<tool id="blastxml_to_gapped_gff3" name="BlastXML to gapped GFF3" version="1.1">
  <description></description>
  <macros>
    <import>macros.xml</import>
  </macros>
  <expand macro="requirements"/>
  <expand macro="stdio"/>
  <version_command>python blastxml_to_gapped_gff3.py --version</version_command>
  <command>
  <![CDATA[
    python '$__tool_directory__/blastxml_to_gapped_gff3.py'
      '$blastxml'
      --min_gap $min_gap
      $trim
      > '$output'
  ]]></command>
  <inputs>
    <param label="Blast XML" name="blastxml" type="data" format="blastxml"/>
    <param label="Maximum gap size before generating a new match_part" name="min_gap" type="integer" value="3"/>

    <param type="select" label="Trim sides of blast hits" name="trim">
        <option value="">None</option>
        <option value="--trim">Trim start and end</option>
        <option value="--trim_end" selected="true">Trim only end</option>
    </param>
  </inputs>
  <outputs>
    <data format="gff3" name="output"/>
  </outputs>
  <tests>
      <test>
          <param name="blastxml" ftype="blastxml" value="input.xml"/>
          <output name="output" file="blast.gff"/>
      </test>
  </tests>
  <help><![CDATA[
**What it does**

Convert BlastXML results into GFF3 format.

**Options**

The trimming option captures an important feature provided in this tool that
isn't provided in most other BlastXML visualization tools: the fact that blast
captures complete alignment location information.

This means that when most blast visualization tools produce output which looks
like this:

.. image:: $PATH_TO_IMAGES/blast2html.png

This tool produces output which shows where the real subject sequence starts and ends relative to your sequence:

.. image:: $PATH_TO_IMAGES/blast-extended.png

This can be a useful feature for examining alternate start locations that are
used by sequences found from your blast query.

The green bars on the very top row of the picture indicate start sites, as you
can see the blast hits and the genome in the visualization share an upstream
start site. You don't lose the information present in your blastxml data.

So, to the end of useful functionality, this options is controllable:

-  you can trim neither end, see where the real protein alignments are.
-  you can trim both ends, like blasts HTML reports
-  you can trim just the end of the sequence, as upstream is generally more
   interesting, and having long tails can result in poor visualizations.
]]></help>
    <citations>
    </citations>
</tool>