Mercurial > repos > iuc > purge_dups

changeset 6:bded6226d503 draft default tip

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/main/tools/purge_dups commit eb64803a7b418d1a02a508920f63449f1669f073
author iuc
date Sun, 25 Jan 2026 10:22:07 +0000
parents e9bd16ba5ebd
children
files macros.xml purge_dups.xml test-data/assembly_test.fasta.gz test-data/get_seqs_hap_10.fa.gz test-data/get_seqs_purged_10.fa.gz test-data/split_04.fasta.gz test-data/split_out.fasta.gz
diffstat 7 files changed, 52 insertions(+), 18 deletions(-) [+]
line wrap: on
line diff
--- a/macros.xml	Mon Jun 06 07:49:11 2022 +0000
+++ b/macros.xml	Sun Jan 25 10:22:07 2026 +0000
@@ -1,6 +1,6 @@
 <macros>
     <token name="@TOOL_VERSION@">1.2.6</token>
-    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@VERSION_SUFFIX@">1</token>
     <xml name="xrefs">
         <xrefs>
             <xref type="bio.tools">purge_dups</xref>
--- a/purge_dups.xml	Mon Jun 06 07:49:11 2022 +0000
+++ b/purge_dups.xml	Sun Jan 25 10:22:07 2026 +0000
@@ -38,7 +38,12 @@
             > dups.bed 2> purge_dups.log
         #else if $function_select.functions == 'split_fa':
             split_fa
-            '${function_select.input}' > split.fasta
+            '${function_select.input}'
+            #if $function_select.input.is_of_type('fasta.gz'):
+                | gzip > split.fasta.gz
+            #else:
+                > split.fasta
+            #end if
         #else if $function_select.functions == 'pbcstat':
             #for $i, $file in enumerate($function_select.input):
                 #if $file.is_of_type('paf'):
@@ -87,6 +92,9 @@
             -m $function_select.advanced_options.min_ratio
             -g $function_select.advanced_options.min_gap
             '${function_select.bed_input}' '${function_select.fasta_input}'
+            #if $function_select.fasta_input.is_of_type('fasta.gz'):
+                && gzip hap.fa && gzip purged.fa
+            #end if
         #end if
     ]]></command>
     <inputs>
@@ -197,7 +205,7 @@
                 <expand macro="calcuts" />
             </when>
             <when value="get_seqs">
-                <param name="fasta_input" type="data" format="fasta" label="Assembly FASTA file"/>
+                <param name="fasta_input" type="data" format="fasta,fasta.gz" label="Assembly FASTA file"/>
                 <param name="bed_input" type="data" format="bed" label="BED input file" help="Generated by the 'purge_dups' function mode."/>
                 <section name="advanced_options" title="Advanced options">
                     <param argument="-c" name="coverage" type="boolean" 
@@ -218,14 +226,14 @@
     </inputs>
     <outputs>
         <!-- Get Seqs -->
-        <data name="get_seqs_hap" format="fasta" from_work_dir="hap.fa" label="${tool.name} on ${on_string}: get_seqs haplotype" >
+        <data name="get_seqs_hap" format_source="function_select|fasta_input" from_work_dir="hap.fa*" label="${tool.name} on ${on_string}: get_seqs haplotype" >
             <filter>function_select['functions'] == 'get_seqs'</filter>
         </data>
-        <data name="get_seqs_purged" format="fasta" from_work_dir="purged.fa" label="${tool.name} on ${on_string}: get_seqs purged sequences">
+        <data name="get_seqs_purged" format_source="function_select|fasta_input" from_work_dir="purged.fa*" label="${tool.name} on ${on_string}: get_seqs purged sequences">
             <filter>function_select['functions'] == 'get_seqs'</filter>
         </data>
         <!-- Split FA -->
-        <data name="split_fasta" format="fasta" from_work_dir="split.fasta" label="${tool.name} on ${on_string}: split FASTA">
+        <data name="split_fasta" format_source="function_select|input" from_work_dir="split.fasta*" label="${tool.name} on ${on_string}: split FASTA">
             <filter>function_select['functions'] == 'split_fa'</filter>
         </data>
         <!-- Ngscstat -->
@@ -332,6 +340,14 @@
             </conditional>
             <output name="split_fasta" value="split_04.fasta" ftype="fasta"/>
         </test>
+        <!-- Test 4b Split fa gzip -->
+        <test expect_num_outputs="1">
+            <conditional name="function_select">
+                <param name="functions" value="split_fa"/>
+                <param name="input" value="assembly_test.fasta.gz" ftype="fasta.gz"/>
+            </conditional>
+            <output name="split_fasta" value="split_04.fasta.gz" ftype="fasta.gz" decompress="true"/>
+        </test>
         <!-- Test 5 pbcstat -->
         <test expect_num_outputs="6">
             <conditional name="function_select">
@@ -342,7 +358,7 @@
                     <param name="min_map_ratio" value="0.01"/>
                     <param name="min_map_qual" value="1"/>
                     <param name="flank" value="1"/>
-                    <param name="primary_alignments" value="-p"/>
+                    <param name="primary_alignments" value="true"/>
                 </section>
                 <section name="section_calcuts">
                     <param name="min_depth" value="0.01"/>
@@ -370,7 +386,7 @@
                     <param name="min_map_ratio" value="0.01"/>
                     <param name="min_map_qual" value="1"/>
                     <param name="flank" value="1"/>
-                    <param name="primary_alignments" value="-p"/>
+                    <param name="primary_alignments" value="true"/>
                 </section>
                 <section name="section_calcuts">
                     <param name="min_depth" value="0.01"/>
@@ -379,8 +395,8 @@
                     <param name="upper_depth" value="100"/>
                     <param name="ploidy" value="-d 0"/>
                 </section>
+                <param name="output_options" value="pbcstat_coverage,calcuts_cutoff"/>
             </conditional>
-            <param name="output_options" value="pbcstat_coverage,calcuts_cutoff"/>
             <output name="calcuts_cutoff" value="calcuts_cutoff_06.tabular" ftype="tabular"/>
             <output name="pbcstat_cov" value="pbcstat_cov_06.tabular" ftype="tabular"/>
         </test>
@@ -396,8 +412,8 @@
                     <param name="upper_depth" value="100"/>
                     <param name="ploidy" value="-d 0"/>
                 </section>
+                <param name="output_options" value="pbcstat_coverage,calcuts_cutoff"/>
             </conditional>
-            <param name="output_options" value="pbcstat_coverage,calcuts_cutoff"/>
             <output name="calcuts_cutoff" value="calcuts_cutoff_07.tabular" ftype="tabular"/>
             <output name="pbcstat_cov" value="pbcstat_cov_07.tabular" ftype="tabular"/>
         </test>
@@ -417,8 +433,8 @@
                     <param name="upper_depth" value="100"/>
                     <param name="ploidy" value="-d 0"/>
                 </section>
+                <param name="output_options" value="ngscstat_coverage,calcuts_cutoff"/>
             </conditional>
-            <param name="output_options" value="ngscstat_coverage,calcuts_cutoff"/>
             <output name="calcuts_cutoff" value="calcuts_cutoff_08.tabular" ftype="tabular"/>
             <output name="ngscstat_cov" value="ngsc_cov_08.tabular" ftype="tabular"/>
         </test>
@@ -445,18 +461,37 @@
                 <param name="fasta_input" value="split_out.fasta"/>
                 <param name="bed_input" value="dups.bed"/>
                 <section name="advanced_options">
-                    <param name="coverage" value="-c"/>
+                    <param name="coverage" value="true"/>
                     <param name="length" value="10"/>
-                    <param name="haplotigs" value="-a"/>
+                    <param name="haplotigs" value="true"/>
                     <param name="min_ratio" value=".01"/>
-                    <param name="end_trim" value="-e"/>
-                    <param name="split" value="-s"/>
+                    <param name="end_trim" value="true"/>
+                    <param name="split" value="true"/>
                     <param name="min_gap" value="100000"/>
                 </section>
             </conditional>
             <output name="get_seqs_purged" value="get_seqs_purged_10.fa" ftype="fasta"/>
             <output name="get_seqs_hap" value="get_seqs_hap_10.fa" ftype="fasta"/>
         </test>
+        <!-- Test 10b Get seqs gzip -->
+        <test expect_num_outputs="2">
+            <conditional name="function_select">
+                <param name="functions" value="get_seqs"/>
+                <param name="fasta_input" value="split_out.fasta.gz" ftype="fasta.gz"/>
+                <param name="bed_input" value="dups.bed"/>
+                <section name="advanced_options">
+                    <param name="coverage" value="true"/>
+                    <param name="length" value="10"/>
+                    <param name="haplotigs" value="true"/>
+                    <param name="min_ratio" value=".01"/>
+                    <param name="end_trim" value="true"/>
+                    <param name="split" value="true"/>
+                    <param name="min_gap" value="100000"/>
+                </section>
+            </conditional>
+            <output name="get_seqs_purged" value="get_seqs_purged_10.fa.gz" ftype="fasta.gz" decompress="true"/>
+            <output name="get_seqs_hap" value="get_seqs_hap_10.fa.gz" ftype="fasta.gz" decompress="true"/>
+        </test>
         <!-- Test 11 pbcstat histogram options-->
         <test expect_num_outputs="1">
             <conditional name="function_select">
@@ -467,7 +502,7 @@
                     <param name="min_map_ratio" value="0.01"/>
                     <param name="min_map_qual" value="1"/>
                     <param name="flank" value="1"/>
-                    <param name="primary_alignments" value="-p"/>
+                    <param name="primary_alignments" value="true"/>
                 </section>
                 <section name="section_calcuts">
                     <param name="min_depth" value="0.01"/>
@@ -479,10 +514,9 @@
                 <section name="section_hist">
                     <param name="ymax" value="100"/>
                     <param name="xmax" value="100"/>
-                    <param name="cutoffs_his" value="calcuts_out.tsv"/>
                 </section>
+                <param name="output_options" value="histogram"/>
             </conditional>
-            <param name="output_options" value="histogram"/>
             <output name="hist" value="hist_11.png" ftype="png" compare="sim_size"/>
         </test>
     </tests>
Binary file test-data/assembly_test.fasta.gz has changed
Binary file test-data/get_seqs_hap_10.fa.gz has changed
Binary file test-data/get_seqs_purged_10.fa.gz has changed
Binary file test-data/split_04.fasta.gz has changed
Binary file test-data/split_out.fasta.gz has changed