From c142110046b4ed562e8240a65f5603640752dc2e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Bj=C3=B6rn=20Gr=C3=BCning?= <bjoern@gruenings.eu>
Date: Mon, 12 Aug 2024 09:06:10 +0200
Subject: [PATCH] udpate, fix pear (#6217)

* udpate, fix pear

* fix shed.yml

* Update tools/pear/pear.xml

* Apply suggestions from code review

Co-authored-by: Marius van den Beek <m.vandenbeek@gmail.com>

---------

Co-authored-by: Marius van den Beek <m.vandenbeek@gmail.com>
---
 tools/pear/.shed.yml |   4 +-
 tools/pear/pear.xml  | 116 ++++++++++++++++---------------------------
 2 files changed, 45 insertions(+), 75 deletions(-)

diff --git a/tools/pear/.shed.yml b/tools/pear/.shed.yml
index e9e0c12611e..2a504195af4 100644
--- a/tools/pear/.shed.yml
+++ b/tools/pear/.shed.yml
@@ -1,4 +1,3 @@
-# repository published to https://toolshed.g2.bx.psu.edu/repos/iuc/pear
 owner: iuc
 name: pear
 description: PEAR evaluates all possible paired-end read overlaps
@@ -7,3 +6,6 @@ long_description: |
   minimizing false-positive results. Together with a highly optimized implementation, it can merge millions of paired end reads within a couple of minutes on a 
   standard desktop computer. Repository-Maintainer: Björn Grüning
 remote_repository_url: https://github.com/galaxyproject/tools-iuc/tree/master/tools/pear
+homepage_url: https://cme.h-its.org/exelixis/web/software/pear/
+categories:
+  - Fastq Manipulation
diff --git a/tools/pear/pear.xml b/tools/pear/pear.xml
index 91152b7b1f9..fd132ae4252 100644
--- a/tools/pear/pear.xml
+++ b/tools/pear/pear.xml
@@ -1,19 +1,19 @@
-<tool id="iuc_pear" name="Pear" version="@TOOL_VERSION@.@VERSION_SUFFIX@" profile="20.09">
+<tool id="iuc_pear" name="Pear" version="@TOOL_VERSION@.@VERSION_SUFFIX@" profile="23.0">
     <description>Paired-End read merger</description>
     <macros>
         <token name="@TOOL_VERSION@">0.9.6</token>
-        <token name="@VERSION_SUFFIX@">2</token>
+        <token name="@VERSION_SUFFIX@">3</token>
         <xml name="format_action">
             <actions>
                 <conditional name="library.type">
                     <when value="paired">
                         <action type="format">
-                            <option type="from_param" name="library.forward" param_attribute="ext" />
+                            <option type="from_param" name="library.forward" param_attribute="ext"/>
                         </action>
                     </when>
                     <when value="paired_collection">
                         <action type="format">
-                            <option type="from_param" name="library.input_collection" param_attribute="forward.ext" />
+                            <option type="from_param" name="library.input_collection" param_attribute="forward.ext"/>
                         </action>
                     </when>
                 </conditional>
@@ -26,18 +26,12 @@
     <requirements>
         <requirement type="package" version="@TOOL_VERSION@">pear</requirement>
     </requirements>
-    <stdio>
-        <exit_code range="1:" />
-        <exit_code range=":-1" />
-        <regex match="Error:" />
-        <regex match="Exception:" />
-    </stdio>
-    <command>
+    <command detect_errors="aggressive">
 <![CDATA[
     pear
         #if str( $library.type ) == "paired":
-            -f "$library.forward"
-            -r "$library.reverse"
+            -f '$library.forward'
+            -r '$library.reverse'
             #if $library.forward.is_of_type( 'fastqillumina' ):
                 --phred-base 64
             #else:
@@ -45,8 +39,8 @@
             #end if
         #else
             ## prepare collection
-            -f "$library.input_collection.forward"
-            -r "$library.input_collection.reverse"
+            -f '$library.input_collection.forward'
+            -r '$library.input_collection.reverse'
             #if $library.input_collection.forward.is_of_type( 'fastqillumina' ):
                 --phred-base 64
             #else:
@@ -57,7 +51,7 @@
         --output pear
         --p-value $pvalue
         --min-overlap $min_overlap
-        #if int($max_assembly_length) > 0:
+        #if $max_assembly_length and int($max_assembly_length) > 0:
             --max-asm-length $max_assembly_length
         #end if
         --min-asm-length $min_assembly_length
@@ -66,7 +60,7 @@
         --max-uncalled-base $max_uncalled_base
         --test-method $test_method
         --empirical-freqs $empirical_freqs
-        -j "\${GALAXY_SLOTS:-8}"
+        --threads "\${GALAXY_SLOTS:-8}"
         --score-method $score_method
         --cap $cap
         $nbase
@@ -75,64 +69,38 @@
     <inputs>
         <conditional name="library">
             <param name="type" type="select" label="Dataset type">
-              <option value="paired">Paired-end</option>
-              <option value="paired_collection">Paired-end Dataset Collection</option>
+                <option value="paired">Paired-end</option>
+                <option value="paired_collection">Paired-end Dataset Collection</option>
             </param>
             <when value="paired">
-                <param name="forward" type="data" format="fastqillumina,fastqsanger"
-                    label="Name of file that contains the forward paired-end reads" help="-f" />
-                <param name="reverse" type="data" format="fastqillumina,fastqsanger"
-                    label="Name of file that contains the reverse paired-end reads" help="-r" />
+                <param name="forward" type="data" format="fastqillumina,fastqsanger" label="Name of file that contains the forward paired-end reads" help="-f"/>
+                <param name="reverse" type="data" format="fastqillumina,fastqsanger" label="Name of file that contains the reverse paired-end reads" help="-r"/>
             </when>
             <when value="paired_collection">
-                <param name="input_collection" format="fastqillumina,fastqsanger"
-                    type="data_collection" collection_type="paired"
-                    label="FASTQ Paired Dataset" help="Nucleotide-space: Must have PHRED-scaled quality values. (-f and -r)" />
+                <param name="input_collection" type="data_collection" format="fastqillumina,fastqsanger" label="FASTQ Paired Dataset" help="Nucleotide-space: Must have PHRED-scaled quality values. (-f and -r)" collection_type="paired"/>
             </when>
         </conditional>
-
         <!-- optional -->
-        <param name="pvalue" type="float" value="0.01" min="0" optional="true" max="1" label="Specify a p-value for the statistical test"
-            help="If the computed p-value of a possible assembly exceeds the specified p-value then the paired-end read will not be assembled. Setting 1.0 disables the test. (--p-value)" />
-
-        <param name="min_overlap" type="integer" value="10" optional="true" label="Minimum overlap size"
-            help="The minimum overlap may be set to 1 when the statistical test is used. However, further restricting the minimum overlap size to a proper value may reduce false-positive assembles. (--min-overlap)" />
-
-        <param name="max_assembly_length" type="integer" value="0" optional="true" label="Maximum possible length of the assembled sequences"
-            help="Setting this value to 0 disables the restriction and assembled sequences may be arbitrary long. (--max-assembly-length)" />
-
-        <param name="min_assembly_length" type="integer" value="50" optional="true" label="Minimum possible length of the assembled sequences"
-            help="Setting this value to 0 disables the restriction and assembled sequences may be arbitrary short. (--min-assembly-length)" />
-
-        <param name="min_trim_length" type="integer" value="1" optional="true" label="Minimum length of reads after trimming the low quality part"
-            help="See option -q. (--min-trim-length)" />
-
-        <param name="quality_threshold" type="integer" value="0" optional="true" label="Quality score threshold for trimming the low quality part of a read"
-                help="If the quality scores of two consecutive bases are strictly less than the specified threshold, the rest of the read will be trimmed. (--quality-threshold)" />
-
-        <param name="max_uncalled_base" type="float" value="1.0" min="0" optional="true" max="1" label="Maximal proportion of uncalled bases in a read"
-            help="Setting this value to 0 will cause PEAR to discard all reads containing uncalled bases. The other extreme setting is 1 which causes PEAR to process all reads independent on the number of uncalled bases. (--max-uncalled-base)" />
-
-        <param name="cap" type="integer" value="40" optional="true" label="Specify  the upper bound for the resulting quality score"
-            help="If set to zero, capping is disabled. (--cap)" />
-
+        <param name="pvalue" argument="--p-value" type="float" min="0" max="1" value="0.01" label="Specify a p-value for the statistical test" help="If the computed p-value of a possible assembly exceeds the specified p-value then the paired-end read will not be assembled. Setting 1.0 disables the test. (--p-value)"/>
+        <param argument="--min-overlap" type="integer" value="10" min="0" label="Minimum overlap size" help="The minimum overlap may be set to 1 when the statistical test is used. However, further restricting the minimum overlap size to a proper value may reduce false-positive assembles. (--min-overlap)"/>
+        <param argument="--max-assembly-length" min="0" type="integer" value="0" label="Maximum possible length of the assembled sequences" help="Setting this value to 0 disables the restriction and assembled sequences may be arbitrary long. (--max-assembly-length)"/>
+        <param argument="--min-assembly-length" min="0" type="integer" value="50" label="Minimum possible length of the assembled sequences" help="Setting this value to 0 disables the restriction and assembled sequences may be arbitrary short."/>
+        <param argument="--min-trim-length" min="0" type="integer" value="1" label="Minimum length of reads after trimming the low quality part" help="See option --quality-threshold."/>
+        <param argument="--quality-threshold" type="integer" value="0" label="Quality score threshold for trimming the low quality part of a read" help="If the quality scores of two consecutive bases are strictly less than the specified threshold, the rest of the read will be trimmed."/>
+        <param argument="--max-uncalled-base" type="float" min="0" max="1" value="1.0" label="Maximal proportion of uncalled bases in a read" help="Setting this value to 0 will cause PEAR to discard all reads containing uncalled bases. The other extreme setting is 1 which causes PEAR to process all reads independent on the number of uncalled bases."/>
+        <param argument="--cap" type="integer" value="40" min="0" label="Specify  the upper bound for the resulting quality score" help="If set to zero, capping is disabled."/>
         <param name="test_method" type="select" label="Type of statistical test" help="(--test-method)">
             <option value="1" selected="true">Given the minimum allowed overlap, test using the highest OES (1)</option>
             <option value="2">Use the acceptance probability (2)</option>
         </param>
-
-        <param name="empirical_freqs" type="boolean" truevalue="-e" falsevalue="" checked="false"
-            label="Disable empirical base frequencies" help="(--empirical-freqs)" />
-        <param name="nbase" type="boolean" truevalue="--nbase" falsevalue="" checked="false"
-            label="Use N base if uncertain" help="When  merging a base-pair that consists of two non-equal bases out of which none is degenerate, set the merged base to N and use the highest quality score of the two bases. (--nbase)" />
-
+        <param name="empirical_freqs" type="boolean" truevalue="-e" falsevalue="" checked="false" label="Disable empirical base frequencies" help="(--empirical-freqs)"/>
+        <param name="nbase" type="boolean" truevalue="--nbase" falsevalue="" checked="false" label="Use N base if uncertain" help="When  merging a base-pair that consists of two non-equal bases out of which none is degenerate, set the merged base to N and use the highest quality score of the two bases. (--nbase)"/>
         <param name="score_method" type="select" label="Scoring method" help="(--score-method)">
             <option value="1">OES with +1 for match and -1 for mismatch</option>
             <option value="2" selected="true">Assembly score (AS) use +1 for match and -1 for mismatch multiplied by base quality scores</option>
             <option value="3">Ignore quality scores and use +1 for a match and -1 for a mismatch</option>
         </param>
-
-        <param name="outputs" type="select" optional="false" multiple="true" label="Output files">
+        <param name="outputs" type="select" optional="false" label="Output files" multiple="true">
             <option value="assembled" selected="true">Assembled reads</option>
             <option value="unassembled_forward">Forward unassembled reads</option>
             <option value="unassembled_reverse">Reverse unassembled reads</option>
@@ -161,13 +129,13 @@
         <test expect_num_outputs="2">
             <conditional name="library">
                 <param name="type" value="paired"/>
-                <param name="forward" value="forward.fastq" ftype="fastqsanger" />
-                <param name="reverse" value="reverse.fastq" ftype="fastqsanger" />
+                <param name="forward" value="forward.fastq" ftype="fastqsanger"/>
+                <param name="reverse" value="reverse.fastq" ftype="fastqsanger"/>
             </conditional>
-            <param name="min_overlap" value="10" />
-            <param name="min_assembly_length" value="50" />
-            <param name="cap" value="0" />
-            <param name="outputs" value="assembled,unassembled_forward" />
+            <param name="min_overlap" value="10"/>
+            <param name="min_assembly_length" value="50"/>
+            <param name="cap" value="0"/>
+            <param name="outputs" value="assembled,unassembled_forward"/>
             <output name="assembled_reads" file="pear_assembled_results1.fastq" ftype="fastqsanger"/>
             <output name="unassembled_forward_reads" file="pear_unassembled_forward_results1.fastq" ftype="fastqsanger"/>
         </test>
@@ -178,15 +146,15 @@
                 <param name="type" value="paired_collection"/>
                 <param name="input_collection">
                     <collection type="paired">
-                        <element name="forward" value="forward.fastq" ftype="fastqillumina" />
-                        <element name="reverse" value="reverse.fastq" ftype="fastqillumina" />
+                        <element name="forward" value="forward.fastq" ftype="fastqillumina"/>
+                        <element name="reverse" value="reverse.fastq" ftype="fastqillumina"/>
                     </collection>
                 </param>
             </conditional>
-            <param name="min_overlap" value="10" />
-            <param name="min_assembly_length" value="50" />
-            <param name="cap" value="0" />
-            <param name="outputs" value="assembled,unassembled_forward" />
+            <param name="min_overlap" value="10"/>
+            <param name="min_assembly_length" value="50"/>
+            <param name="cap" value="0"/>
+            <param name="outputs" value="assembled,unassembled_forward"/>
             <output name="assembled_reads" ftype="fastqillumina">
                 <assert_contents>
                     <has_line_matching expression="@.*"/>
@@ -222,7 +190,7 @@ and that commercial partners should obtain a license.
 
 ]]>
   </help>
-  <citations>
-      <citation type="doi">10.1093/bioinformatics/btt593</citation>
-  </citations>
+    <citations>
+        <citation type="doi">10.1093/bioinformatics/btt593</citation>
+    </citations>
 </tool>