Mercurial > repos > iuc > data_manager_gtdbtk_database_installer

changeset 7:3b1d503c6260 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_gtdbtk_database_installer commit 671e8c706fa211b6ec1c476d3d1a36d098822fe5
author iuc
date Thu, 03 Oct 2024 12:42:22 +0000
parents df84aaed4769
children 750d902de22c
files data_manager/gtdbtk_database_installer.py data_manager/gtdbtk_database_installer.xml test-data/gtdbtk_database_metadata_versioned.loc test-data/gtdbtk_database_versioned.loc
diffstat 3 files changed, 66 insertions(+), 76 deletions(-) [+]
line wrap: on
line diff
--- a/data_manager/gtdbtk_database_installer.py	Mon Sep 09 09:01:39 2024 +0000
+++ b/data_manager/gtdbtk_database_installer.py	Thu Oct 03 12:42:22 2024 +0000
@@ -7,7 +7,7 @@
 import shutil
 import sys
 import tarfile
-from datetime import datetime
+from datetime import date
 from urllib.parse import urlparse
 from urllib.request import HTTPError, Request, urlopen
 
@@ -118,7 +118,19 @@
         return target_directory
 
 
-def download(database_name, release, meta, test, out_file):
+def create_data_manager_entry(database_name, release, file_path):
+    time = date.today().strftime("%Y-%m-%d")
+    data_manager_entry = {}
+    data_manager_entry["value"] = (
+        f"{database_name.replace(' ', '_').lower()}_release_{release}_downloaded_{time}"
+    )
+    data_manager_entry["name"] = f"{database_name} - release {release} ({time})"
+    data_manager_entry["path"] = file_path
+    data_manager_entry["version"] = release
+    return data_manager_entry
+
+
+def download(release, meta, test, out_file):
 
     with open(out_file) as fh:
         params = json.load(fh)
@@ -137,72 +149,56 @@
             for url in items.values():
                 assert is_urlfile(url)
 
-    # download both taxonomy metadata tables
+    data_manager_json = {"data_tables": {}}
+
+    # download taxonomy metadata tables
     if meta:
         url = urls[release]["meta_ar"]
-        file_path = url_download(url, target_directory, meta)
+        url_download(url, target_directory, meta)
         url = urls[release]["meta_bac"]
         file_path = url_download(url, target_directory, meta)
+
+        data_manager_json["data_tables"]["gtdbtk_database_metadata_versioned"] = [
+            create_data_manager_entry("Metadata Tables", release, file_path)
+        ]
     # download the full DB
     else:
         url = urls[release]["full"]
         file_path = url_download(url, target_directory, meta)
-
-    time = datetime.utcnow().strftime("%Y-%m-%d")
-
-    data_manager_json = {"data_tables": {}}
-    data_manager_entry = {}
-    data_manager_entry["value"] = f"{database_name}_release_{release}_downloaded_{time}"
-    data_manager_entry["name"] = database_name
-    data_manager_entry["path"] = file_path
-    data_manager_entry["version"] = release
+        data_manager_json["data_tables"]["gtdbtk_database_versioned"] = [
+            create_data_manager_entry("Full Database", release, file_path)
+        ]
 
     # store in dedicated metadata table
-    if meta:
-        data_manager_json["data_tables"][
-            "gtdbtk_database_metadata_versioned"
-        ] = data_manager_entry
-    else:
-        data_manager_json["data_tables"][
-            "gtdbtk_database_versioned"
-        ] = data_manager_entry
-
     with open(out_file, "w") as fh:
         json.dump(data_manager_json, fh, sort_keys=True)
 
 
-parser = argparse.ArgumentParser()
-
-parser.add_argument(
-    "--database_name", dest="database_name", help="GTDB-Tk database display name"
-)
-
-parser.add_argument("--version", dest="version", help="DB version")
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
 
-parser.add_argument(
-    "--release", dest="release", help="Release of the GTDB-Tk database version"
-)
-parser.add_argument("--out_file", dest="out_file", help="JSON output file")
-parser.add_argument(
-    "--meta",
-    dest="meta",
-    action="store_true",
-    help="Store meta data flag",
-)
+    parser.add_argument("--version", dest="version", help="DB version")
+    parser.add_argument(
+        "--release", dest="release", help="Release of the GTDB-Tk database version"
+    )
+    parser.add_argument("--out_file", dest="out_file", help="JSON output file")
+    parser.add_argument(
+        "--meta",
+        dest="meta",
+        action="store_true",
+        help="Store meta data flag",
+    )
+    parser.add_argument(
+        "--test",
+        dest="test",
+        action="store_true",
+        help="Run test",
+    )
+    args = parser.parse_args()
 
-parser.add_argument(
-    "--test",
-    dest="test",
-    action="store_true",
-    help="Run test",
-)
-
-args = parser.parse_args()
-
-download(
-    args.database_name,
-    args.release,
-    args.meta,
-    args.test,
-    args.out_file,
-)
+    download(
+        args.release,
+        args.meta,
+        args.test,
+        args.out_file,
+    )
--- a/data_manager/gtdbtk_database_installer.xml	Mon Sep 09 09:01:39 2024 +0000
+++ b/data_manager/gtdbtk_database_installer.xml	Thu Oct 03 12:42:22 2024 +0000
@@ -2,7 +2,7 @@
     <description></description>
     <macros>
         <token name="@TOOL_VERSION@">202</token>
-        <token name="@VERSION_SUFFIX@">2</token>
+        <token name="@VERSION_SUFFIX@">3</token>
         <token name="@PROFILE@">20.09</token>
     </macros>
     <requirements>
@@ -10,18 +10,16 @@
     </requirements>
     <command>
     <![CDATA[
-        python '$__tool_directory__/gtdbtk_database_installer.py'
-          --database_name '$database_name'
-          --release '$release'
-          --out_file '$out_file'
-          $meta
-          $test
+python '$__tool_directory__/gtdbtk_database_installer.py'
+    --release '$release'
+    --out_file '$out_file'
+    $meta
+    $test
     ]]>
     </command>
     <inputs>
-	    <param name="database_name" type="text" value="" label="Database name or description" help="This value will be displayed in the GTDB-Tk Database select list"/>
-        <param name="meta" type="boolean" truevalue="--meta" falsevalue="" checked="false" label="Only store GTDBTK metadata in a dedicated data table. " />
-        <param name="test" type="hidden" value="" checked="false" label="Run a dry test run !" />
+	    <param name="meta" type="boolean" truevalue="--meta" falsevalue="" checked="false" label="Only store GTDBTK metadata in a dedicated data table. " />
+        <param name="test" type="hidden" value="" checked="false" label="Run a dry test run!" />
         <param name="release" type="select" multiple="false" label="GTDB Release">
             <option value="202">202</option>
             <option value="207">207</option>
@@ -34,40 +32,36 @@
     </outputs>
     <tests>
         <test>
-            <!-- TODO -->
             <!-- Not actually installing a huge GTDB-Tk database -->
             <!-- but it will check if all urls exist -->
             <param name="release" value="202"/>
-            <param name="database_name" value="GTDB-Tk database release 202"/>
             <param name="test" value="--test"/>
             <output name="out_file">
                 <assert_contents>
-                    <has_text text="GTDB-Tk database release 202"/>
-                    <has_text text="release_202"/>
+                    <has_text text="Full Database - release 202 "/>
+                    <has_text text="full_database_release_202_downloaded"/>
                 </assert_contents>
             </output>
         </test>
         <test>
             <!-- Test meta data download with tsv.gz-->
             <param name="release" value="220"/>
-            <param name="database_name" value="GTDB-Tk database release 220 metadata"/>
             <param name="meta" value="true"/>
             <output name="out_file">
                 <assert_contents>
-                    <has_text text="GTDB-Tk database release 220 metadata"/>
-                    <has_text text="release_220"/>
+                    <has_text text="Metadata Tables - release 220"/>
+                    <has_text text="metadata_tables_release_220_downloaded_"/>
                 </assert_contents>
             </output>
         </test>
         <test>
             <!-- Test meta data download with tar.gz -->
             <param name="release" value="207"/>
-            <param name="database_name" value="GTDB-Tk database release 207 metadata"/>
             <param name="meta" value="true"/>
             <output name="out_file">
                 <assert_contents>
-                    <has_text text="GTDB-Tk database release 207 metadata"/>
-                    <has_text text="release_207"/>
+                    <has_text text="Metadata Tables - release 207"/>
+                    <has_text text="metadata_tables_release_207_downloaded_"/>
                 </assert_contents>
             </output>
         </test>
--- a/test-data/gtdbtk_database_metadata_versioned.loc	Mon Sep 09 09:01:39 2024 +0000
+++ b/test-data/gtdbtk_database_metadata_versioned.loc	Thu Oct 03 12:42:22 2024 +0000
@@ -2,4 +2,4 @@
 # to use a directory of GTDB-Tk databases.  The gtdbtk_databases.loc
 # file has this format (longer white space characters are TAB characters):
 #
-# <unique_build_id> <display_name> <version> <directory_path>
+# <unique_build_id> <display_name> <version> <directory_path>
\ No newline at end of file