.toarray(), get_from_registry(adata, _CONSTANTS.X_KEY).toarray()
-    )
-    assert np.array_equal(
-        adata.obsm["protein_expression"].toarray(),
-        get_from_registry(adata, _CONSTANTS.PROTEIN_EXP_KEY).toarray(),
-    )
-
     # if obsm is dataframe, make it C_CONTIGUOUS if it isnt
     adata = synthetic_iid()
     pe = np.asfortranarray(adata.obsm["protein_expression"])