Merge pull request #98 from cjnolet/rapids-22.08-updates

Dockerfile

@@ -1,4 +1,4 @@
-ARG BASE_IMAGE=rapidsai/rapidsai:22.02-cuda11.2-runtime-ubuntu18.04-py3.8
+ARG BASE_IMAGE=rapidsai/rapidsai:22.08-cuda11.5-runtime-ubuntu18.04-py3.8
 
 FROM ${BASE_IMAGE}
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
@@ -10,15 +10,15 @@ RUN git clone \
 
 
 RUN /opt/conda/envs/rapids/bin/pip install \
-    scanpy==1.8.1 wget pytabix dash-daq \
+    scanpy==1.9.1 wget pytabix dash-daq \
     dash-html-components dash-bootstrap-components dash-core-components
 
 RUN cd atacworks && /opt/conda/envs/rapids/bin/pip install .
 
 WORKDIR /workspace
 ENV HOME /workspace
 RUN git clone \
-    https://github.com/clara-parabricks/rapids-single-cell-examples.git \
+    https://github.com/NVIDIA-Genomics-Research/rapids-single-cell-examples.git \
     rapids-single-cell-examples
 
 ARG GIT_BRANCH=master
@@ -34,5 +34,5 @@ CMD jupyter-lab \
 		--NotebookApp.token="" \
 		--NotebookApp.password_required=False
 
-# ENV LD_LIBRARY_PATH /usr/local/cuda-10.2/compat
+# ENV LD_LIBRARY_PATH /usr/local/cuda-11.5/compat
 # RUN echo "export PATH=$PATH:/workspace/data" >> ~/.bashrc

conda/rapidgenomics_cuda11.0.viz.yml → conda/rapidgenomics_cuda11.5.viz.yml

@@ -7,9 +7,9 @@ channels:
   - defaults
 dependencies:
   - python=3.8
-  - cuml=22.02*
-  - cudf=22.02*
-  - cugraph=22.02*
+  - cuml=22.08*
+  - cudf=22.08*
+  - cugraph=22.08*
   - tabix
   - pytabix
   - cudatoolkit=11.5

conda/rapidgenomics_cuda11.0.yml → conda/rapidgenomics_cuda11.5.yml

@@ -7,9 +7,9 @@ channels:
   - defaults
 dependencies:
   - python=3.8
-  - cuml=22.02*
-  - cudf=22.02*
-  - cugraph=22.02*
+  - cuml=22.08*
+  - cudf=22.08*
+  - cugraph=22.08*
   - cudatoolkit=11.5
   - python-wget
   - cupy=9*

notebooks/rapids_scanpy_funcs.py

@@ -308,7 +308,7 @@ def select_groups(labels, groups_order_subset='all'):
         groups_ids = []
         for name in groups_order_subset:
             groups_ids.append(
-                cp.where(cp.array(labels.cat.categories.to_array().astype("int32")) == int(name))[0][0]
+                cp.where(cp.array(labels.cat.categories.to_numpy().astype("int32")) == int(name))[0][0]
             )
         if len(groups_ids) == 0:
             # fallback to index retrieval
@@ -321,9 +321,9 @@ def select_groups(labels, groups_order_subset='all'):
 
         groups_ids = [groups_id.item() for groups_id in groups_ids]
         groups_masks = groups_masks[groups_ids]
-        groups_order_subset = labels.cat.categories[groups_ids].to_array().astype(int)
+        groups_order_subset = labels.cat.categories[groups_ids].to_numpy().astype(int)
     else:
-        groups_order_subset = groups_order.to_array()
+        groups_order_subset = groups_order.to_numpy()
     return groups_order_subset, groups_masks
 
 
@@ -427,7 +427,7 @@ def rank_genes_groups(
     y = labels.loc[grouping]
 
     clf = LogisticRegression(**kwds)
-    clf.fit(X.get(), grouping.to_array().astype('float32'))
+    clf.fit(X.get(), grouping.astype('float32').to_numpy())
     scores_all = cp.array(clf.coef_).T
 
     for igroup, group in enumerate(groups_order):
@@ -660,7 +660,7 @@ def highly_variable_genes_filter(client,
     dispersion = variance / mean
 
     df = pd.DataFrame()
-    df['genes'] = genes.to_array()
+    df['genes'] = genes.to_numpy()
     df['means'] = mean.tolist()
     df['dispersions'] = dispersion.tolist()
     df['mean_bin'] = pd.cut(
@@ -699,7 +699,7 @@ def _cellranger_hvg(mean, mean_sq, genes, n_cells, n_top_genes):
 
     df = pd.DataFrame()
     # Note - can be replaced with cudf once 'cut' is added in 21.08
-    df['genes'] = genes.to_array()
+    df['genes'] = genes.to_numpy()
     df['means'] = mean.tolist()
     df['dispersions'] = dispersion.tolist()
     df['mean_bin'] = pd.cut(

notebooks/visualize.py

@@ -261,10 +261,10 @@ def start_graph(self, df):
             gdf = df.query(query)
             fig.add_trace(
                 go.Scattergl({
-                'x': gdf['x'].to_array(),
-                'y': gdf['y'].to_array(),
-                'text': gdf['labels'].to_array(),
-                'customdata': gdf['point_index'].to_array(),
+                'x': gdf['x'].to_numpy(),
+                'y': gdf['y'].to_numpy(),
+                'text': gdf['labels'].to_numpy(),
+                'customdata': gdf['point_index'].to_numpy(),
                 'name': 'Cluster ' + si,
                 'mode': 'markers',
                 'marker': {'size': 3, 'color': colors[i % len(colors)]}
@@ -286,10 +286,10 @@ def update_graph(self, df):
             gdf = df.query(query)
             fig = {
                 'type':'scattergl',
-                'x': gdf['x'].to_array(),
-                'y': gdf['y'].to_array(),
-                'text': gdf['labels'].to_array(),
-                'customdata': gdf['point_index'].to_array(),
+                'x': gdf['x'].to_numpy(),
+                'y': gdf['y'].to_numpy(),
+                'text': gdf['labels'].to_numpy(),
+                'customdata': gdf['point_index'].to_numpy(),
                 'name': 'Cluster ' + si,
                 'mode': 'markers',
                 'marker': {'size': 3, 'color': colors[i % len(colors)]}        }
@@ -341,7 +341,7 @@ def graph_violin(self, df, marker):
             query = 'labels == ' + si
             gdf = df.query(query)
 
-            y = gdf[marker_val].to_array()
+            y = gdf[marker_val].to_numpy()
             x = [i] * len(y)
             fig.add_trace(
                 go.Violin({