<tool id="snpSift_annotate" name="SnpSift Annotate" version="@WRAPPER_VERSION@.galaxy1">
    <description>SNPs from dbSnp</description>
    <!--
        You can change the amount of memory used, just change the -Xmx parameter (e.g. use -Xmx2G for 2Gb of memory)
    -->
    <macros>
        <import>snpSift_macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <expand macro="stdio" />
    <expand macro="version_command" />
    <command><![CDATA[
ln -s '$dbSnp' 'dbSnp.vcf' &&
SnpSift -Xmx8G annotate
${annotate.no_info}
#if str($annotate.info_ids).strip():
    -info '$annotate.info_ids'
#end if
-q 'dbSnp.vcf' '$input' > '$output'
    ]]></command>
    <inputs>
        <param name="input" type="data" format="vcf" label="Variant input file in VCF format"/>
        <param name="dbSnp" type="data" format="vcf" label="VCF File with ID field annotated (e.g. dbSNP.vcf)"
            help="The ID field for a variant in input will be assigned from a matching variant in this file."/>
        <conditional name="annotate">
            <param name="no_info" type="select" label="Fields to annotate">
                <option value="-noInfo">Only annotate ID field (no INFO fields will be added)</option>
                <option value="">Add also INFO fields</option>
            </param>
            <when value="-noInfo">
                <param name="info_ids" type="hidden" value="" />
            </when>
            <when value="">
                <param name="info_ids" type="text" value="" label="Limit INFO annotation to these INFO IDs" help="List is a comma separated list of fields. When blank, all INFO fields are included">
                    <validator type="regex" message="IDs separated by commas">^(([a-zA-Z][a-zA-Z0-9_-]*)(,[a-zA-Z][a-zA-Z0-9_-]*)*)?$</validator>
                </param>
            </when>
        </conditional>
    </inputs>
    <outputs>
        <data name="output" format="vcf" />
    </outputs>
    <tests>
        <test>
            <param name="input" ftype="vcf" value="annotate_1.vcf"/>
            <param name="dbSnp" ftype="vcf" value="db_test_1.vcf"/>
            <conditional name="annotate">
                <param name="no_info" value=""/>
            </conditional>
            <output name="output">
                <assert_contents>
                    <has_line_matching expression="1&#009;872687&#009;rs76166080&#009;C&#009;G&#009;.&#009;.&#009;.+;.+" />
                </assert_contents>
            </output>
        </test>
        <test>
            <param name="input" ftype="vcf" value="annotate_1.vcf"/>
            <param name="dbSnp" ftype="vcf" value="db_test_1.vcf"/>
            <conditional name="annotate">
                <param name="no_info" value=""/>
                <param name="info_ids" value="dbSNPBuildID" />
            </conditional>
            <output name="output">
                <assert_contents>
                    <has_line_matching expression="1&#009;872687&#009;rs76166080&#009;C&#009;G&#009;.&#009;.&#009;dbSNPBuildID=137" />
                </assert_contents>
            </output>
        </test>
        <test>
            <param name="input" ftype="vcf" value="annotate_1.vcf"/>
            <param name="dbSnp" ftype="vcf" value="db_test_1.vcf"/>
            <conditional name="annotate">
                <param name="no_info" value="-noInfo"/>
            </conditional>
            <output name="output">
                <assert_contents>
                    <has_line line="1&#009;872687&#009;rs76166080&#009;C&#009;G&#009;.&#009;.&#009;." />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[
This is typically used to annotate IDs from dbSnp.

Annotating only the ID field from dbSnp137.vcf ::

    Input VCF:
    CHROM  POS         ID           REF  ALT  QUAL   FILTER  INFO
    22      16157571    .            T    G    0.0    FAIL    NS=53
    22      16346045    .            T    C    0.0    FAIL    NS=244
    22      16350245    .            C    A    0.0    FAIL    NS=192

    Annotated Output VCF:
    #CHROM  POS         ID           REF  ALT  QUAL   FILTER  INFO
    22      16157571    .            T    G    0.0    FAIL    NS=53
    22      16346045    rs56234788   T    C    0.0    FAIL    NS=244
    22      16350245    rs2905295    C    A    0.0    FAIL    NS=192

Annotating both the ID and INFO fields from dbSnp137.vcf ::

    Input VCF:
    #CHROM  POS         ID           REF  ALT  QUAL   FILTER  INFO
    22      16157571    .            T    G    0.0    FAIL    NS=53
    22      16346045    .            T    C    0.0    FAIL    NS=244
    22      16350245    .            C    A    0.0    FAIL    NS=192

    Annotated Output VCF:
    #CHROM  POS         ID           REF  ALT  QUAL   FILTER  INFO
    22      16157571    .            T    G    0.0    FAIL    NS=53
    22      16346045    rs56234788   T    C    0.0    FAIL    NS=244;RSPOS=16346045;GMAF=0.162248628884826;dbSNPBuildID=129;SSR=0;SAO=0;VP=050100000000000100000100;WGT=0;VC=SNV;SLO;GNO
    22      16350245    rs2905295    C    A    0.0    FAIL    NS=192;RSPOS=16350245;GMAF=0.230804387568556;dbSNPBuildID=101;SSR=1;SAO=0;VP=050000000000000100000140;WGT=0;VC=SNV;GNO

@EXTERNAL_DOCUMENTATION@
- http://snpeff.sourceforge.net/SnpSift.html#annotate
    ]]></help>
    <expand macro="citations" />
</tool>