<tool id="pm_muairss_read" name="PyMuonSuite AIRSS Cluster" version="@PYMUONSUITE_VERSION@+galaxy@WRAPPER_VERSION@" python_template_version="3.5" profile="22.05" license="MIT">
    <description>run clustering for optimised structures</description>
    <macros>
        <!-- version of this tool wrapper (integer) -->
        <token name="@WRAPPER_VERSION@">0</token>
        <import>muon_macros.xml</import>
        <xml name="supercell">
            <param type="text" argument="supercell" value="1" label="Supercell" help="Supercell size and shape to use when saving cluster structures. This can either be a single int, a list of three integers or a 3x3 matrix of integers. For a single number a diagonal matrix will be generated with the integer repeated on the diagonals. For a list of three numbers a diagonal matrix will be generated where the diagonal elements are set to the list. A matrix will be used directly as is. Default is a 3x3 identity matrix.">
                <validator type="regex" message="Input should only contain whitespace, '[', ']', ',' and digits.">^[\s\d,\[\]]+$</validator>
            </param>
        </xml>
    </macros>
    <creator>
        <person givenName="Jyothish" familyName="Thomas" identifier="https://orcid.org/0000-0003-4724-6924"/>
        <person givenName="Eli" familyName="Chadwick" url="https://github.com/elichad" identifier="https://orcid.org/0000-0002-0035-6475"/>
        <person givenName="Patrick" familyName="Austin" url="https://github.com/patrick-austin" identifier="https://orcid.org/0000-0002-6279-7823"/>
        <organization url="https://muon-spectroscopy-computational-project.github.io/index.html" name="The Muon Spectroscopy Computational Project"/>
    </creator>
    <requirements>
        <requirement type="package" version="@PYMUONSUITE_VERSION@">pymuonsuite</requirement>
        <requirement type="package" version="3.0">zip</requirement>
    </requirements>
    <required_files>
        <include type="literal" path="get_out_folder.py"/>
        <include type="literal" path="config.py"/>
    </required_files>
    <command detect_errors="exit_code"><![CDATA[
        unzip '$optimisation_results' &&
        if test -f "params.yaml"; then echo "params.yaml present"; else echo "params.yaml missing" && exit 64; fi
        && if ( test -f input_structure.* ) ; then echo "input structure present"; else echo "input structure missing" && exit 64; fi
        && out_folder="`python '${__tool_directory__}/get_out_folder.py'`"
        && python '${__tool_directory__}/config.py' 
        && pm-muairss -t r input_structure.* params.yaml
        #if $clustering_save.clustering_save_format != "none":
            && ln -s *_clusters/* minimal_clusters
        #end if
    ]]></command>
    <configfiles>
        <configfile filename="incoming_params">clustering_method: $clustering.clustering_method
#if $clustering.clustering_method == "hier":
clustering_hier_t: $clustering.clustering_hier_t
#else if $clustering.clustering_method == "kmeans":
clustering_kmeans_k: $clustering.clustering_kmeans_k
#end if
#if $clustering_save.clustering_save_format != "none":
clustering_save_type: structures
clustering_save_format: $clustering_save.clustering_save_format
supercell: $clustering_save.supercell.replace('__ob__', '[').replace('__cb__', ']')
#if $clustering_save.clustering_save_format == "cell":
mu_symbol: $clustering_save.mu_symbol
#end if
#end if</configfile>
    </configfiles>
    <inputs>
        <param type="data" name="optimisation_results" label="optimised muonated structures (.zip)" format="zip" help="A zip folder containing a set of optimised muonated structures, the original structure, and a YAML parameter file. See below for the expected folder structure."/>
        <conditional name="clustering">
            <param type="select" argument="clustering_method" display="radio" label="Clustering method" value="hier">
                <option value="hier">hierarchical</option>
                <option value="kmeans">k-means</option>
            </param>
            <when value="hier">
                <param type="float" argument="clustering_hier_t" value="0.3" min="0.0" optional="true" label="t parameter for hierarchical clustering"/>
            </when>
            <when value="kmeans">
                <param type="integer" argument="clustering_kmeans_k" value="4" min="0" optional="true" label="Number of clusters for k-means clustering"/>
            </when>
        </conditional>
        <conditional name="clustering_save">
            <param argument="clustering_save_format" type="select" label="Clustering Save Format" help="If set, for each cluster a structural file with the specified format will be generated, corresponding to the structure in the cluster with minimal energy.">
                <option value="none" selected="true">Do not generate</option>
                <option value="cell">CELL</option>
                <option value="cif">CIF</option>
                <option value="xyz">XYZ</option>
            </param>
            <when value="none"/>
            <when value="cell">
                <expand macro="supercell"/>
                <param argument="mu_symbol" type="text" value="H:mu" optional="true" label="Muon symbol" help="The symbol to use for the muon when writing out the CASTEP custom species. Should be a valid chemical symbol followed by `:`, then custom text."/>
            </when>
            <when value="cif">
                <expand macro="supercell"/>
            </when>
            <when value="xyz">
                <expand macro="supercell"/>
            </when>
        </conditional>
    </inputs>
    <outputs>
        <data label="Cluster report for $optimisation_results.name" name="cluster_report" format="txt" from_work_dir="${out_folder}/*clusters.txt"/>
        <data label="Cluster data for $optimisation_results.name" name="cluster_data" format="txt" from_work_dir="${out_folder}/*clusters.dat"/>
        <collection name="saved_structures" type="list" label="Minimal energy structures" format="txt">
            <filter>clustering_save["clustering_save_format"] != "none"</filter>
            <discover_datasets pattern="__name_and_ext__" directory="minimal_clusters"/>
        </collection>
    </outputs>
    <tests>
        <test expect_num_outputs="2">
            <param name="optimisation_results" value="uep-out.zip" ftype="zip"/>
            <conditional name="clustering">
                <param name="clustering_hier_t" value="0.2"/>
            </conditional>
            <output name="cluster_report" file="clustout-uep.txt" ftype="txt" lines_diff="2"/>
            <output name="cluster_data" file="clustout-uep.dat" ftype="txt"/>
        </test>
        <test expect_num_outputs="3">
            <param name="optimisation_results" value="uep-out.zip" ftype="zip"/>
            <conditional name="clustering">
                <param name="clustering_hier_t" value="0.2"/>
            </conditional>
            <conditional name="clustering_save">
                <param name="clustering_save_format" value="cell"/>
            </conditional>
            <output name="cluster_report" file="clustout-uep.txt" ftype="txt" lines_diff="2"/>
            <output name="cluster_data" file="clustout-uep.dat" ftype="txt"/>
            <output_collection name="saved_structures" type="list" count="2"/>
        </test>
        <!-- Test optional save params -->
        <test expect_num_outputs="3">
            <param name="optimisation_results" value="uep-out.zip" ftype="zip"/>
            <conditional name="clustering">
                <param name="clustering_hier_t" value="0.2"/>
            </conditional>
            <conditional name="clustering_save">
                <param name="clustering_save_format" value="cell"/>
                <param name="supercell" value="2"/>
                <param name="mu_symbol" value="H:nu"/>
            </conditional>
            <output name="cluster_report" file="clustout-uep.txt" ftype="txt" lines_diff="2"/>
            <output name="cluster_data" file="clustout-uep.dat" ftype="txt"/>
            <output_collection name="saved_structures" type="list" count="2">
                <element name="Si_uep_min_cluster_1" file="Si_uep_min_cluster_1.cell"/>
                <element name="Si_uep_min_cluster_2" file="Si_uep_min_cluster_2.cell"/>
            </output_collection>
        </test>
        <test expect_num_outputs="2">
            <param name="optimisation_results" value="dftb-out.zip" ftype="zip"/>
            <conditional name="clustering">
                <param name="clustering_hier_t" value="0.2"/>
            </conditional>
            <output name="cluster_report" file="clustout-dftb.txt" ftype="txt" lines_diff="2"/>
            <output name="cluster_data" file="clustout-dftb.dat" ftype="txt"/>
        </test>
        <!-- Test overwrite of values in the existing params.yaml -->
        <test expect_num_outputs="2">
            <param name="optimisation_results" value="uep-out.zip" ftype="zip"/>
            <conditional name="clustering">
                <param name="clustering_method" value="kmeans"/>
                <param name="clustering_kmeans_k" value="2"/>
            </conditional>
            <!-- kmeans gives predictable clusters, but orders them randomly -->
            <!-- To prevent this from affecting tests, sort the lines -->
            <!-- in the report and use regex matching for the dat file -->
            <output name="cluster_report" file="clustout-uep-kmeans.txt" ftype="txt" lines_diff="2" sort="true"/>
            <output name="cluster_data" file="clustout-uep-kmeans.dat" ftype="txt" compare="re_match"/>
        </test>
    </tests>
    <help><![CDATA[
        Given a set of optimised muonated structures, clusters the structures based on similarity.

        This is intended to be used with the output from '`PyMuonSuite AIRSS UEP Optimise`_'.
        If you used the 'Run UEP optimisation' tool for optimisation, the output from that tool can be used as the input to this one.
        If you are bringing files from elsewhere (such as running PyMuonSuite from the command line, or with a different optimisation calculator), make sure they fit the expected zip folder structure:

        .. code-block::

            optimised-structures.zip/
            ├─ muon-airss-out/
            │  ├─ uep/
            │  │  ├─ struct_1/
            │  │  │  ├─ <files for optimised struct_1 go here>
            │  │  ├─ struct_2/
            │  │  │  ├─ <files for optimised struct_2 go here>
            │  │  ├─ struct_N/
            │  │  │  ├─ <files for optimised struct_N go here>
            ├─ input.cell
            ├─ params.yaml

        This folder represents a run with the output folder name ``muon-airss-out``, the structure name ``struct``, and the ``uep`` calculator.
        If you use different parameters, these names should match your own settings (``dftb+`` and ``castep`` are used as folder names for the respective calculators).
        The zip folder itself has been named ``optimised-structures.zip`` in this example, but it doesn't matter what you call it.

        The files ``input.cell`` and ``params.yaml`` must have those exact names. ``input.cell`` should be your original structure file,
        and ``params.yaml`` is the YAML parameter file used to configure the muonation and optimisation.
        This file is created automatically when using '`PyMuonSuite AIRSS UEP Optimise`_'.
        Any parameters relating to the clustering process present in the original file will be overwritten with the values selected here.

        Command-line usage: pm-muairss [-h] -t r structures parameter_file

        .. _PyMuonSuite AIRSS UEP Optimise: /tool_runner?tool_id=pm_uep_opt

        PyMuonSuite is distributed under the GPLv3 license. This tool wrapper is distributed under the MIT license.
    ]]></help>
    <citations>
        <citation type="doi">@PYMUONSUITE_CITATION@</citation>
        <citation type="doi">10.1063/1.5024450</citation>
        <citation type="doi">10.1063/1.5085197</citation>
        <citation type="doi">10.1063/5.0012381</citation>
        <citation type="bibtex">
            @article {castep,
                author = {Clark, S. J. and Segall, M. D. and Pickard, C. J. and Hasnip, P. J. and Probert, M. I. J. and Refson, K. and Payne, M. C.},
                title = {First principles methods using CASTEP},
                journal = {Zeitschrift fuer Kristallographie},
                volume = {220},
                pages = {567},
                year = {2005}
            }
        </citation>
    </citations>
</tool>