<tool id="snpEff_databases" name="SnpEff databases:" version="@wrapper_version@.1">
    <description> list available databases</description>
    <macros>
        <import>snpEff_macros.xml</import>
    </macros>
    <requirements>
        <expand macro="requirement" />
    </requirements>
    <expand macro="stdio" />
    <expand macro="version_command" />
    <command><![CDATA[

        snpEff databases | grep -v '^---' | sed 's/^Genome/#Genome/' | sed 's/  *//g' 

        #if $include_pattern:
             | grep '${include_pattern}' 
        #end if

        #if $exclude_pattern:
             | grep -v '${exclude_pattern}' 
        #end if

        > '${snpeff_dbs}'
            
    ]]></command>
    <inputs>
        <param name="include_pattern" size="40" type="text" value="" optional="True" label="List entries matching the following expression" help="Databases matching this expression will be listed. Here you can enter text or regular expression. For example, to show only mouse databases use 'Mouse'. Note that this parameter is case-sensitive.">
            <sanitizer>
                <valid initial="string.digits,string.letters">
                    <add value="^" />
                </valid>
            </sanitizer>
        </param>
        <param name="exclude_pattern" size="40" type="text" value="" optional="True" label="Do not output entries matching the following expression" help="Databases matching this expression WILL NOT BE listed. Here you can enter text or regular expression. For excample, to avoid all ENSEMBL bundles enter 'ENSEMBL'. Note that this parameter is case-sensitive.">
            <sanitizer>
                <valid initial="string.digits,string.letters">
                    <add value="^" />
                </valid>
            </sanitizer>
        </param>

    </inputs>
    <outputs>
        <data format="tabular" name="snpeff_dbs" label="${tool.name} @snpeff_version@ available databases" />
    </outputs>
    <tests>
        <test>
            <output name="snpeff_dbs">
                <assert_contents>
                    <has_text text="ebola_zaire" />
                </assert_contents>
            </output>
        </test>
         <test>
            <param name="include_pattern" value="ebola"/>
            <output name="snpeff_dbs">
                <assert_contents>
                    <has_text text="ebola_zaire" />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help><![CDATA[

**What it does**

This tool downloads the master list of snpEff databases from @snpeff_database_url@. You can then look at this list and decide which database to use for your analysis. For example, if **List entries matching the following expression** parameter of this tool is set to *Mouse* the it will produce a tabular dataset with the following content::

    mm10  Mouse  http://downloads.sourceforge.net/project/snpeff/databases/v4_3/snpEff_v4_3_mm10.zip
    mm9   Mouse  http://downloads.sourceforge.net/project/snpeff/databases/v4_3/snpEff_v4_3_mm9.zip

This means that there two available snpEff databases for mouse genome versions mm9 and mm10. In order to download these databases you should use identifier from the first column (e.g., mm9 or mm10 in this case). 

-------

.. class:: infomark

**The usage scenario**

There are two ways to use names of databases obtained with this tool in Galaxy's version on snpEff:

    #. Use **SnpEff download** tool. It will download the database to the history and you will be able to use it in **SnpEff eff** tool using *Downloaded snpEff database in your history* option of the **Genome source** parameter.
    #. Use *Download on demand* option of the **SnpEff eff** tool (again, **Genome source** parameter). In this case snpEff will download the database before performing annotation. 

@snpeff_in_galaxy_info@
@external_documentation@
    ]]></help>
    <expand macro="citations" />
</tool>