comparison cutadapt.xml @ 16:8f1df515eb87 draft

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/cutadapt commit 034c93600ce3279644d05f3675b4820ef7180851

15:660cffd8d92a

<tool id="cutadapt" name="Cutadapt" version="1.16.3" profile="17.09">
    <description>Remove adapter sequences from Fastq/Fasta</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <requirements>

    ln -f -s '${input_2}' '$read2' &&
#end if

## Run Cutadapt

cutadapt

## cutadapt (up to version 1.16) can't be run in multicore mode with these options
#if not any(($output_options.info_file, $output_options.rest_file, $output_options.wildcard_file, $output_options.too_short_file, $output_options.too_long_file, $output_options.untrimmed_file))
     -j \${GALAXY_SLOTS:-1}
#end if

#if str( $library.type ) == "single":
    @read1_options@
    --output='$out1'
#else:
    @read1_options@
    @read2_options@
    --output='$out1'
    --paired-output='$out2'

--error-rate=$adapter_options.error_rate
--times=$adapter_options.count
--overlap=$adapter_options.overlap
$adapter_options.no_indels
$adapter_options.match_read_wildcards


$filter_options.discard
$filter_options.discard_untrimmed
$filter_options.no_trim
$filter_options.mask_adapter
#if str($filter_options.min) != '0':
    --minimum-length=$filter_options.min
#end if
#if str($filter_options.max) != '0':
    --maximum-length=$filter_options.max

        <section name="adapter_options" title="Adapter Options">
            <param name="error_rate" argument="--error-rate" type="float" min="0" max="1" value="0.1" label="Maximum error rate" help="Maximum allowed error rate (no. of errors divided by the length of the matching region)." />
            <param name="no_indels" argument="--no-indels" type="boolean" value="False" truevalue="--no-indels" falsevalue="" label="Do not allow indels (Use ONLY with anchored 5' (front) adapters)." help="Do not allow indels in the alignments. That is, allow only mismatches. This option is currently only supported for anchored 5' adapters ('^ADAPTER') (default: both mismatches and indels are allowed)." />
            <param name="count" argument="--times" type="integer" min="1" value="1" label="Match times" help="Try to remove adapters at most COUNT times. Useful when an adapter gets appended multiple times." />
            <param name="overlap" argument="--overlap" type="integer" min="1" value="3" label="Minimum overlap length" help="Minimum overlap length. If the overlap between the adapter and the sequence is shorter than LENGTH, the read is not modified. This reduces the number of bases trimmed purely due to short random adapter matches." />
            <param name="match_read_wildcards" argument="--match-read-wildcards" type="boolean" value="False" truevalue="--match-read-wildcards" falsevalue="" label="Match Read Wildcards" help="Allow 'N's in the read as matches to the adapter." />
        </section>

        <!-- Filter Options -->
        <section name="filter_options" title="Filter Options">
            <param name="discard" argument="--discard-trimmed" type="boolean" value="False" truevalue="--discard" falsevalue="" label="Discard Trimmed Reads" help="Discard reads that contain the adapter instead of trimming them. Use the 'Minimum overlap length' option in order to avoid throwing away too many randomly matching reads!" />
            <param name="discard_untrimmed" argument="--discard_untrimmed" type="boolean" value="False" truevalue="--discard-untrimmed" falsevalue="" label="Discard Untrimmed Reads" help="Discard reads that do not contain the adapter." />
            <param name="min" argument="--minimum-length" type="integer" min="0" optional="True" value="0" label="Minimum length" help="Discard trimmed reads that are shorter than LENGTH.  Reads that are too short even before adapter removal are also discarded. In colorspace, an initial primer is not counted. Value of 0 means no minimum length." />
            <param name="max" argument="--maximum-length" type="integer" min="0" optional="True" value="0" label="Maximum length" help="Discard trimmed reads that are longer than LENGTH.  Reads that are too long even before adapter removal are also discarded. In colorspace, an initial primer is not counted. Value of 0 means no maximum length." />
            <param name="no_trim" argument="--no-trim" type="boolean" value="False" truevalue="--no-trim" falsevalue="" label="Do not trim adapters" help="Match and redirect reads to output/untrimmed-output as usual, but don't remove the adapters (default: trim the adapters)." />
            <param name="mask_adapter" argument="--mask-adapter" type="boolean" value="False" truevalue="--mask-adapter" falsevalue="" label="Mask Adapters" help="Mask adapter bases with 'N' instead of trimming them (default: trim adapters)." />
            <param name="max_n" argument="--max-n" type="float" min="0" optional="True" label="Max N" help="Discard reads with more than this number of 'N' bases. A number between 0 and 1 is interpreted as a fraction of the read length." />
            <param name="pair_filter" argument="--pair-filter" type="select" optional="True" label="Pair filter" help="Which of the reads in a paired-end read have to match the filtering criterion in order for the pair to be filtered. Default: any">
                <option value="any" selected="True">any</option>
                <option value="both">both</option>
            </param>

            <param name="rest_file" argument="--rest-file" type="boolean" value="False" label="Rest of Read" help="When the adapter matches in the middle of a read, write the rest (after the adapter) into a file."/>
            <param name="wildcard_file" argument="--wildcard-file" type="boolean" value="False" label="Wildcard File" help="When the adapter has wildcard bases ('N's) write adapter bases matching wildcard positions to file."/>
            <param name="too_short_file" argument="--too-short-output" type="boolean" value="False" label="Too Short Reads" help="Write reads that are too short (according to minimum length specified) to a file. (default: discard reads)"/>
            <param name="too_long_file" argument="--too-long-output" type="boolean" value="False" label="Too Long Reads" help="Write reads that are too long (according to maximum length specified) to a file. (default: discard reads)"/>
            <param name="untrimmed_file" argument="--untrimmed-output" type="boolean" value="False" label="Untrimmed Reads" help="Write reads that do not contain the adapter to a separate file, instead of writing them to the regular output file.  (default: output to same file as trimmed)"/>
        </section>

    </inputs>

    <outputs>

        <data name="too_long_paired_output" format="fastqsanger" metadata_source="input_2" label="${tool.name} on ${on_string}: Too Long Read 2" >
            <filter>(library['type'] == 'paired' or library['type'] == 'paired_collection')</filter>
            <filter>(output_options['too_long_file'] is True)</filter>
            <expand macro="inherit_format_2" />
        </data>
    </outputs>

    <tests>
        <!-- Ensure fastq works -->
        <test>

                    <has_text text="Summary"/>
                </assert_contents>
            </output>
            <output name="info_file" value="cutadapt_info_out.txt" ftype="txt"/>
        </test>
    </tests>

    <help><![CDATA[

.. class:: infomark

16:8f1df515eb87

<tool id="cutadapt" name="Cutadapt" version="1.16.4" profile="17.09">
    <description>Remove adapter sequences from Fastq/Fasta</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <requirements>

    ln -f -s '${input_2}' '$read2' &&
#end if

## Run Cutadapt

#if $output_options.multiple_output:
    mkdir split &&
#end if

cutadapt

## cutadapt (up to version 1.16) can't be run in multicore mode with these options
#if not any(($output_options.info_file, $output_options.rest_file, $output_options.wildcard_file, $output_options.too_short_file, $output_options.too_long_file, $output_options.untrimmed_file))
     -j \${GALAXY_SLOTS:-1}
#end if

#if str( $library.type ) == "single":
    @read1_options@
    #if $output_options.multiple_output:
        --output='split/{name}.fastq'
    #else:
        --output='$out1'
    #end if
#else:
    @read1_options@
    @read2_options@
    --output='$out1'
    --paired-output='$out2'

--error-rate=$adapter_options.error_rate
--times=$adapter_options.count
--overlap=$adapter_options.overlap
$adapter_options.no_indels
$adapter_options.match_read_wildcards
$adapter_options.no_trim
$adapter_options.mask_adapter

$filter_options.discard
$filter_options.discard_untrimmed
#if str($filter_options.min) != '0':
    --minimum-length=$filter_options.min
#end if
#if str($filter_options.max) != '0':
    --maximum-length=$filter_options.max

        <section name="adapter_options" title="Adapter Options">
            <param name="error_rate" argument="--error-rate" type="float" min="0" max="1" value="0.1" label="Maximum error rate" help="Maximum allowed error rate (no. of errors divided by the length of the matching region)." />
            <param name="no_indels" argument="--no-indels" type="boolean" value="False" truevalue="--no-indels" falsevalue="" label="Do not allow indels (Use ONLY with anchored 5' (front) adapters)." help="Do not allow indels in the alignments. That is, allow only mismatches. This option is currently only supported for anchored 5' adapters ('^ADAPTER') (default: both mismatches and indels are allowed)." />
            <param name="count" argument="--times" type="integer" min="1" value="1" label="Match times" help="Try to remove adapters at most COUNT times. Useful when an adapter gets appended multiple times." />
            <param name="overlap" argument="--overlap" type="integer" min="1" value="3" label="Minimum overlap length" help="Minimum overlap length. If the overlap between the adapter and the sequence is shorter than LENGTH, the read is not modified. This reduces the number of bases trimmed purely due to short random adapter matches." />
            <param name="match_read_wildcards" type="select"  label="Match Wildcards" help="Allow 'N's as matches. Default: In the adapters but not in the reads">
                <option value=" " selected="True">In the adapters but not in the reads</option>
                <option value="--match-read-wildcards">In the adapters and in the reads</option>
                <option value="--no-match-adapter-wildcards">Nowhere</option>
            </param>
            <param name="no_trim" argument="--no-trim" type="boolean" value="False" truevalue="--no-trim" falsevalue="" label="Do not trim adapters" help="Match and redirect reads to output/untrimmed-output as usual, but don't remove the adapters (default: trim the adapters)." />
            <param name="mask_adapter" argument="--mask-adapter" type="boolean" value="False" truevalue="--mask-adapter" falsevalue="" label="Mask Adapters" help="Mask adapter bases with 'N' instead of trimming them (default: trim adapters)." />
        </section>

        <!-- Filter Options -->
        <section name="filter_options" title="Filter Options">
            <param name="discard" argument="--discard-trimmed" type="boolean" value="False" truevalue="--discard" falsevalue="" label="Discard Trimmed Reads" help="Discard reads that contain the adapter instead of trimming them. Use the 'Minimum overlap length' option in order to avoid throwing away too many randomly matching reads!" />
            <param name="discard_untrimmed" argument="--discard_untrimmed" type="boolean" value="False" truevalue="--discard-untrimmed" falsevalue="" label="Discard Untrimmed Reads" help="Discard reads that do not contain the adapter." />
            <param name="min" argument="--minimum-length" type="integer" min="0" optional="True" value="0" label="Minimum length" help="Discard trimmed reads that are shorter than LENGTH.  Reads that are too short even before adapter removal are also discarded. In colorspace, an initial primer is not counted. Value of 0 means no minimum length." />
            <param name="max" argument="--maximum-length" type="integer" min="0" optional="True" value="0" label="Maximum length" help="Discard trimmed reads that are longer than LENGTH.  Reads that are too long even before adapter removal are also discarded. In colorspace, an initial primer is not counted. Value of 0 means no maximum length." />
            <param name="max_n" argument="--max-n" type="float" min="0" optional="True" label="Max N" help="Discard reads with more than this number of 'N' bases. A number between 0 and 1 is interpreted as a fraction of the read length." />
            <param name="pair_filter" argument="--pair-filter" type="select" optional="True" label="Pair filter" help="Which of the reads in a paired-end read have to match the filtering criterion in order for the pair to be filtered. Default: any">
                <option value="any" selected="True">any</option>
                <option value="both">both</option>
            </param>

            <param name="rest_file" argument="--rest-file" type="boolean" value="False" label="Rest of Read" help="When the adapter matches in the middle of a read, write the rest (after the adapter) into a file."/>
            <param name="wildcard_file" argument="--wildcard-file" type="boolean" value="False" label="Wildcard File" help="When the adapter has wildcard bases ('N's) write adapter bases matching wildcard positions to file."/>
            <param name="too_short_file" argument="--too-short-output" type="boolean" value="False" label="Too Short Reads" help="Write reads that are too short (according to minimum length specified) to a file. (default: discard reads)"/>
            <param name="too_long_file" argument="--too-long-output" type="boolean" value="False" label="Too Long Reads" help="Write reads that are too long (according to maximum length specified) to a file. (default: discard reads)"/>
            <param name="untrimmed_file" argument="--untrimmed-output" type="boolean" value="False" label="Untrimmed Reads" help="Write reads that do not contain the adapter to a separate file, instead of writing them to the regular output file.  (default: output to same file as trimmed)"/>
            <param name="multiple_output" argument="" type="boolean" value="False" label="Multiple output" help="Create a separate file for each adapter trimmed  (default: all trimmed reads are in a single file)"/>
        </section>

    </inputs>

    <outputs>

        <data name="too_long_paired_output" format="fastqsanger" metadata_source="input_2" label="${tool.name} on ${on_string}: Too Long Read 2" >
            <filter>(library['type'] == 'paired' or library['type'] == 'paired_collection')</filter>
            <filter>(output_options['too_long_file'] is True)</filter>
            <expand macro="inherit_format_2" />
        </data>

        <collection name="split_output" type="list" format_source="input" label="${tool.name} on ${on_string}: Split outputs">
            <discover_datasets pattern="__designation_and_ext__" directory="split" visible="false"   />
            <filter>(output_options['multiple_output'] is True)</filter>
        </collection>

    </outputs>

    <tests>
        <!-- Ensure fastq works -->
        <test>

                    <has_text text="Summary"/>
                </assert_contents>
            </output>
            <output name="info_file" value="cutadapt_info_out.txt" ftype="txt"/>
        </test>


        <test>
            <conditional name="library">
                <param name="type" value="single" />
                <param name="input_1" ftype="fastq" value="cutadapt_in_split.fastq" />
                <section name="r1" >
                    <repeat name="front_adapters">
                        <conditional name="front_adapter_source">
                            <param name="front_adapter_source_list" value="user"/>
                            <param name="front_adapter_name" value="A1" />
                            <param name="front_adapter" value="^GTCGGTAA" />
                        </conditional>
                    </repeat>
                    <repeat name="front_adapters">
                        <conditional name="front_adapter_source">
                            <param name="front_adapter_source_list" value="user"/>
                            <param name="front_adapter_name" value="A2" />
                            <param name="front_adapter" value="^AGGTCACT" />
                        </conditional>
                    </repeat>
                </section>
            </conditional>
            <param name="report" value="True" />
            <param name="info_file" value="False" />
            <param name="multiple_output" value="True" />
            <output_collection name="split_output" type="list" count="3">
                   <element name="A1" value="A1.fastq" ftype="fastq">
                   </element>
                   <element name="A2" value="A2.fastq" ftype="fastq">
                   </element>
                   <element name="unknown" value="unknown.fastq" ftype="fastq">
                   </element>
             </output_collection>
            <output name="report">
                <assert_contents>
                    <has_text text="Summary"/>
                </assert_contents>
            </output>
        </test>


    </tests>

    <help><![CDATA[

.. class:: infomark