<tool id="tracy_assemble" name="tracy Assemble" version="@TOOL_VERSION@+galaxy0" profile="20.09">
    <description>genomic region from tiled, overlapping Sanger Chromatogram trace files</description>
    <macros>
        <import>macros.xml</import>
    </macros>    
    <expand macro="requirements" />
    <expand macro="version_command" />
    <command detect_errors="exit_code"><![CDATA[
        tracy assemble 
          #if $ref.useref == 'yes':
            --reference '$ref.reference'
            $ref.incref
          #end if
          --pratio $pratio 
          --trim $trim 
          --fracmatch $fracmatch
          --called $called
          --format $format
          $inccons
          --gapopen $alignment.gapopen
          --gapext $alignment.gapext
          --match $alignment.match
          --mismatch $alignment.mismatch
          #for tracefile in $tracefiles
            '$tracefile'
          #end for
          &&
          #if $format == 'fastq'
            mv out.cons.fq '$out_consensus'
          #else
            mv out.cons.fa '$out_consensus'
          #end if
    ]]></command>
    <inputs>
        <param name="tracefiles" type="data" format="ab1,scf" multiple="true" optional="false" label="Sanger chromatogram tracefiles to assemble" />
        <param argument="--pratio" type="float" value="0.33" label="Peak ratio to call base" />
        <param argument="--trim" type="integer" value="4" min="1" label="Trimming stringency" />
        <param argument="--fracmatch" type="float" value="0.5" min="0" max="1" label="Minimum fraction of matches [0:1]" />
        <param argument="--called" type="float" value="0.1" min="0" label="Fraction of traces required for consensus" />
        <param argument="--format" type="select" label="Output format">
            <option value="fasta" selected="true">FASTA</option>
            <option value="fastq">FASTQ</option>
        </param>
        <param argument="--inccons" type="boolean" truevalue="--inccons" falsevalue="" label="Include consensus in FASTA alignment" />
        <conditional name="ref" label="Reference guided assembly options">
            <param type="select" name="useref" label="Use reference to guide assembly">
                <option value="yes">Yes</option>
                <option value="no" selected="true">No</option>
            </param>
            <when value="yes">
                <param argument="--reference" type="data" format="fasta" label="FASTA or ABI format genome" />
                <param argument="--incref" type="boolean" truevalue="--incref" falsevalue="" label="Include reference in consensus" />
            </when>
            <when value="no">
            </when>
        </conditional>
        <expand macro="alignment_options" />
        <param name="json_output" type="boolean" label="Produce JSON output" />
    </inputs>
    <outputs>
        <data name="out_json" format="json" from_work_dir="out.json" label="tracy assemble JSON on ${on_string}">
            <filter>json_output</filter>
        </data>
        <data name="out_consensus" format="fasta" label="tracy assemble consensus on ${on_string}">
            <change_format>
                <when input="format" value="fasta" format="fasta" />
                <when input="format" value="fastq" format="fastq" />
            </change_format>
        </data>
        <data name="out_alignment" format="fasta" from_work_dir="out.align.fa" label="tracy assemble alignment on ${on_string}" />
    </outputs>
    <tests>
        <test expect_num_outputs="2">
            <param name="tracefiles" value="input1.ab1,input1.ab1" ftype="ab1" />
            <output name="out_consensus" value="out1.consensus.fasta" ftype="fasta" />    
            <output name="out_alignment" value="out1.alignment.fasta" ftype="fasta" lines_diff="4" />    
        </test>
        <test expect_num_outputs="2">
            <param name="tracefiles" value="input1.ab1,input1.ab1" ftype="ab1" />
            <conditional name="ref">
                <param name="useref" value="yes" />
                <param name="reference" value="reference1.fasta" />
            </conditional>
            <output name="out_consensus" value="out1.consensus.fasta" ftype="fasta" />    
            <output name="out_alignment" value="out2.alignment.fasta" ftype="fasta" lines_diff="4" />    
        </test>
        <test expect_num_outputs="3">
            <param name="tracefiles" value="input1.ab1,input1.ab1" ftype="ab1" />
            <param name="json_output" value="true" />
            <output name="out_json" ftype="json">
                <assert_contents>
                    <has_text text='"peakA": [5, 5, 5, 5, 5, 4' />
                </assert_contents>
            </output>
            <output name="out_consensus" value="out1.consensus.fasta" ftype="fasta" />    
            <output name="out_alignment" value="out1.alignment.fasta" ftype="fasta" lines_diff="4" />    
        </test>
    </tests>
    <help><![CDATA[
**What this does**

This tool assembles trace files that are from overlapping regions (using tracy_), optionally using a reference genome.
The assembled sequence is output either as FASTQ or FASTA.

@pratio@

Prior to assembling the traces, they are trimmed using an algorithm that automatically finds left and right trimming
positions. This algorithm can be tuned using the *trimming stringency* (higher values mean stricter trimmer).

Trace files have to match (either each other or, when using reference guided assembly, the reference sequence) 
at at least *fracmatch* positions (by default 10%) to be considered as part of the assembly.

Read more here_.

.. _tracy: https://github.com/gear-genomics/tracy
.. _here: https://www.gear-genomics.com/docs/tracy/cli/#trace-assembly

    ]]></help>
    <expand macro="citations" />
</tool>