view NSPDK_sparseVect.xml @ 5:9b5fb987a484 draft

planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust/NSPDK commit b'7e5cd452018ae9507c2d1cd13dd688a747550393\n'
author rnateam
date Fri, 17 Mar 2017 08:28:51 -0400
parents 0625db6b6e45
children 2aaf391798a3
line wrap: on
line source

<tool id="nspdk_sparse" name="NSPDK_sparseVect" version="9.2">
    <requirements>
        <requirement type="package" version="0.1.11">graphclust-wrappers</requirement>
        <requirement type="package" version="9.2">nspdk</requirement>
    </requirements>
    <stdio>
        <exit_code range="1:" />
    </stdio>
    <command>
        <![CDATA[

        NSPDK_sparseVect.pl
            '$data_fasta'
            '$gspan_file'
            $max_rad
            $max_dist_relations

]]>
    </command>
    <inputs>
        <param type="data" name="gspan_file" format="txt" label="gspan file"/>
        <param type="data" name="data_fasta" format="fasta" />
        <param name="max_rad" type="integer" value="3" label="maximum radius " help="-R"/>
        <param name="max_dist_relations" type="integer" value="3" label="maximum distance relations" help="-D"/>
    </inputs>
    <outputs>
        <data name="data_svector" format="zip" from_work_dir="SVECTOR/data.svector" label="data_svector"/>
    </outputs>
    <tests>
        <test>
            <param name="data_fasta" value="data.fasta"/>
            <param name="gspan_file" value="1.group.gspan.bz2" ftype="searchgui_archive"/>
            <param name="max_rad" value="3"/>
            <param name="max_dist_relations" value="3"/>
            <output name="data_svector" file="SVECTOR/data.svector" ftype="zip" />
        </test>
    </tests>
    <help>
        <![CDATA[

**What it does**

Produces an explicit sparse feature encoding.
Integer code for the invariant graph encoding is used as a feature indicator. In this way,
the integer associated to each feature (i.e. each pair or neighborhood subgraphs of radius r whose
roots are at distance d) can be interpreted as the feature key and the (normalized) count of occurrences as its value.
This allows to obtain an explicit feature encoding for a given graph G as a sparse vector in ℝ^m (with a very high dimension m).

**Parameters**

+    **-R** <max radius> (default: 1)
+    **-D** <max distance relations> (default: 4)


    ]]>
    </help>
    <citations>
        <citation type="doi">10.1093/bioinformatics/bts224</citation>
        <citation type="bibtex">@inproceedings{costa2010fast,
          title={Fast neighborhood subgraph pairwise distance kernel},
          author={Costa, Fabrizio and De Grave, Kurt},
          booktitle={Proceedings of the 26th International Conference on Machine Learning},
          pages={255--262},
          year={2010},
          organization={Omnipress}
        }
        </citation>
    </citations>
</tool>