<tool id="schicexplorer_schiccluster" name="@BINARY@" version="@TOOL_VERSION@.1" profile="@PROFILE@">
    <description>clusters single-cell Hi-C interaction matrices on the raw data</description>
    <macros>
        <token name="@BINARY@">scHicCluster</token>
        <import>macros.xml</import>
    </macros>
    <expand macro="requirements" />
    <command detect_errors="exit_code"><![CDATA[
        @BINARY@

        --matrix '$matrix_scooler'

        --numberOfClusters $numberOfClusters

        --clusterMethod $clusterMethod_selector

        #if $chromosomes:
            #set $chromosome = ' '.join([ '\'%s\'' % $chrom for $chrom in str($chromosomes).split(' ') ])
            --chromosomes $chromosome
        #end if

        --dimensionReductionMethod $dimensionReductionMethod_selector
        --numberOfNearestNeighbors $numberOfNearestNeighbors
        --outFileName cluster_list.txt

        --threads @THREADS@



    ]]></command>
    <inputs>
        
        <expand macro="matrix_scooler_macro"/>
        <param name="clusterMethod_selector" type="select" label="Cluster method:">
                <option value="kmeans" selected="True">K-means</option>
                <option value="spectral" >Spectral clustering</option>
        </param>
        <param name="dimensionReductionMethod_selector" type="select" label="Apply dimension reduction:">
                <option value="none" selected="True">No dimension reduction</option>
                <option value="knn" >Exact k-nearest neighbors</option>
                <option value="pca" >Principal components</option>
        </param>
        <param name="numberOfClusters" type="integer" value="7"  label="Number of clusters" help='How many clusters should be computed by kmeans or spectral clustering' />   
        <param name="numberOfNearestNeighbors" type="integer" value="100"  label="Number of nearest neighbors" help='How many nearest neighbors should be computed for the k-nn graph?' />   

        <param name='chromosomes' type='text' label='List of chromosomes to consider' help='Please separate the chromosomes by space'/>
    </inputs>
    <outputs>
        <data name="outFileName" from_work_dir="cluster_list.txt" format="txt" label="${tool.name} on ${on_string}: Cluster results"/>
    </outputs>
    <tests>
        <test>
            <param name='matrix_scooler' value='test_matrix.scool' />
            <param name='clusterMethod_selector' value='kmeans' />
            <param name='dimensionReductionMethod_selector' value='none' />
            <param name='numberOfClusters' value='7' />
            <output name="outFileName" file="scHicCluster/cluster_kmeans.txt" ftype="txt" compare="sim_size" delta="4000"/>
        </test>
        <test>
            <param name='matrix_scooler' value='test_matrix.scool' />
            <param name='clusterMethod_selector' value='kmeans' />
            <param name='dimensionReductionMethod_selector' value='none' />
            <param name='numberOfClusters' value='7' />
            <param name='chromosomes' value='chr1 chr2' />

            <output name="outFileName" file="scHicCluster/cluster_kmeans_chromosomes.txt" ftype="txt" compare="sim_size" delta="4000"/>
        </test>
        <test>
            <param name='matrix_scooler' value='test_matrix.scool' />
            <param name='clusterMethod_selector' value='spectral' />
            <param name='dimensionReductionMethod_selector' value='knn' />
            <param name='numberOfClusters' value='7' />
            <output name="outFileName" file="scHicCluster/cluster_spectral_knn.txt" ftype="txt" compare="sim_size" delta="4000"/>
        </test>
        <test>
            <param name='matrix_scooler' value='test_matrix.scool' />
            <param name='clusterMethod_selector' value='spectral' />
            <param name='dimensionReductionMethod_selector' value='pca' />
            <param name='numberOfClusters' value='7' />
            <output name="outFileName" file="scHicCluster/cluster_spectral_pca.txt" ftype="txt" compare="sim_size" delta="4000"/>
        </test>
    </tests>
    <help><![CDATA[

Clustering on raw data
======================

scHicCluster uses kmeans or spectral clustering to associate each cell to a cluster and therefore to its cell cycle. 
The clustering can be run on the raw data, on a kNN computed via the exact euclidean distance or via PCA. 
Please consider also the other clustering and dimension reduction approaches of the scHicExplorer suite such as `scHicCluster`, `scHicClusterMinHash` and `scHicClusterSVL`. 
They can give you better results, can be faster or less memory demanding.

For more information about scHiCExplorer please consider our documentation on readthedocs.io_

.. _readthedocs.io: http://schicexplorer.readthedocs.io/
]]></help>
    <expand macro="citations" />

</tool>