annotate get.oturep.xml @ 0:e6758f2f744b draft

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/mothur commit a9d1e0debcd357d8080a1c6c5f1d206dd45a7a4d

<tool profile="16.07" id="mothur_get_oturep" name="Get.oturep" version="@WRAPPER_VERSION@.0">
    <description>Generate a fasta with a representative sequence for each OTU</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements"/>
    <expand macro="stdio"/>
    <expand macro="version_command"/>
    <command><![CDATA[
        @SHELL_OPTIONS@

        ## create symlinks to input datasets
        ln -s "$otu" otu.dat &&
        #if $method_condition.method_switch == "distance":
            ln -s "$method_condition.matrix_condition.dist" dist.dat &&
            ln -s "$method_condition.matrix_condition.nameOrCount" nameOrCount.dat &&
            ln -s "$method_condition.fasta" fasta.dat &&
            #if $method_condition.pick.type == "yes":
                ln -s "$method_condition.pick.group" group.dat &&
            #end if
        #elif $method_condition.method_switch == "abundance":
            ln -s "$method_condition.nameOrCount" nameOrCount.dat &&
        #end if

        echo 'get.oturep(
            list=otu.dat,
            #if $sorted:
                sorted=$sorted,
            #end if
            #if $label:
                label=${ str($label).replace(",","-") },
            #end if
            #if $method_condition.method_switch == "distance":
                method=distance,
                #if $method_condition.matrix_condition.matrix_format == "column":
                    column=dist.dat,
                    #if $method_condition.matrix_condition.nameOrCount.is_of_type("mothur.names"):
                        name=nameOrCount.dat,
                    #elif $method_condition.matrix_condition.nameOrCount.is_of_type("mothur.count_table"):
                        count=nameOrCount.dat,
                    #end if
                #elif $method_condition.matrix_condition.matrix_format == "phylip":
                    phylip=dist.dat,
                    #if $method_condition.matrix_condition.nameOrCount:
                        #if $method_condition.matrix_condition.nameOrCount.is_of_type("mothur.names"):
                            name=nameOrCount.dat,
                        #elif $method_condition.matrix_condition.nameOrCount.is_of_type("mothur.count_table"):
                            count=nameOrCount.dat,
                        #end if
                    #end if
                #end if
                #if $method_condition.fasta:
                    fasta=fasta.dat,
                #end if
                #if $method_condition.pick.type == "yes":
                    #if $method_condition.pick.group:
                        group=group.dat,
                    #end if
                    #if $method_condition.pick.groups:
                        groups=${ str($method_condition.pick.groups).replace(",","-") },
                    #end if
                #end if
                large=$method_condition.large
            #elif $method_condition.method_switch == "abundance":
                method=abundance,
                #if $method_condition.nameOrCount.is_of_type("mothur.names"):
                    name=nameOrCount.dat
                #elif $method_condition.nameOrCount.is_of_type("mothur.count_table"):
                    count=nameOrCount.dat
                #end if
            #end if
        )'
        | sed 's/ //g'  ## mothur trips over whitespace
        | mothur
        | tee mothur.out.log
    ]]></command>
    <inputs>
        <param name="otu" type="data" format="mothur.list" label="list - OTU List"/>
        <conditional name="method_condition">
            <param name="method_switch" type="select" label="Method to use for the selection of the representative sequences" help="The distance method finds the sequence with the smallest maximum distance to other sequences. The abundance method chooses the most abundant sequence in the OTU as the representative.">
                <option value="distance" selected="true">distance</option>
                <option value="abundance">abundance</option>
            </param>
            <when value="distance">
                <conditional name="matrix_condition">
                    <param name="matrix_format" type="select" label="Select a Distance Matrix Format" help="">
                        <option value="column">Pairwise Column Distance Matrix</option>
                        <option value="phylip">Phylip Distance Matrix</option>
                    </param>
                    <when value="column">
                        <param name="dist" type="data" format="mothur.pair.dist" label="column - Distance Matrix"/>
                        <param name="nameOrCount" type="data" format="mothur.names,mothur.count_table" label="name file or count table - Sequences Name reference"/>
                    </when>
                    <when value="phylip">
                        <param name="dist" type="data" format="mothur.dist,mothur.lower.dist,mothur.square.dist" label="phylip - Distance Matrix"/>
                        <param name="nameOrCount" type="data" format="mothur.names,mothur.count_table" optional="true" label="name file or count table - Sequences Name reference"/>
                    </when>
                </conditional>
                <param name="fasta" type="data" format="fasta" optional="true" label="fasta - Fasta"/>
                <conditional name="pick">
                    <param name="type" type="select" label="Parse a group file into abundant and rare groups?" help="">
                        <option value="no" selected="true">No</option>
                        <option value="yes">Yes</option>
                    </param>
                    <when value="no"/>
                    <when value="yes">
                        <param name="group" type="data" format="mothur.groups" label="group - Group file for the OTU List"/>
                        <param name="groups" type="select"  multiple="true" label="groups - Group Selection (all used if none are selected)">
                            <options>
                                <filter type="data_meta" ref="group" key="groups"/>
                            </options>
                        </param>
                    </when>
                </conditional>
                <param name="large" type="boolean" checked="false" truevalue="true" falsevalue="false" label="large - Distance Matrix is very Large" help="Set this parameter to Yes if your distance matric might not fit in RAM"/>
            </when>
            <when value="abundance">
                <param name="nameOrCount" type="data" format="mothur.names,mothur.count_table" label="name file or count table - Sequences Name reference"/>
            </when>
        </conditional>
        <param name="label" type="select" multiple="true" label="label - OTU Labels">
            <expand macro="labeloptions"/>
        </param>
        <param name="sorted" type="select" label="sorted - Sort Sequences by">
            <option value="">Don't sort</option>
            <option value="name">Sequence Name</option>
            <option value="number">Bin Number</option>
            <option value="size">Bin Size</option>
            <option value="group">Group</option>
        </param>
    </inputs>
    <outputs>
        <expand macro="logfile-output"/>
        <collection name="names_out" type="list" label="${tool.name} on ${on_string}: rep.names">
            <filter>method_condition['nameOrCount'].ext == "mothur.names"</filter>
            <filter>method_condition['matrix_condition']['nameOrCount'].ext == "mothur.names"</filter>
            <discover_datasets pattern=".*?\.(?P&lt;designation&gt;.*)\.rep\.names" format="mothur.names"/>
        </collection>
        <collection name="fasta_out" type="list" label="${tool.name} on ${on_string}: rep.fasta">
            <filter>method_condition['method_switch'] == "distance"</filter>
            <filter>method_condition['fasta']</filter>
            <discover_datasets pattern=".*?\.(?P&lt;designation&gt;.*)\.rep\.fasta" format="fasta"/>
        </collection>
        <collection name="count_out" type="list" label="${tool.name} on ${on_string}: rep.count_table">
            <filter>method_condition['nameOrCount'].ext == "mothur.count_table"</filter>
            <filter>method_condition['matrix_condition']['nameOrCount'].ext == "mothur.count_table"</filter>
            <discover_datasets pattern=".*?\.(?P&lt;designation&gt;.*)\.rep\.count_table" format="mothur.count_table"/>
        </collection>
    </outputs>
    <tests>
        <!-- test with distance / phylip -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="distance"/>
            <param name="matrix_format" value="phylip"/>
            <param name="dist" value="98_sq_phylip_amazon.dist" ftype="mothur.square.dist"/>
            <param name="nameOrCount" value="amazon1.names" ftype="mothur.names"/>
            <output_collection name="names_out" count="36">
                <element name="0.27" md5="39ff2858909d49633871d5a625585de5" ftype="mothur.names"/>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
        <!-- test with distance / column / label select -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="distance"/>
            <param name="matrix_format" value="column"/>
            <param name="dist" value="amazon.pair.dist" ftype="mothur.pair.dist"/>
            <param name="nameOrCount" value="amazon1.names" ftype="mothur.names"/>
            <param name="label" value="unique,0.05,0.27,0.45"/>
            <output_collection name="names_out" count="4">
                <element name="0.27" md5="be6dccdd3d4619c3ac465246c826209a" ftype="mothur.names"/>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
        <!-- test with distance / fasta / sorting -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="distance"/>
            <param name="matrix_format" value="phylip"/>
            <param name="dist" value="98_sq_phylip_amazon.dist" ftype="mothur.square.dist"/>
            <param name="fasta" value="amazon.fasta" ftype="fasta"/>
            <param name="label" value="0.05,0.27"/>
            <param name="sorted" value="name"/>
            <output_collection name="names_out" count="2">
                <element name="0.27" md5="39ff2858909d49633871d5a625585de5" ftype="mothur.names"/>
            </output_collection>
            <output_collection name="fasta_out" count="2">
                <element name="0.27" md5="a9d9622b6f5fcbe8acdc05df5484ffc4" ftype="fasta"/>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
        <!-- test with distance / group file -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="distance"/>
            <param name="matrix_format" value="column"/>
            <param name="dist" value="amazon.pair.dist" ftype="mothur.pair.dist"/>
            <param name="nameOrCount" value="amazon1.names" ftype="mothur.names"/>
            <param name="label" value="unique,0.05,0.27,0.45"/>
            <param name="type" value="yes"/>
            <param name="group" value="amazon.groups" ftype="mothur.groups"/>
            <param name="groups" value="forest,pasture"/>
            <output_collection name="names_out" count="8">
                <element name="0.27.pasture" md5="34f87d93fd9ebbefed9754f627f502fd" ftype="mothur.names"/>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
        <!-- test with distance / count table -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="distance"/>
            <param name="matrix_format" value="phylip"/>
            <param name="dist" value="98_sq_phylip_amazon.dist" ftype="mothur.square.dist"/>
            <param name="label" value="unique,0.05,0.27"/>
            <param name="nameOrCount" value="amazon1.count_table" ftype="mothur.count_table"/>
            <output_collection name="count_out" count="3">
                <element name="0.27" md5="6b1afd712825af4bf8cb4ab8304df903" ftype="mothur.count_table"/>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
        <!-- test with abundance / count / label -->
        <!-- It seems that the content of the result file can vary from one execution to another with the same parameters -->
        <test>
            <param name="otu" value="amazon.an.list" ftype="mothur.list"/>
            <param name="method_switch" value="abundance"/>
            <param name="nameOrCount" value="amazon1.count_table" ftype="mothur.count_table"/>
            <param name="label" value="0.27,unique"/>
            <param name="sorted" value="name"/>
            <output_collection name="count_out" count="2">
                <element name="0.27" ftype="mothur.count_table">
                    <assert_contents>
                        <has_text_matching expression="U68\w+\t36" />
                    </assert_contents>
                </element>
            </output_collection>
            <expand macro="logfile-test"/>
        </test>
    </tests>
    <help>
<![CDATA[

@MOTHUR_OVERVIEW@


**Command Documentation**

The get.oturep_ command generates a fasta-formatted sequence file containing only a representative sequence for each OTU.  The opposite of the bin.seqs command.

.. _get.oturep: https://www.mothur.org/wiki/Get.oturep

v1.23.0: Updated to Mothur 1.33, added count and method parameter
]]>
    </help>
    <expand macro="citations"/>
</tool>