Mercurial > repos > iuc > data_manager_genomic_super_signature_ravmodels

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/README	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,1 @@
+Adds new entries to data tables for GenomicSuperSignature Replicable Axes of Variation (RAV) models (RAVmodels). See also https://www.biorxiv.org/content/10.1101/2021.05.26.445900v1.
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager/data_manager_genomic_super_signature_ravmodels.py	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,64 @@
+#!/usr/bin/env python
+# Dan Blankenberg
+
+
+import argparse
+import json
+import os
+import re
+import shutil
+import urllib.request
+
+
+OUTPUT_BASE_NAME = 'RAVmodel.rds'
+
+BUCKET_NAME = "genomic_super_signature"
+RAVMODEL_NAME = "RAVmodel_{prior}.rds"
+PRIOR_URL = "https://storage.googleapis.com/{bucket_name}/{prior_filename}"
+
+
+def get_prior_url(prior):
+    return PRIOR_URL.format(bucket_name=BUCKET_NAME, prior_filename=RAVMODEL_NAME.format(prior=prior))
+
+
+def clean(text):
+    return re.sub(r'[^\w\-_\.]', '-', text)
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--output', dest='output', action='store', type=str, default=None)
+    parser.add_argument('--json', dest='json', action='store', type=str, default=None)
+    parser.add_argument('--dbkey', dest='dbkey', action='store', type=str, default=None, help='dbkey')
+    parser.add_argument('--name', dest='name', action='store', type=str, default=None)
+    parser.add_argument('--id', dest='id', action='store', type=str, default=None)
+    parser.add_argument('--prior', dest='prior', action='store', type=str, default=None)
+    parser.add_argument('--url', dest='url', action='store', type=str, default=None)
+    parser.add_argument('--input', dest='input', action='store', type=str, default=None)
+    parser.add_argument('--symlink', dest='symlink', action='store_true', default=False)
+    parser.add_argument('--version', dest='version', action='store', type=str, default="0")
+    parser.add_argument('--tablename', dest='tablename', action='store', type=str, default="genomic_super_signature_ravmodels")
+    args = parser.parse_args()
+
+    os.makedirs(args.output, exist_ok=True)
+    output_filename = os.path.join(args.output, OUTPUT_BASE_NAME)
+    if args.symlink:
+        assert args.input, "You must provide a filename when using symlink functionality."
+        os.symlink(args.input, output_filename)
+    else:
+        url = args.url
+        if args.prior:
+            url = get_prior_url(args.prior)
+        if url:
+            urllib.request.urlretrieve(url, output_filename)
+        else:
+            assert args.input, "You must provide a filename, prior, or URL."
+            shutil.copyfile(args.input, output_filename)
+    data_manager_dict = {'data_tables': {args.tablename: [dict(value=clean(args.id), dbkey=args.dbkey, version=args.version, name=args.name, path=OUTPUT_BASE_NAME)]}}
+
+    with open(args.json, 'w') as fh:
+        json.dump(data_manager_dict, fh, sort_keys=True)
+
+
+if __name__ == "__main__":
+    main()
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager/data_manager_genomic_super_signature_ravmodels.xml	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,89 @@
+<tool id="data_manager_genomic_super_signature_ravmodels" name="Load GenomicSuperSignature RAVmodels" version="0.0.1" tool_type="manage_data" profile="19.01">
+    <description>into built-in cache</description>
+    <requirements>
+        <requirement type="package" version="3.8">python</requirement>
+    </requirements>
+    <command detect_errors="exit_code"><![CDATA[
+       #set $guess_name = ""
+       #set $guess_id = ""
+       python '$__tool_directory__/data_manager_genomic_super_signature_ravmodels.py'
+       --tablename 'genomic_super_signature_ravmodels'
+       --version '0'
+       --output '${out_file.extra_files_path}'
+       --json '${out_file}'
+       --dbkey '${rav_dbkey}'
+       #if str($reference_source.reference_source_selector) == 'prior':
+           --prior '${reference_source.prior_name}'
+           #if str($reference_source.prior_name) == 'C2':
+               #set $this_guess = "MSigDB C2 (curated gene sets)"
+           #elif str($reference_source.prior_name) == 'PLIERpriors':
+               #set $this_guess = "PLIERpriors: bloodCellMarkersIRISDMAP, svmMarkers, and canonicalPathways"
+           #else:
+               #set $this_guess = $reference_source.prior_name
+           #end if
+           #set $guess_name = $this_guess
+           #set $guess_id = $reference_source.prior_name
+       #elif str($reference_source.reference_source_selector) == 'url':
+           --url '${reference_source.user_url}'
+           #set $guess_name = $reference_source.user_url
+           #set $guess_id = $reference_source.user_url
+       #elif str($reference_source.reference_source_selector) == 'history':
+           --input '${reference_source.input_rav}'
+           #set $guess_name = $reference_source.input_rav.element_identifier
+           #set $guess_id = $reference_source.input_rav.id
+       #else:
+            #raise('Bad Selection')
+       #end if
+       --name '${$rav_name or $guess_name}'
+       --id '${rav_id or $guess_id}'
+    ]]></command>
+    <inputs>
+        <param name="rav_name" type="text" value="" label="Name of RAV" help="Leave blank to generate automatically" />
+        <param name="rav_id" type="text" value="" label="ID for RAV" help="Leave blank to generate automatically" />
+        <param name="rav_dbkey" type="text" value="hg38" label="Identifier for reference genome for RAV" />
+        <conditional name="reference_source">
+            <param name="reference_source_selector" type="select" label="Choose the source for the RAV">
+                <option value="prior" selected="True">Published GenomicSuperSignature RAVmodel</option>
+                <option value="url">URL</option>
+                <option value="history">History</option>
+            </param>
+            <when value="prior">
+                <param name="prior_name" type="select" optional="false" label="Published GenomicSuperSignature model">
+                    <option value="C2" selected="True">MSigDB C2 (curated gene sets)</option>
+                    <option value="PLIERpriors">PLIERpriors: bloodCellMarkersIRISDMAP, svmMarkers, and canonicalPathways</option>
+                </param>
+            </when>
+            <when value="url">
+                <param name="user_url" type="text" area="True" value="https://" optional="false" label="URL" />
+            </when>
+            <when value="history">
+                <param name="input_rav" type="data" format="rds" label="RAV Model file" />
+            </when>
+        </conditional>
+    </inputs>
+    <outputs>
+        <data name="out_file" format="data_manager_json"/>
+    </outputs>
+    <tests>
+        <!-- TODO: need some way to test that new entry was added to data table -->
+        <test>
+            <param name="rav_name" value=""/>
+            <param name="rav_id" value=""/>
+            <param name="rav_dbkey" value="hg38"/>
+            <param name="dbkey_source|reference_source_selector" value="prior"/>
+            <param name="dbkey_source|prior_name" value="C2"/>
+            <output name="out_file" file="RAVmodel_prior_C2.data_manager_json"/>
+        </test>
+    </tests>
+    <help>
+**What it does**
+
+Fetches a reference RAVmodel from various sources (Google Cloud, URL, Galaxy History, or a server file) and populates the "genomic_super_signature_ravmodels" data table.
+
+------
+
+.. class:: infomark
+
+**Notice:** If you leave name or id blank, it will be generated automatically.
+    </help>
+</tool>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/data_manager_conf.xml	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,21 @@
+<?xml version="1.0"?>
+<data_managers>
+    <data_manager tool_file="data_manager/data_manager_genomic_super_signature_ravmodels.xml" id="data_manager_genomic_super_signature_ravmodels">
+        <data_table name="genomic_super_signature_ravmodels">
+            <output>
+                <column name="value" />
+                <column name="dbkey" />
+                 <column name="version" />
+                <column name="name" />
+                <column name="path" output_ref="out_file" >
+                    <move type="directory" relativize_symlinks="True">
+                        <!-- <source>${path}</source>--> <!-- out_file.extra_files_path is used as base by default --> <!-- if no source, eg for type=directory, then refers to base -->
+                        <target base="${GALAXY_DATA_MANAGER_DATA_PATH}">${dbkey}/genomic_super_signature_ravmodels/${version}/${value}</target>
+                    </move>
+                    <value_translation>${GALAXY_DATA_MANAGER_DATA_PATH}/${dbkey}/genomic_super_signature_ravmodels/${version}/${value}/${path}</value_translation>
+                    <value_translation type="function">abspath</value_translation>
+                </column>
+            </output>
+        </data_table>
+    </data_manager>
+</data_managers>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/RAVmodel_prior_C2.data_manager_json	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,1 @@
+{"data_tables": {"genomic_super_signature_ravmodels": [{"dbkey": "hg38", "name": "MSigDB C2 (curated gene sets)", "path": "RAVmodel.rds", "value": "C2", "version": "0"}]}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tool-data/genomic_super_signature_ravmodels.loc.sample	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,1 @@
+#<value/id>	<dbkey>	<RAV format version>	<display_name>	<path>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tool_data_table_conf.xml.sample	Wed Jan 12 19:07:19 2022 +0000
@@ -0,0 +1,7 @@
+<tables>
+    <!-- Locations of RAVmodels for GenomicSuperSignature -->
+    <table name="genomic_super_signature_ravmodels" comment_char="#">
+        <columns>value, dbkey, version, name, path</columns>
+        <file path="tool-data/genomic_super_signature_ravmodels.loc" />
+    </table>
+</tables>