Read Anndata objects in backed mode (from disk) wherever possible

tools/anndata/export.xml

@@ -12,7 +12,7 @@
     <configfiles>
         <configfile name="script_file"><![CDATA[
 @CMD_imports@
-adata = ad.read_h5ad('$input')
+adata = ad.read_h5ad('$input', backed='r')
 adata.write_csvs('.', sep="\t", skip_data = False)
     ]]></configfile>
 </configfiles>

tools/anndata/import.xml

@@ -1,4 +1,4 @@
-<tool id="anndata_import" name="Import Anndata" version="@TOOL_VERSION@+galaxy2" profile="@PROFILE@">
+<tool id="anndata_import" name="Import Anndata" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
     <description>from different formats</description>
     <macros>
         <import>macros.xml</import>

tools/anndata/inspect.xml

@@ -17,7 +17,7 @@ from scipy import io
 
 pd.options.display.precision = 15
 
-adata = ad.read_h5ad('$input')
+adata = ad.read_h5ad('$input', backed='r')
 
 #if $inspect.info == 'general'
 with open('$general', 'w', encoding="utf-8") as f:
@@ -202,7 +202,13 @@ pd.DataFrame(adata.varm['PCs']).to_csv("$varm_PCs", sep="\t", index = False)
             <conditional name="inspect">
                 <param name="info" value="general"/>
             </conditional>
-            <output name="general" value="inspect.general.txt"/>
+            <output name="general">
+                <assert_contents>
+                    <has_text_matching expression="AnnData object with n_obs .* n_vars = 500 .* 11" />
+                    <has_text_matching expression="obs: 'cell_type'" />
+                    <has_text_matching expression="uns: 'highlights', 'iroot'" />
+                </assert_contents>            
+            </output>
         </test>
         <test expect_num_outputs="1">
             <!-- test 2: X -->

tools/anndata/macros.xml

@@ -1,6 +1,6 @@
 <macros>
     <token name="@TOOL_VERSION@">0.11.4</token>
-    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@VERSION_SUFFIX@">3</token>
     <token name="@PROFILE@">21.09</token>
     <xml name="requirements">
         <requirements>
@@ -39,7 +39,7 @@ AnnData stores a data matrix `X` together with annotations of observations `obs`
 
 
 AnnData stores observations (samples) of variables (features) in the rows of a matrix. This is the convention of the modern classics
-of statistics (`Hastie et al., 2009 <https://web.stanford.edu/~hastie/ElemStatLearn/>`__)  and machine learning (Murphy, 2012), the convention of dataframes both in R and Python and the established statistics
+of statistics (Hastie et al., 2009)  and machine learning (Murphy, 2012), the convention of dataframes both in R and Python and the established statistics
 and machine learning packages in Python (statsmodels, scikit-learn).
 
 More details on the `AnnData documentation

tools/anndata/manipulate.xml

@@ -21,11 +21,12 @@
         <configfile name="script_file"><![CDATA[
 @CMD_imports@
 
-adata = ad.read_h5ad('$input')
+adata = ad.read_h5ad('$input', backed='r')
 
 #if $manipulate.function == 'concatenate'
+adata = adata.to_memory()
     #for i, filepath in enumerate($manipulate.other_adatas)
-adata_$i = ad.read_h5ad('$filepath')
+adata_$i = ad.read_h5ad('$filepath', backed='r').to_memory()
     #end for
 adata = adata.concatenate(
     #for i, filepath in enumerate($manipulate.other_adatas)
@@ -118,6 +119,7 @@ del adata.var['${from_var}']
 adata.strings_to_categoricals()
 
 #else if $manipulate.function == 'transpose'
+adata = adata.to_memory()
 adata = adata.transpose()
 
 #else if $manipulate.function == 'add_annotation'
@@ -137,14 +139,15 @@ adata.obs = obs
 
 #else if $manipulate.function == 'split_on_obs'
 import os
+adata = adata.to_memory()
 res_dir = "output_split"
 os.makedirs(res_dir, exist_ok=True)
 for s,field_value in enumerate(adata.obs["${manipulate.key}"].unique()):
-    ad_s = adata[adata.obs.${manipulate.key} == field_value]
+    ad_s = adata[adata.obs.${manipulate.key} == field_value].copy()
     ad_s.write(f"{res_dir}/${manipulate.key}_{s}.h5ad", compression='gzip')
 
 #else if $manipulate.function == 'copy_obs'
-source_adata = ad.read_h5ad('$source_adata')
+source_adata = ad.read_h5ad('$source_adata', backed='r')
     #for $key in $manipulate.keys
 if '$key.source_key' in source_adata.obs:
         #if str($key.target_key) == '':
@@ -157,7 +160,7 @@ else:
     #end for
 
 #else if $manipulate.function == 'copy_uns'
-source_adata = ad.read_h5ad('$source_adata')
+source_adata = ad.read_h5ad('$source_adata', backed='r')
     #for $key in $manipulate.keys
 if '$key.source_key' in source_adata.uns:
         #if str($key.target_key) == '':
@@ -170,7 +173,7 @@ else:
     #end for
 
 #else if $manipulate.function == 'copy_embed'
-source_adata = ad.read_h5ad('$source_adata')
+source_adata = ad.read_h5ad('$source_adata', backed='r')
     #for $key in $manipulate.keys
 if '$key.source_key' in source_adata.obsm:
         #if $key.target_key is None
@@ -183,7 +186,7 @@ else:
     #end for
 
 #else if $manipulate.function == 'copy_layers'
-source_adata = ad.read_h5ad('$source_adata')
+source_adata = ad.read_h5ad('$source_adata', backed='r')
     #for $key in $manipulate.keys
 if '$key.source_key' in source_adata.layers:
         #if $key.target_key is None
@@ -196,14 +199,15 @@ else:
     #end for
 
 #else if $manipulate.function == 'copy_X'
-source_adata = ad.read_h5ad('$source_adata')
+source_adata = ad.read_h5ad('$source_adata', backed='r')
     #if $target_key is None
 adata.X = source_adata.X
     #else
 adata.layers['$target_key'] = source_adata.X
     #end if
 
 #else if $manipulate.function == 'save_raw'
+adata = adata.to_memory()
 adata.raw = adata
 
 #end if
@@ -924,7 +928,7 @@ This tool takes an AnnData dataset, manipulates it and returns it.
 
 The possible manipulations are:
 
-- Concatenate along the observations axis (`concatenate method <https://anndata.readthedocs.io/en/latest/generated/anndata.AnnData.concatenate.html>`__)
+- Concatenate along the observations axis (`concatenate method <https://anndata.readthedocs.io/en/stable/generated/anndata.AnnData.concatenate.html>`__)
 
     The `uns`, `varm` and `obsm` attributes are ignored.