update write_loom to write layers (#53)

anndata/readwrite/write.py

@@ -71,15 +71,15 @@ def write_loom(filename: PathLike, adata: AnnData):
     row_attrs['var_names'] = adata.var_names.values
     col_attrs = adata.obs.to_dict('list')
     col_attrs['obs_names'] = adata.obs_names.values
-    X = adata.X.T
-    if issparse(X):
-        logg.info(
-            '... writing to \'.loom\' file densifies sparse matrix')
-        X = X.tocoo()
+
+    layers = {'': adata.X.T}
+    for key in adata.layers.keys():
+        layers[key] = adata.layers[key].T
+
     from loompy import create
     if filename.exists():
         filename.unlink()
-    create(fspath(filename), X, row_attrs=row_attrs, col_attrs=col_attrs)
+    create(fspath(filename), layers, row_attrs=row_attrs, col_attrs=col_attrs)
 
 # TODO: store can be a MutableMapping or a string - accept PathLike too?
 def write_zarr(store, adata: AnnData, **kwargs):

anndata/tests/layers.py

@@ -1,3 +1,5 @@
+from importlib.util import find_spec
+import pytest
 import numpy as np
 import anndata as ad
 import os
@@ -38,14 +40,24 @@ def test_views():
 def test_readwrite():
     adata = ad.AnnData(X=X, layers={'L':L.copy()})
     adata.write('test.h5ad')
-
     adata_read = ad.read_h5ad('test.h5ad')
 
     assert adata.layers.keys() == adata_read.layers.keys()
     assert (adata.layers['L'] == adata_read.layers['L']).all()
 
     os.remove('test.h5ad')
 
+@pytest.mark.skipif(find_spec('loompy') is None, reason="loompy not installed")
+def test_readwrite_loom():
+    adata = ad.AnnData(X=X, layers={'L': L.copy()})
+    adata.write_loom('test.loom')
+    adata_read = ad.read_loom('test.loom', X_name='')
+
+    assert adata.layers.keys() == adata_read.layers.keys()
+    assert (adata.layers['L'] == adata_read.layers['L']).all()
+
+    os.remove('test.loom')
+
 def test_backed():
     #backed mode for layers isn't implemented, layers stay in memory
     pass

requires_tests.txt

@@ -1,4 +1,6 @@
-loompy>=2.0; python_version>='3.6'
+loompy>=2.0.12; python_version>='3.6'
 pytest-cov
 codecov
 zarr
+matplotlib
+sklearn