<tool id="seqkit_fx2tab" name="SeqKit fx2tab" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
    <description>convert FASTA/Q to tabular</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="bio_tools"/>
    <expand macro="requirements"/>
    <command detect_errors="exit_code"><![CDATA[
#import re

#set input_identifier = re.sub('[^\s\w\-]', '_', str($input.element_identifier))
ln -s '${input}' '${input_identifier}' &&

seqkit fx2tab 
'${input_identifier}'
$alphabet
$avg_qual
$gc
$gc_skew
$header_line
$length
$name
$no_qual
$only_id
#if str($input.ext) in ['fastqsanger', 'fastqsanger.gz']:
    --qual-ascii-base 33
#end if
$seq_hash
> '$output'
]]></command>
    <inputs>
        <param name="input" type="data" format="fasta,fasta.gz,fastqsanger,fastqsanger.gz" label="Input file"/>
        <param argument="--alphabet" type="boolean" truevalue="--alphabet" falsevalue="" checked="false" label="Output alphabet letters?"/>
        <param argument="--avg_qual" type="boolean" truevalue="--avg-qual" falsevalue="" checked="false" label="Output average quality of the read?"/>
        <param argument="--gc" type="boolean" truevalue="--gc" falsevalue="" checked="false" label="Output GC content?"/>
        <param argument="--gc_skew" type="boolean" truevalue="--gc-skew" falsevalue="" checked="false" label="Output GC-Skew?"/>
        <param argument="--header_line" type="boolean" truevalue="--header-line" falsevalue="" checked="false" label="Output header line?"/>
        <param argument="--length" type="boolean" truevalue="--length" falsevalue="" checked="false" label="Output sequence length?"/>
        <param argument="--name" type="boolean" truevalue="--name" falsevalue="" checked="false" label="Output only names (no sequences and qualities)?"/>
        <param argument="--no_qual" type="boolean" truevalue="--no-qual" falsevalue="" checked="false" label="Eliminate qualities from output?"/>
        <param argument="--only_id" type="boolean" truevalue="--only-id" falsevalue="" checked="false" label="Output id instead of full head?"/>
        <param argument="--seq_hash" type="boolean" truevalue="--seq-hash" falsevalue="" checked="false" label="Output md5 hash of sequence?"/>
    </inputs>
    <outputs>
        <data name="output" format="tabular"/>
    </outputs>
    <tests>
        <test expect_num_outputs="1">
            <param name="input" value="input1.fastq.gz" ftype="fastqsanger.gz"/>
            <output name="output" file="fx2tab_output1.tabular" ftype="tabular"/>
        </test>
        <test expect_num_outputs="1">
            <param name="input" value="input1.fastq.gz" ftype="fastqsanger.gz"/>
            <param name="avg_qual" value="--avg-qual"/>
            <param name="header_line" value="--header-line"/>
            <param name="length" value="--length"/>
            <param name="name" value="--name"/>
            <param name="only_id" value="--only-id"/>
            <output name="output" file="fx2tab_output2.tabular" ftype="tabular"/>
        </test>
        <test expect_num_outputs="1">
            <param name="input" value="input1.fasta.gz" ftype="fasta.gz"/>
            <output name="output" file="fx2tab_output3.tabular" ftype="tabular"/>
        </test>
        <test expect_num_outputs="1">
            <param name="input" value="input1.fasta.gz" ftype="fasta.gz"/>
            <param name="alphabet" value="--alphabet"/>
            <param name="avg_qual" value="--avg-qual"/>
            <param name="gc" value="--gc"/>
            <param name="gc_skew" value="--gc-skew"/>
            <param name="header_line" value="--header-line"/>
            <param name="length" value="--length"/>
            <param name="name" value="--name"/>
            <param name="no_qual" value="--no-qual"/>
            <param name="only_id" value="--only-id"/>
            <param name="seq_hash" value="--seq-hash"/>
            <output name="output" file="fx2tab_output4.tabular" ftype="tabular"/>
        </test>
    </tests>
    <help>
**What it does**

Converts FASTA/Q to tabular format and adds columns for sequence length and GC content/GC skew.  The output
consists of 3 columns; id, sequence and quality unless one of the options to output only names or to eliminate
qualities is selected or additional columns (e.g., md5 sequence hash) are requested.
    </help>
    <expand macro="citations"/>
</tool>