Mercurial > repos > devteam > kraken_report

changeset 6:1902d3de4a4f 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/blob/master/tool_collections/kraken/ commit 06345505a91f3dcfa8a37dceb6f25e95806dddc8"
author iuc
date Wed, 04 Dec 2019 06:52:03 -0500
parents 6c09bed3444f
children 1c677354ab18
files kraken-mpa-report.xml kraken-report.xml macros.xml test-data/kraken-report/kraken_mpa_report_input1.tab test-data/kraken-report/kraken_mpa_report_input2.tab test-data/kraken-report/kraken_mpa_report_test1_output.tab test-data/kraken-report/kraken_report_test1.tab test-data/kraken-report/kraken_report_test1_output.tab test-data/kraken-report/test_database.loc test-data/kraken_mpa_report_input1.tab test-data/kraken_mpa_report_input2.tab test-data/kraken_mpa_report_test1_output.tab test-data/kraken_report_test1.tab test-data/kraken_report_test1_output.tab test-data/test_database.loc test-data/test_db/database.idx test-data/test_db/database.kdb test-data/test_db/taxonomy/names.dmp test-data/test_db/taxonomy/nodes.dmp tool-data/kraken_databases.loc.sample
diffstat 20 files changed, 80 insertions(+), 204 deletions(-) [+]
line wrap: on
line diff
--- a/kraken-mpa-report.xml	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,81 +0,0 @@
-<?xml version="1.0"?>
-<tool id="kraken-mpa-report" name="Kraken-mpa-report" version="@WRAPPER_VERSION@">
-    <description>view report of classification for multiple samples</description>
-    <macros>
-        <import>macros.xml</import>
-    </macros>
-    <expand macro="requirements" />
-    <expand macro="version_command" />
-    <command detect_errors="exit_code"><![CDATA[
-    #set $names = []
-
-    #for $input_count, $input_classification in enumerate( $classification ):
-        #set $name_base = str( getattr( $input_classification, 'element_identifier', 'sample' ) ).replace( "/", '-' ).replace( "\t", "-" )
-        #set $name = $name_base
-        #set $i = 1
-        #while $name in $names:
-            #set $name = "%s_%s" % ( $name_base, $i )
-            #set $i = $i + 1
-        #end while
-        #silent $names.append( $name )
-        ln -s '${input_classification}' '${name}' &&
-    #end for
-
-    @SET_DATABASE_PATH@ &&
-
-    kraken-mpa-report
-        @INPUT_DATABASE@
-        #for $name in $names:
-            '${name}'
-        #end for
-
-        ${show_zeros}
-        ${header_line}
-
-        > '$output_report'
-    ]]></command>
-    <inputs>
-        <param name="classification" format="tabular" label="Kraken output" multiple="True" type="data" />
-        <param name="show_zeros" argument="--show-zeros" type="boolean" falsevalue=""  truevalue="--show-zeros" checked="False"
-               label="Display taxa even if they lack a read in any sample" />
-        <param name="header_line" argument="--header-line" type="boolean" truevalue="--header-line" falsevalue="" checked="False"
-               label="Display a header line indicating sample IDs"/>
-
-        <expand macro="input_database" />
-    </inputs>
-    <outputs>
-        <data format="tabular" name="output_report" />
-    </outputs>
-    <tests>
-        <test>
-            <param name="classification" value="kraken_mpa_report_input1.tab,kraken_mpa_report_input2.tab" ftype="tabular"/>
-            <param name="show_zeros" value="--show-zeros"/>
-            <param name="header_line" value="--header-line"/>
-            <param name="kraken_database" value="test_db"/>
-
-            <output name="output_report" ftype="tabular" file="kraken_mpa_report_test1_output.tab" />
-        </test>
-    </tests>
-    <help>
-<![CDATA[
-
-.. class:: warningmark
-
-**Note**: the database used must be the same as the one used in the original Kraken run
-
------
-
-**What is Does**
-
-Kraken-mpa-report summarizes read counts across taxonomic ranks for multiple samples. This is convenient for comparing results across multiple experiments, conditions, locations, etc.
-
------
-
-**Output**
-
-The output of kraken-mpa-report is a tab-delimited table, with one line per taxon.
-
-]]>
-    </help>
-    <expand macro="citations" />
-</tool>
--- a/kraken-report.xml	Thu Sep 13 09:45:15 2018 -0400
+++ b/kraken-report.xml	Wed Dec 04 06:52:03 2019 -0500
@@ -24,9 +24,9 @@
     </outputs>
     <tests>
         <test>
-            <param name="kraken_output" value="kraken_report_test1.tab" ftype="tabular"/>
-            <param name="kraken_database" value="test_db"/>
-            <output name="output_report" ftype="tabular" file="kraken_report_test1_output.tab"/>
+            <param name="kraken_output" value="kraken-report/kraken_report_test1.tab" ftype="tabular"/>
+            <param name="kraken_database" value="new_style_test_entry"/>
+            <output name="output_report" ftype="tabular" file="kraken-report/kraken_report_test1_output.tab"/>
         </test>
     </tests>
 
--- a/macros.xml	Thu Sep 13 09:45:15 2018 -0400
+++ b/macros.xml	Wed Dec 04 06:52:03 2019 -0500
@@ -1,9 +1,9 @@
 <?xml version="1.0"?>
 <macros>
-    <token name="@WRAPPER_VERSION@">1.2.4</token>
+    <token name="@WRAPPER_VERSION@">1.3.1</token>
     <xml name="requirements">
         <requirements>
-            <requirement type="package" version="0.10.6_eaf8fb68">kraken</requirement>
+            <requirement type="package" version="1.1.1">kraken</requirement>
         </requirements>
     </xml>
     <xml name="version_command">
@@ -23,6 +23,10 @@
             <citation type="doi">10.1186/gb-2014-15-3-r46</citation>
         </citations>
     </xml>
-    <token name="@INPUT_DATABASE@">--db '${kraken_database.fields.name}'</token>
-    <token name="@SET_DATABASE_PATH@">export KRAKEN_DB_PATH='${kraken_database.fields.path}'</token>
+    <!-- See the kraken_databases.loc.sample documentation,
+         this if statement is for backward compatibility as early
+         versions of the wrapper assumed the UI facing field name
+         was also part of the directory path -->
+    <token name="@SET_DATABASE_PATH@">if [ -d '${kraken_database.fields.path}/${kraken_database.fields.name}' ]; then export KRAKEN_DEFAULT_DB='${kraken_database.fields.path}/${kraken_database.fields.name}'; else export KRAKEN_DEFAULT_DB='${kraken_database.fields.path}'; fi</token>
+    <token name="@INPUT_DATABASE@">--db "\$KRAKEN_DEFAULT_DB"</token>
 </macros>
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/kraken_mpa_report_input1.tab	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,4 @@
+C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
+C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
+C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
+C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/kraken_mpa_report_input2.tab	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,4 @@
+C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
+C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
+C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
+C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/kraken_mpa_report_test1_output.tab	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,8 @@
+#Sample ID	kraken_mpa_report_input1.tab	kraken_mpa_report_input2.tab
+d__Bacteria	4	4
+d__Bacteria|p__Proteobacteria	4	4
+d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria	4	4
+d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales	4	4
+d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae	4	4
+d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae|g__Escherichia	4	4
+d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae|g__Escherichia|s__Escherichia_coli	4	4
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/kraken_report_test1.tab	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,4 @@
+C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
+C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
+C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
+C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/kraken_report_test1_output.tab	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,11 @@
+  0.00	0	0	U	0	unclassified
+100.00	4	0	-	1	root
+100.00	4	0	-	131567	  cellular organisms
+100.00	4	0	D	2	    Bacteria
+100.00	4	0	P	1224	      Proteobacteria
+100.00	4	0	C	1236	        Gammaproteobacteria
+100.00	4	0	O	91347	          Enterobacteriales
+100.00	4	0	F	543	            Enterobacteriaceae
+100.00	4	0	G	561	              Escherichia
+100.00	4	0	S	562	                Escherichia coli
+100.00	4	4	-	83333	                  Escherichia coli K-12
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/kraken-report/test_database.loc	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,20 @@
+# Tab separated with three columns:
+# - value (Galaxy records this in the Galaxy DB)
+# - name (Galaxy shows this in the UI)
+# - path (folder name containing the Kraken DB)
+#
+# Modern usage assumes $path is the folder containing
+# the Kraken database of interest:
+#
+new_style_test_entry	"Test Database"	${__HERE__}/test_db
+
+# The kraken wrapper still supports an alternative
+# legacy usage where $path/$name is the folder
+# containing the Kraken database of interest, meaning
+# $name must be both a folder name and human readable
+# description of the DB to show in the Galaxy UI:
+old_style_test_entry	test_db	${__HERE__}
+
+# NOTE: This legacy style should be avoided as other
+# tools also using the kraken_database.loc file may
+# not understand it!
--- a/test-data/kraken_mpa_report_input1.tab	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,4 +0,0 @@
-C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
-C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
-C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
-C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- a/test-data/kraken_mpa_report_input2.tab	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,4 +0,0 @@
-C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
-C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
-C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
-C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- a/test-data/kraken_mpa_report_test1_output.tab	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,8 +0,0 @@
-#Sample ID	kraken_mpa_report_input1.tab	kraken_mpa_report_input2.tab
-d__Bacteria	4	4
-d__Bacteria|p__Proteobacteria	4	4
-d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria	4	4
-d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales	4	4
-d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae	4	4
-d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae|g__Escherichia	4	4
-d__Bacteria|p__Proteobacteria|c__Gammaproteobacteria|o__Enterobacteriales|f__Enterobacteriaceae|g__Escherichia|s__Escherichia_coli	4	4
--- a/test-data/kraken_report_test1.tab	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,4 +0,0 @@
-C	gi|145231|gb|M33724.1|ECOALPHOA	83333	171	83333:162
-C	gi|145232|gb|M33725.1|ECOALPHOB	83333	183	83333:174
-C	gi|145234|gb|M33727.1|ECOALPHOE	83333	97	83333:88
-C	gi|146195|gb|J01619.1|ECOGLTA	83333	3850	83333:3841
--- a/test-data/kraken_report_test1_output.tab	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,11 +0,0 @@
-  0.00	0	0	U	0	unclassified
-100.00	4	0	-	1	root
-100.00	4	0	-	131567	  cellular organisms
-100.00	4	0	D	2	    Bacteria
-100.00	4	0	P	1224	      Proteobacteria
-100.00	4	0	C	1236	        Gammaproteobacteria
-100.00	4	0	O	91347	          Enterobacteriales
-100.00	4	0	F	543	            Enterobacteriaceae
-100.00	4	0	G	561	              Escherichia
-100.00	4	0	S	562	                Escherichia coli
-100.00	4	4	-	83333	                  Escherichia coli K-12
--- a/test-data/test_database.loc	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,1 +0,0 @@
-test_db	test_db	${__HERE__}
\ No newline at end of file
Binary file test-data/test_db/database.idx has changed
Binary file test-data/test_db/database.kdb has changed
--- a/test-data/test_db/taxonomy/names.dmp	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,74 +0,0 @@
-83333	|	Escherichia coli K-12	|		|	scientific name	|
-83333	|	Escherichia coli K12	|		|	equivalent name	|
-562	|	"Bacillus coli" Migula 1895	|		|	authority	|
-562	|	"Bacterium coli commune" Escherich 1885	|		|	authority	|
-562	|	"Bacterium coli" (Migula 1895) Lehmann and Neumann 1896	|		|	authority	|
-562	|	ATCC 11775	|		|	type material	|
-562	|	Bacillus coli	|		|	synonym	|
-562	|	Bacterium coli	|		|	synonym	|
-562	|	Bacterium coli commune	|		|	synonym	|
-562	|	CCUG 24	|		|	type material	|
-562	|	CCUG 29300	|		|	type material	|
-562	|	CIP 54.8	|		|	type material	|
-562	|	DSM 30083	|		|	type material	|
-562	|	Enterococcus coli	|		|	synonym	|
-562	|	Escherchia coli	|		|	misspelling	|
-562	|	Escherichia coli	|		|	scientific name	|
-562	|	Escherichia coli (Migula 1895) Castellani and Chalmers 1919	|		|	authority	|
-562	|	Escherichia sp. MAR	|		|	includes	|
-562	|	Escherichia/Shigella coli	|		|	equivalent name	|
-562	|	Eschericia coli	|		|	misspelling	|
-562	|	JCM 1649	|		|	type material	|
-562	|	LMG 2092	|		|	type material	|
-562	|	NBRC 102203	|		|	type material	|
-562	|	NCCB 54008	|		|	type material	|
-562	|	NCTC 9001	|		|	type material	|
-562	|	bacterium 10a	|		|	includes	|
-562	|	bacterium E3	|		|	includes	|
-561	|	Escherchia	|		|	misspelling	|
-561	|	Escherichia	|		|	scientific name	|
-561	|	Escherichia Castellani and Chalmers 1919	|		|	authority	|
-543	|	Enterobacteraceae	|		|	synonym	|
-543	|	Enterobacteraceae (ex Lapage 1979) Lapage 1982, fam. nov., nom. rev.	|		|	synonym	|
-543	|	Enterobacteriaceae	|		|	scientific name	|
-543	|	Enterobacteriaceae (ex Rahn 1937) Ewing et al. 1980, fam. nov., nom. rev.	|		|	synonym	|
-543	|	Enterobacteriaceae Rahn 1937	|		|	synonym	|
-543	|	gamma-3 proteobacteria	|	gamma-3 proteobacteria <#1>	|	in-part	|
-91347	|	'Enterobacteriales'	|		|	synonym	|
-91347	|	Enterobacteriaceae and related endosymbionts	|		|	synonym	|
-91347	|	Enterobacteriaceae group	|		|	synonym	|
-91347	|	Enterobacteriales	|		|	scientific name	|
-91347	|	enterobacteria	|	enterobacteria<blast91347>	|	blast name	|
-91347	|	gamma-3 proteobacteria	|	gamma-3 proteobacteria <#5>	|	in-part	|
-1236	|	Gammaproteobacteria	|		|	scientific name	|
-1236	|	Gammaproteobacteria Garrity et al. 2005	|		|	synonym	|
-1236	|	Proteobacteria gamma subdivision	|		|	synonym	|
-1236	|	Purple bacteria, gamma subdivision	|		|	synonym	|
-1236	|	g-proteobacteria	|	gamma proteos<blast1236>	|	blast name	|
-1236	|	gamma proteobacteria	|		|	synonym	|
-1236	|	gamma subdivision	|		|	synonym	|
-1236	|	gamma subgroup	|		|	synonym	|
-1224	|	Proteobacteria	|		|	scientific name	|
-1224	|	Proteobacteria Garrity et al. 2005	|		|	authority	|
-1224	|	Proteobacteria [class] Stackebrandt et al. 1988	|		|	authority	|
-1224	|	not Proteobacteria Cavalier-Smith 2002	|		|	authority	|
-1224	|	proteobacteria	|	proteobacteria<blast1224>	|	blast name	|
-1224	|	purple bacteria	|		|	common name	|
-1224	|	purple bacteria and relatives	|		|	common name	|
-1224	|	purple non-sulfur bacteria	|		|	common name	|
-1224	|	purple photosynthetic bacteria	|		|	common name	|
-1224	|	purple photosynthetic bacteria and relatives	|		|	common name	|
-2	|	Bacteria	|	Bacteria <prokaryote>	|	scientific name	|
-2	|	Monera	|	Monera <Bacteria>	|	in-part	|
-2	|	Procaryotae	|	Procaryotae <Bacteria>	|	in-part	|
-2	|	Prokaryota	|	Prokaryota <Bacteria>	|	in-part	|
-2	|	Prokaryotae	|	Prokaryotae <Bacteria>	|	in-part	|
-2	|	bacteria	|	bacteria <blast2>	|	blast name	|
-2	|	eubacteria	|		|	genbank common name	|
-2	|	not Bacteria Haeckel 1894	|		|	synonym	|
-2	|	prokaryote	|	prokaryote <Bacteria>	|	in-part	|
-2	|	prokaryotes	|	prokaryotes <Bacteria>	|	in-part	|
-1	|	all	|		|	synonym	|
-1	|	root	|		|	scientific name	|
-131567	|	biota	|		|	synonym	|
-131567	|	cellular organisms	|		|	scientific name	|
--- a/test-data/test_db/taxonomy/nodes.dmp	Thu Sep 13 09:45:15 2018 -0400
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,10 +0,0 @@
-83333	|	562	|	no rank	|		|	0	|	1	|	11	|	1	|	0	|	1	|	1	|	0	|		|
-562	|	561	|	species	|	EC	|	0	|	1	|	11	|	1	|	0	|	1	|	1	|	0	|		|
-561	|	543	|	genus	|		|	0	|	1	|	11	|	1	|	0	|	1	|	0	|	0	|		|
-543	|	91347	|	family	|		|	0	|	1	|	11	|	1	|	0	|	1	|	0	|	0	|		|
-91347	|	1236	|	order	|		|	0	|	1	|	11	|	1	|	0	|	1	|	0	|	0	|		|
-1236	|	1224	|	class	|		|	0	|	1	|	11	|	1	|	0	|	1	|	0	|	0	|		|
-1224	|	2	|	phylum	|		|	0	|	1	|	11	|	1	|	0	|	1	|	0	|	0	|		|
-2	|	131567	|	superkingdom	|		|	0	|	0	|	11	|	0	|	0	|	0	|	0	|	0	|		|
-131567	|	1	|	no rank	|		|	8	|	1	|	1	|	1	|	0	|	1	|	1	|	0	|		|
-1	|	1	|	no rank	|		|	8	|	0	|	1	|	0	|	0	|	0	|	0	|	0	|		|
--- a/tool-data/kraken_databases.loc.sample	Thu Sep 13 09:45:15 2018 -0400
+++ b/tool-data/kraken_databases.loc.sample	Wed Dec 04 06:52:03 2019 -0500
@@ -0,0 +1,18 @@
+# Expect three columns, tab separated, as follows:
+# - value (Galaxy records this in the Galaxy DB)
+# - name (Galaxy shows this in the UI)
+# - path (folder name containing the Kraken DB)
+#
+# e.g.
+# plants2018<tab>Plant genomes (2018)<tab>/path/to/krakenDB/plants_2018
+#
+# For backward compatibility the Kraken wrapper also accepts:
+# - value (Galaxy records this in the Galaxy DB)
+# - name (Final part of folder name, Galaxy shows this in the UI)
+# - path (Parent folder name, where $path/$name contains the Kraken DB)
+#
+# e.g.
+# plants2018<tab>plants_2018<tab>/path/to/krakenDB
+#
+# Please avoid this legacy usage, as not all tools using
+# the kraken_databases.loc file will understand it.