Mercurial > repos > matthias > data_manager_dada2

changeset 4:51b90d282704 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bernt-matthias/mb-galaxy-tools/tree/master/data_managers/data_manager_dada2 commit 2f47c03d1b775ab796c75c806c1ab247bb51359d
author matthias
date Fri, 10 May 2019 11:52:03 -0400
parents 3a4ee8bf012a
children b899349a9c2b
files data_manager/.dada2_fetcher.xml.swp data_manager/.data_manager.py.swp data_manager/dada2_fetcher.xml data_manager/test-data/PR24.11.1_json data_manager/test-data/RefSeq_RDP2018_json data_manager/test-data/greengenes13.84_json data_manager/test-data/gtdb2018_json data_manager/test-data/hitdb1_json data_manager/test-data/rdp16_json data_manager/test-data/silva132_json data_manager/test-data/silvaeuk132_json data_manager/test-data/unite8fungi_json test-data/PR24.11.1_json test-data/RefSeq_RDP2018_json test-data/dada2_species.loc test-data/dada2_taxonomy.loc test-data/greengenes13.84_json test-data/gtdb2018_json test-data/hitdb1_json test-data/rdp16_json test-data/silva132_json test-data/silvaeuk132_json test-data/unite8fungi_json tool_data_table_conf.xml.test
diffstat 24 files changed, 66 insertions(+), 21 deletions(-) [+]
line wrap: on
line diff
Binary file data_manager/.dada2_fetcher.xml.swp has changed
Binary file data_manager/.data_manager.py.swp has changed
--- a/data_manager/dada2_fetcher.xml	Tue Apr 09 07:18:57 2019 -0400
+++ b/data_manager/dada2_fetcher.xml	Fri May 10 11:52:03 2019 -0400
@@ -1,10 +1,11 @@
 <?xml version="1.0"?>
-<tool id="dada_fetcher" name="dada2 data manager" version="0.0.5">
+<tool id="dada_fetcher" name="dada2 data manager" tool_type="manage_data" version="0.0.5">
     <description>Download reference databases</description>
     <command detect_errors="exit_code"><![CDATA[
     python '$__tool_directory__/data_manager.py'
-    --out '${out_file}'
-    --dataset '$db_cond.db_select'_'$db_cond.version_select'
+	--out '$out_file'
+	#set dataset = str($db_cond.db_select) + '_' + str($db_cond.version_select)
+    --dataset '$dataset'
     ]]>
     </command>
     <inputs>
@@ -76,47 +77,47 @@
     <tests>
         <test>
             <param name="db_cond|db_select" value="silva"/>
-            <param name="db_cond|db_version" value="132"/>
+            <param name="db_cond|version_select" value="132"/>
             <output name="out_file" file="silva132_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="rdp"/>
-            <param name="db_cond|db_version" value="16"/>
+            <param name="db_cond|version_select" value="16"/>
             <output name="out_file" file="rdp16_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="greengenes"/>
-            <param name="db_cond|db_version" value="13.84"/>
+            <param name="db_cond|version_select" value="13.84"/>
             <output name="out_file" file="greengenes13.84_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="unite"/>
-            <param name="db_cond|db_version" value="8.0_fungi"/>
+            <param name="db_cond|version_select" value="8.0_fungi"/>
             <output name="out_file" file="unite8fungi_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="RefSeq_RDP"/>
-            <param name="db_cond|db_version" value="2018_05"/>
+            <param name="db_cond|version_select" value="2018_05"/>
             <output name="out_file" file="RefSeq_RDP2018_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="gtdb"/>
-            <param name="db_cond|db_version" value="2018_11"/>
+            <param name="db_cond|version_select" value="2018_11"/>
             <output name="out_file" file="gtdb2018_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="hitdb"/>
-            <param name="db_cond|db_version" value="1"/>
+            <param name="db_cond|version_select" value="1"/>
             <output name="out_file" file="hitdb1_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="silva_euk_18S"/>
-            <param name="db_cond|db_version" value="132"/>
+            <param name="db_cond|version_select" value="132"/>
             <output name="out_file" file="silvaeuk132_json"/>
         </test>
         <test>
             <param name="db_cond|db_select" value="PR2"/>
-            <param name="db_cond|db_version" value="4.11.1"/>
+            <param name="db_cond|version_select" value="4.11.1"/>
             <output name="out_file" file="PR24.11.1_json"/>
         </test>
     </tests>
--- a/data_manager/test-data/PR24.11.1_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "PR2_4.11.1.taxonomy", "name": "Protist Ribosomal Reference database (PR2) 4.11.1", "value": "PR2_4.11.1", "taxlevels": "Kingdom,Supergroup,Division,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/RefSeq_RDP2018_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "RefSeq_RDP_2018_05.taxonomy", "name": "NCBI RefSeq 16S rRNA database supplemented by RDP (05/2018)", "value": "RefSeq_RDP_2018_05", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/greengenes13.84_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "greengenes_13.84.taxonomy", "name": "GreenGenes version 13.84", "value": "greengenes_13.84", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/gtdb2018_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "gtdb_2018_11.taxonomy", "name": "GTDB: Genome Taxonomy Database (Bacteria &amp; Archaea) (11/2018)", "value": "gtdb_2018_11", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/hitdb1_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "hitdb_1.taxonomy", "name": "HitDB version 1 (Human InTestinal 16S rRNA)", "value": "hitdb_1", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/rdp16_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_species": {"path": "rdp_16.species", "name": "RDP trainset 16", "value": "rdp_16"}, "dada2_taxonomy": {"path": "rdp_16.taxonomy", "name": "RDP trainset 16", "value": "rdp_16", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/silva132_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_species": {"path": "silva_132.species", "name": "Silva version 132", "value": "silva_132"}, "dada2_taxonomy": {"path": "silva_132.taxonomy", "name": "Silva version 132", "value": "silva_132", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/silvaeuk132_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "silva_euk_18S_132.taxonomy", "name": "Silva version 132 Eukaryotic 18S", "value": "silva_euk_18S_132", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- a/data_manager/test-data/unite8fungi_json	Tue Apr 09 07:18:57 2019 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-{"data_tables": {"dada2_taxonomy": {"path": "unite_8.0_fungi.taxonomy", "name": "UNITE: General Fasta release 8.0 for Fungi", "value": "unite_8.0_fungi", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/PR24.11.1_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "PR2_4.11.1.taxonomy", "name": "Protist Ribosomal Reference database (PR2) 4.11.1", "value": "PR2_4.11.1", "taxlevels": "Kingdom,Supergroup,Division,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/RefSeq_RDP2018_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "RefSeq_RDP_2018_05.taxonomy", "name": "NCBI RefSeq 16S rRNA database supplemented by RDP (05/2018)", "value": "RefSeq_RDP_2018_05", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/dada2_species.loc	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,13 @@
+# This is a sample file distributed with Galaxy that is used to define a
+# list of dada2 reference data sets for species assignment, using three
+# tab separated columns:
+#
+# <unique_build_id>	<display_name>	<fasta_file_path>
+#
+# Datasets can be retrieved from http://busco.ezlab.org/frame_wget.html
+#
+# Datasets can be retrieved from https://benjjneb.github.io/dada2/training.html
+silva_132	Silva version 132	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/silva_132.species
+rdp_16	RDP trainset 16	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/rdp_16.species
+rdp_16	RDP trainset 16	/tmp/tmpLsjmNt/tmppH8mTW/tmpqUnkc6/database/data_manager_tool-data1EI7iU/dada2/rdp_16.species
+silva_132	Silva version 132	/tmp/tmpLsjmNt/tmppH8mTW/tmpqUnkc6/database/data_manager_tool-data1EI7iU/dada2/silva_132.species
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/dada2_taxonomy.loc	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,20 @@
+# This is a sample file distributed with Galaxy that is used to define a
+# list of dada2 reference data sets for taxonomy assignment, using three
+# tab separated columns:
+#
+# <unique_build_id>	<display_name>	<fasta_file_path>	<taxlevels>
+#
+# Datasets can be retrieved from https://benjjneb.github.io/dada2/training.html
+# 
+# taxlevels is a comma separated list of taxonomy levels
+silva_132	Silva version 132	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/silva_132.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+rdp_16	RDP trainset 16	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/rdp_16.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+greengenes_13.84	GreenGenes version 13.84	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/greengenes_13.84.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+unite_8.0_fungi	UNITE: General Fasta release 8.0 for Fungi	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/unite_8.0_fungi.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+RefSeq_RDP_2018_05	NCBI RefSeq 16S rRNA database supplemented by RDP (05/2018)	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/RefSeq_RDP_2018_05.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+gtdb_2018_11	GTDB: Genome Taxonomy Database (Bacteria &amp; Archaea) (11/2018)	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/gtdb_2018_11.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+hitdb_1	HitDB version 1 (Human InTestinal 16S rRNA)	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/hitdb_1.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+silva_euk_18S_132	Silva version 132 Eukaryotic 18S	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/silva_euk_18S_132.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+PR2_4.11.1	Protist Ribosomal Reference database (PR2) 4.11.1	/tmp/tmpC8Yc3F/tmpzwAZwA/tmpZp3_e0/database/data_manager_tool-datal4tFOr/dada2/PR2_4.11.1.taxonomy	Kingdom,Supergroup,Division,Class,Order,Family,Genus,Species
+rdp_16	RDP trainset 16	/tmp/tmpLsjmNt/tmppH8mTW/tmpqUnkc6/database/data_manager_tool-data1EI7iU/dada2/rdp_16.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
+silva_132	Silva version 132	/tmp/tmpLsjmNt/tmppH8mTW/tmpqUnkc6/database/data_manager_tool-data1EI7iU/dada2/silva_132.taxonomy	Kingdom,Phylum,Class,Order,Family,Genus,Species
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/greengenes13.84_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "greengenes_13.84.taxonomy", "name": "GreenGenes version 13.84", "value": "greengenes_13.84", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/gtdb2018_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "gtdb_2018_11.taxonomy", "name": "GTDB: Genome Taxonomy Database (Bacteria &amp; Archaea) (11/2018)", "value": "gtdb_2018_11", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/hitdb1_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "hitdb_1.taxonomy", "name": "HitDB version 1 (Human InTestinal 16S rRNA)", "value": "hitdb_1", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/rdp16_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_species": {"path": "rdp_16.species", "name": "RDP trainset 16", "value": "rdp_16"}, "dada2_taxonomy": {"path": "rdp_16.taxonomy", "name": "RDP trainset 16", "value": "rdp_16", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/silva132_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_species": {"path": "silva_132.species", "name": "Silva version 132", "value": "silva_132"}, "dada2_taxonomy": {"path": "silva_132.taxonomy", "name": "Silva version 132", "value": "silva_132", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/silvaeuk132_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "silva_euk_18S_132.taxonomy", "name": "Silva version 132 Eukaryotic 18S", "value": "silva_euk_18S_132", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/unite8fungi_json	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,1 @@
+{"data_tables": {"dada2_taxonomy": {"path": "unite_8.0_fungi.taxonomy", "name": "UNITE: General Fasta release 8.0 for Fungi", "value": "unite_8.0_fungi", "taxlevels": "Kingdom,Phylum,Class,Order,Family,Genus,Species"}}}
\ No newline at end of file
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tool_data_table_conf.xml.test	Fri May 10 11:52:03 2019 -0400
@@ -0,0 +1,11 @@
+<?xml version="1.0"?>
+<tables>
+    <table name="dada2_species" comment_char="#">
+        <columns>value, name, path</columns>
+		<file path="${__HERE__}/test-data/dada2_species.loc" />
+    </table>
+    <table name="dada2_taxonomy" comment_char="#">
+        <columns>value, name, path, taxlevels</columns>
+		<file path="${__HERE__}/test-data/dada2_taxonomy.loc" />
+    </table>
+</tables>