# HG changeset patch
# User iuc
# Date 1736716952 0
# Node ID f2fd2d1b96e3adf6b29913af86f75f4cc097ad52

planemo upload for repository https://github.com/bluenote-1577/fairy commit 54ed1c9745f1b74e1d0a0a8353fc47f9818708e8

diff -r 000000000000 -r f2fd2d1b96e3 fairy_cov.xml
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/fairy_cov.xml	Sun Jan 12 21:22:32 2025 +0000
@@ -0,0 +1,114 @@
+<tool id="fairy_cov" name="Fairy coverage" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
+    <description>Create coverage file for specific binners</description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+    <expand macro="requirements"/>
+    <command detect_errors="exit_code">
+        <![CDATA[
+
+            #import re
+
+            #set $file = re.sub('[^\s\w\-\\.]', '_', str($contig.element_identifier))
+            #set $bcsp = re.sub('[^\s\w\-\\.]', '_', str($bcsp_file.element_identifier))
+
+            ln -s '$contig' '$file' &&
+            ln -s '$bcsp_file' '$bcsp' &&
+
+            fairy coverage
+            '$file'
+            '$bcsp'
+            -t "\${GALAXY_SLOTS:-3}"
+            -m ${minimum_ani}
+            -M ${min_number_kmers}
+            -c ${c}
+            -k ${k}
+            --min-spacing ${min_spacing}
+            ${full_contig_name}
+            #if $output_type == 'semi':
+                --aemb-format
+            #end if
+            #if $output_type == 'max':
+                --maxbin-format
+            #end if
+            -o '$output'
+
+        ]]>
+    </command>
+    <inputs>
+        <param name="contig" type="data" format="fasta,fasta.gz" label="Input fasta contig file" help="Input the RAW FASTA contig file. It can be gzip!"/>
+        <param name="bcsp_file" type="data" format="bcsp" label="Input the pre-sketched file (.bcsp file)" help="This file will be generated with the fairy sketch tool."/>
+        <param argument="--minimum-ani" type="integer" optional="true" min="0" max="100" value="95" label="Set minimum ANI" help="Set the minimum adjusted ANI for the coverage calculation"/>
+        <param argument="--min-number-kmers" type="integer" value="8" optional="true" label="Genome filter" help="Filter out genomes with less then x k-mer sampled."/>
+        <param argument="-c" type="integer" value="50" optional="true" label="Set subsampling rate" help="This value does not interact with the .bcsp file which was used as input."/>
+        <param argument="-k" type="select" label="Select k-mer size" help="This value does not interact with the .bcsp file which was used as input.">
+            <option value="31">31</option>
+            <option value="21">21</option>
+        </param>
+        <param argument="--min-spacing" type="integer" value="30" label="Set spacing between k-mers" help=" Minimum spacing between selected k-mers on the contigs."/>
+        <param argument="--full-contig-name" type="boolean" falsevalue="" truevalue="--full-contig-name" label="Full contig name"
+            help ="When a contig has a space in there name this option allows to use the full name instead only the name till the first space"/>
+        <param name="output_type" type="select" label="Select for which binner the output should be generated">
+            <option value="meta">MetaBAT2</option>
+            <option value="semi">SemiBin2</option>
+            <option value="max">MaxBin2</option>
+        </param>
+    </inputs>
+    <outputs>
+        <data name="output" format="tabular" label="${tool.name} on ${on_string}"/>
+    </outputs>
+    <tests>
+        <test>
+            <param name="contig" value="single_test.fasta.gz" ftype="fasta.gz"/>
+            <param name="bcsp_file" value="single_test.fasta.gz.bcsp" ftype="bcsp"/>
+            <output name="output" value="normal_test.tsv"/>
+        </test>
+        <test>
+            <param name="contig" value="single_test.fasta.gz" ftype="fasta.gz"/>
+            <param name="bcsp_file" value="single_test.fasta.gz.bcsp" ftype="bcsp"/>
+            <param name="minimum-ani" value="99"/>
+            <param name="min-number-kmers" value="2"/>
+            <param name="full-contig-name" value="true"/>
+            <param name="output_type" value="semi"/>
+            <output name="output" value="test_2.tsv"/>
+        </test>
+        <test>
+            <param name="contig" value="single_test.fasta.gz" ftype="fasta.gz"/>
+            <param name="bcsp_file" value="single_test.fasta.gz.bcsp" ftype="bcsp"/>
+            <param name="k" value="21"/>
+            <param name="c" value="45"/>
+            <param name="min-spacing" value="10"/>
+            <param name="output_type" value="max"/>
+            <output name="output" value="test_3.tsv"/>
+        </test>
+    </tests>
+    <help>
+        <![CDATA[
+
+            Fairy computes multi-sample contig coverage for metagenome-assembled genome (MAG) binning.
+
+            Fairy is used after metagenomic assembly and before binning. It can
+
+            - Calculate coverage 100x-1000x faster than read alignment (e.g. BWA)
+            - Give comparable bins for multi-sample binning (short read or nanopore reads)
+            - Output formats that are compatible with MetaBAT2, MaxBin2, SemiBin2, and more
+            
+            Caveats:
+
+            - Don't use fairy for single-sample binning
+            - Don't use fairy for PacBio HiFi
+
+            For more information visit `the wiki site on GitHub <https://github.com/bluenote-1577/fairy/wiki/Introduction-to-fairy>`_.,
+
+            .. class:: infomark
+
+            Fairy usage for SemiBin2 is different than other tools: SemiBin2 requires separate coverage files for each read sample -- other tools require a single coverage matrix.
+
+            .. class:: infomark
+
+            The default output format from Fairy is the MetaBAT2 format. Any tool using this or the format from the other 2 binners work also with Fairys coverage files!
+        
+        ]]>
+    </help>
+    <expand macro="citations"/>
+</tool>
diff -r 000000000000 -r f2fd2d1b96e3 macros.xml
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/macros.xml	Sun Jan 12 21:22:32 2025 +0000
@@ -0,0 +1,17 @@
+<macros>
+    <xml name="requirements">
+        <requirements>
+            <requirement type="package" version="@TOOL_VERSION@">fairy</requirement>
+            <yield/>
+        </requirements>
+    </xml>
+    <token name="@TOOL_VERSION@">0.5.7</token>
+    <token name="@VERSION_SUFFIX@">0</token>
+    <token name="@PROFILE@">24.1</token>
+    <xml name="citations">
+        <citations>
+            <citation type="doi">10.1101/2024.04.23.590803</citation>
+            <yield/>
+        </citations>
+    </xml>
+</macros>
\ No newline at end of file
diff -r 000000000000 -r f2fd2d1b96e3 test-data/forward.paired.bcsp
Binary file test-data/forward.paired.bcsp has changed
diff -r 000000000000 -r f2fd2d1b96e3 test-data/normal_test.tsv
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/normal_test.tsv	Sun Jan 12 21:22:32 2025 +0000
@@ -0,0 +1,2 @@
+contigName	contigLen	totalAvgDepth	single_test.fasta.gz	single_test.fasta.gz-var
+NZ_CP017438.1	3123040	0.05509718146076748	0.05509718146076748	0.014618167653679848
diff -r 000000000000 -r f2fd2d1b96e3 test-data/single_test.fasta.gz
Binary file test-data/single_test.fasta.gz has changed
diff -r 000000000000 -r f2fd2d1b96e3 test-data/single_test.fasta.gz.bcsp
Binary file test-data/single_test.fasta.gz.bcsp has changed
diff -r 000000000000 -r f2fd2d1b96e3 test-data/test_2.tsv
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/test_2.tsv	Sun Jan 12 21:22:32 2025 +0000
@@ -0,0 +1,1 @@
+NZ_CP017438.1	0.05509718146076748
diff -r 000000000000 -r f2fd2d1b96e3 test-data/test_3.tsv
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/test-data/test_3.tsv	Sun Jan 12 21:22:32 2025 +0000
@@ -0,0 +1,2 @@
+contigName
+NZ_CP017438.1
diff -r 000000000000 -r f2fd2d1b96e3 test-data/test_paired_1.fq.gz
Binary file test-data/test_paired_1.fq.gz has changed
diff -r 000000000000 -r f2fd2d1b96e3 test-data/test_paired_1.fq.gz.paired.bcsp
Binary file test-data/test_paired_1.fq.gz.paired.bcsp has changed
diff -r 000000000000 -r f2fd2d1b96e3 test-data/test_paired_2.fq.gz
Binary file test-data/test_paired_2.fq.gz has changed