changeset 4:828d961faea7 draft

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_busco/ commit d46d2ec37b9b8f08d58472978c51b4c46e7ed18c

--- a/data_manager/busco_fetcher.xml	Thu Feb 20 20:14:38 2025 +0000
+++ b/data_manager/busco_fetcher.xml	Fri Apr 04 10:12:36 2025 +0000
@@ -2,15 +2,27 @@
     <description>dataset dowloader</description>
         <macros>
         <token name="@TOOL_VERSION@">5.8.0</token>
-        <token name="@VERSION_SUFFIX@">0</token>
+        <token name="@VERSION_SUFFIX@">1</token>
     </macros>
     <requirements>
         <requirement type="package" version="@TOOL_VERSION@">busco</requirement>
     </requirements>
     <command detect_errors="exit_code"><![CDATA[
-        mkdir '$out_file.extra_files_path' &&
         busco --download_path '$out_file.extra_files_path' --download '$lineage' &&
-        cp '$dmjson' '$out_file' 
+        ## remove unwanted lineage folders and placement files
+        ## https://gitlab.com/ezlab/busco/-/issues/784
+        if [ -d '$out_file.extra_files_path'/lineages ]; then
+            find '$out_file.extra_files_path'/lineages/ -mindepth 1 -maxdepth 1 ! -name '*_$datasets_version*' -exec rm -rf {} \;;
+        fi &&
+        if [ -d '$out_file.extra_files_path'/placement_files ]; then
+            find '$out_file.extra_files_path'/placement_files/ -mindepth 1 -maxdepth 1 ! -name '*_$datasets_version*' -delete;
+        fi &&
+        ## unpack faa.gz files ready to use
+        ## https://gitlab.com/ezlab/busco/-/issues/789
+        if [ -d '$out_file.extra_files_path'/lineages ]; then
+            find '$out_file.extra_files_path'/lineages/ -name "*.faa.gz" -exec gunzip {} \;;
+        fi
+        && cp '$dmjson' '$out_file' 
     ]]></command>
     <configfiles>
         <configfile name="dmjson"><![CDATA[
@@ -19,8 +31,8 @@
       "data_tables":{
         "busco_database":[
           {
-            "value":"$lineage #echo datetime.now().strftime('%Y-%m-%d-%H%M%S')#",
-            "name":"$lineage #echo datetime.now().strftime('%Y-%m-%d-%H%M%S')#",
+            "value":"$lineage $datasets_version #echo datetime.now().strftime('%Y-%m-%d-%H%M%S')#",
+            "name":"$lineage $datasets_version #echo datetime.now().strftime('%Y-%m-%d-%H%M%S')#",
             "version":"@TOOL_VERSION@",
             "path":"$out_file.extra_files_path"
           }
@@ -30,12 +42,16 @@
             </configfile>
         </configfiles>
     <inputs>
-        <param name="lineage" type="select" label="Select the lineage to be downloaded">
+        <param name="lineage" argument="--download" type="select" label="Select the lineage to be downloaded">
             <option value="all">All</option>
             <option value="prokaryota">Prokaryota</option>
             <option value="eukaryota">Eukaryota</option>
             <option value="virus">Virus</option>
         </param>
+        <param argument="--datasets_version" type="select" label="OrthoDB version" help="By March 2025 OrthoDB 12 datasets were only available for Prokaryota (https://busco-data2.ezlab.org/v5/data/)">
+            <option value="odb10" selected="true">OrthoDB 10</option>
+            <option value="odb12">OrthoDB 12</option>
+        </param>
     </inputs>
     <outputs>
         <data name="out_file" format="data_manager_json" label="BUSCO data manager: JSON"/>