<tool id="seurat_reduce_dimension" name="Seurat Run Dimensional Reduction" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
    <description>- PCA, tSNE or UMAP</description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="bio_tools"/>
    <expand macro="requirements"/>
    <expand macro="version_command"/>
    <command detect_errors="exit_code"><![CDATA[
@CMD@
    ]]></command>
    <configfiles>
        <configfile name="script_file"><![CDATA[
@CMD_imports@
@CMD_read_inputs@

#if $method.method == 'RunPCA'

    #if $method.features
    features_list<-paste(readLines('$method.features'), collapse=",")
    #end if

seurat_obj<-RunPCA(
    seurat_obj,
    #if $method.assay != ''
    assay = '$method.assay',
    #end if
    #if $method.features
    features = c(unlist(strsplit(features_list, ","))),
    #end if
    npcs = $method.npcs,
    rev.pca = $method.rev_pca,
    weight.by.var = $method.weight_by_var,
    reduction.key = '$method.reduction_key',
    seed.use = $method.seed_use,
    approx = $method.approx,
    reduction.name = '$method.reduction_name'
)

    #if $method.print_pcs.print_pcs == 'true'
    sink('print_pcs.txt')
    print(
        seurat_obj[['pca']],
        dims = 1:$method.print_pcs.dims,
        nfeatures = $method.print_pcs.nfeatures
    )
    sink()
    #end if

#else if $method.method == 'RunTSNE'
seurat_obj<-RunTSNE(
    seurat_obj,
    #if $method.reduction != ''
    reduction = '$method.reduction',
    #end if
    #if $method.dims
    dims = 1:$method.dims,
    #end if
    #if $method.assay != ''
    assay = '$method.assay',
    #end if
    seed.use = $method.seed_use,
    tsne.method = '$method.tsne_method',
    dim.embed = $method.dim_embed,
    reduction.key = '$method.reduction_key',
    reduction.name = '$method.reduction_name'
)

#else if $method.method == 'RunUMAP'
@reticulate_hack@

    #if $method.run_umap_on.run_umap_on == 'features'
    features_list<-paste(readLines('$method.run_umap_on.features'), collapse=",")
    #end if

seurat_obj<-RunUMAP(
    seurat_obj,
    #if $method.reduction != ''
    reduction = '$method.reduction',
    #end if
    #if $method.assay != ''
    assay = '$method.assay',
    #end if
    reduction.key = '$method.adv.reduction_key',
    reduction.name = '$method.adv.reduction_name',
    #if $method.adv.reduction_model != ''
    reduction.model = '$method.adv.reduction_model',
    #end if
    umap.method ='$method.umap_method.umap_method',
    #if $method.umap_method.umap_method == 'umap-learn'
    densmap = $method.umap_method.densmap.densmap,
        #if $method.umap_method.densmap.densmap == 'TRUE'
        dens.lambda = $method.umap_method.densmap.dens_lambda,
        dens.frac = $method.umap_method.densmap.dens_frac,
        dens.var.shift = $method.umap_method.densmap.dens_var_shift,
        #end if
    #end if
    n.neighbors = $method.n_neighbors,
    n.components = $method.n_components,
    metric = '$method.metric',
    #if $method.adv.n_epochs
    n.epochs = $method.adv.n_epochs,
    #end if
    learning.rate = $method.adv.learning_rate,
    min.dist = $method.adv.min_dist,
    spread = $method.adv.spread,
    set.op.mix.ratio = $method.adv.set_op_mix_ratio,
    local.connectivity = $method.adv.local_connectivity,
    repulsion.strength = $method.adv.repulsion_strength,
    negative.sample.rate = $method.adv.negative_sample_rate,
    #if $method.adv.a != ''
    a = '$method.adv.a',
    #end if
    #if $method.adv.b != ''
    b = '$method.adv.b',
    #end if
    uwot.sgd = $method.adv.uwot_sgd,
    seed.use = $method.adv.seed_use,
    angular.rp.forest = $method.adv.angular_rp_forest,
    #if $method.run_umap_on.run_umap_on == 'dims'
    dims = 1:$method.run_umap_on.dims,
    #else if $method.run_umap_on.run_umap_on == 'features'
    features = c(unlist(strsplit(features_list, ","))),
    slot = '$method.run_umap_on.slot',
    #else if $method.run_umap_on.run_umap_on == 'graph'
    graph = '$method.run_umap_on.graph',
    #else if $method.run_umap_on.run_umap_on == 'nn_name'
    nn.names = '$method.run_umap_on.nn_name',
    #end if
)

#end if

@CMD_rds_write_outputs@

]]></configfile>
    </configfiles>
    <inputs>
        <expand macro="input_rds"/>
        <conditional name="method">
            <param name="method" type="select" label="Method used">
                <option value="RunPCA">Run a PCA dimensionality reduction using 'RunPCA'</option>
                <option value="RunTSNE">Run a t-SNE dimensionality reduction using 'RunTSNE'</option>
                <option value="RunUMAP">Run a UMAP dimensional reduction using 'RunUMAP'</option>
            </param>
            <when value="RunPCA">
                <expand macro="select_assay"/>
                <param argument="npcs" type="integer" value="50" label="Number of PCs to compute"/>
                <param name="rev_pca" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Compute the PCA on the 'gene x cell' matrix instead of 'cell x gene' matrix" help="(rev.pca)"/>
                <param name="weight_by_var" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Weight the cell embeddings by the variance of each PC" help="(weight.by.var)"/>
                <param name="seed_use" type="integer" value="42" label="Set a random seed" help="(seed.use)"/>
                <param argument="approx" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="true" label="Use truncated singular value decomposition to approximate PCA"/>
                <param argument="features" type="data" format="txt,tabular" optional="true" label="List of features to compute PCA on" help="text file with one feature on each line, default is variable features"/>
                <param name="reduction_name" type="text" value="pca" label="Name for dimensional reduction" help="(reduction.name)">
                    <expand macro="valid_name"/>
                </param>
                <param name="reduction_key" type="text" value="PC_" label="Dimensional reduction key" help="specifies the string before the number for the dimension names (reduction.key)">
                        <expand macro="valid_reduction_key"/>
                </param>
                <conditional name="print_pcs">
                    <param name="print_pcs" type="select" label="Output list of top genes">
                        <option value="false" selected="true">No</option>
                        <option value="true">Yes</option>
                    </param>
                    <when value="false">
                    </when>
                    <when value="true">
                        <param argument="dims" type="integer" value="5" label="Number of PCs to print genes for"/>
                        <param argument="nfeatures" type="integer" value="5" label="Number of top genes to print for each PC"/>
                    </when>
                </conditional>
            </when>
            <when value="RunTSNE">
                <expand macro="select_reduction_pca"/>
                <expand macro="set_dims"/>
                <expand macro="select_assay"/>
                <param name="tsne_method" type="select" label="Select the method to use to compute the tSNE" help="(tsne.method)">
                    <option value="Rtsne" selected="true">Rtsne</option>
                    <option value="FIt-SNE">FIt-SNE</option>
                </param>
                <param name="seed_use" type="integer" value="1" label="Set a random seed" help="(seed.use)"/>
                <param name="dim_embed" type="integer" value="2" label="Set the dimensional space of the tSNE embedding" help="(dim.embed)"/>
                <param name="reduction_name" type="text" value="tsne" label="Name for dimensional reduction" help="(reduction.name)">
                    <expand macro="valid_name"/>
                </param>
                <param name="reduction_key" type="text" value="tSNE_" label="Dimensional reduction key" help="specifies the string before the number for the dimension names (reduction.key)">
                    <expand macro="valid_reduction_key"/>
                </param>
            </when>
            <when value="RunUMAP">
                <expand macro="select_reduction_pca"/>
                <expand macro="select_assay"/>
                <conditional name="umap_method">
                    <param name="umap_method" type="select" label="UMAP implementation to run" help="(umap.method)">
                        <option value="uwot" selected="true">uwot</option>
                        <option value="uwot-learn">uwot-learn</option>
                        <option value="umap-learn">umap-learn</option>
                    </param>
                    <when value="uwot">
                    </when>
                    <when value="uwot-learn">
                    </when>
                    <when value="umap-learn">
                        <conditional name="densmap">
                            <param argument="densmap" type="select" label="Use the density-augmented objective of densMAP">
                                <option value="TRUE">Yes</option>
                                <option value="FALSE" selected="true">No</option>
                            </param>
                            <when value="TRUE">
                                <param name="dens_lambda" type="integer" optional="true" value="2" label="Control the regularization weight of densMAP density correlation term" help="(dens.lambda)"/>
                                <param name="dens_frac" type="float" optional="true" min="0" max="1" value="0.3" label="Control the fraction of epochs where density-augmented objective is used in densMAP" help="(dens.frac)"/>
                                <param name="dens_var_shift" type="float" optional="true" value="0.1" label="Constant to add to variance of local radii when calculating density correlation objective" help="(dens.var.shift)"/>
                            </when>
                            <when value="FALSE">
                            </when>
                        </conditional>
                    </when>
                </conditional>
                <param name="n_neighbors" type="integer" value="30" label="Number of neighbors to use" help="Usually in range 5-50 (n.neighbors)"/>
                <param name="n_components" type="integer" value="2" label="Dimension of the space to embed into" help="(n.components)"/>
                <param argument="metric" type="select" label="Choose metric to use to measure distances in the input space">
                    <option value="cosine" selected="true">cosine</option>
                    <option value="correlation">correlation</option>
                    <option value="euclidean">euclidean</option>
                    <option value="manhattan">manhattan</option>
                    <option value="hamming">hamming</option>
                </param>
                <conditional name="run_umap_on">
                    <param name="run_umap_on" type="select" label="Run UMAP on dimensions, features, graph or KNN output">
                        <option value="dims" selected="true">dims</option>
                        <option value="features">features</option>
                        <option value="graph">graph</option>
                        <option value="nn_name">nn.name</option>
                    </param>
                    <when value="dims">
                        <expand macro="set_dims"/>
                    </when>
                    <when value="features">
                        <param argument="features" type="data" format="txt,tabular" optional="true" label="Run UMAP on this subset of features" help="text file with one feature on each line"/>
                        <expand macro="select_slot_data"/>
                    </when>
                    <when value="graph">
                        <param argument="graph" type="text" optional="true" value="" label="Name of graph on which to run UMAP">
                            <expand macro="valid_name"/>
                        </param>
                    </when>
                    <when value="nn_name">
                        <param name="nn_name" type="text" optional="true" value="" label="Name of KNN output on which to run UMAP" help="(nn.name)">
                            <expand macro="valid_name"/>
                        </param>
                    </when>
                </conditional>
                <section name="adv" title="Advanced Options">
                    <param name="reduction_model" type="text" optional="true" value="" label="DimReduct object that contains the umap model" help="(reduction.model)"/>
                    <param name="n_epochs" type="integer" optional="true" value="" label="Set the number of training epochs to use to optimise the embedding" help="if left empty, will be set as 500 for small datasets, 200 for large datasets (n.epochs)"/>
                    <param name="learning_rate" type="integer" value="1" label="Initial learning rate for the embedding optimization" help="(learning.rate)"/>
                    <param name="min_dist" type="float" value="0.3" label="Control how tightly points are compressed in the embedding" help="sensible values are in the range 0.001 to 0.5 (min.dist)"/>
                    <param argument="spread" type="integer" value="1" label="Effective spread of embedded points"/>
                    <param name="set_op_mix_ratio" type="float" value="1.0" label="Interpolate between fuzzy union and intersection" help="(set.op.mix.ratio)"/>
                    <param name="local_connectivity" type="integer" value="1" label="Set local connectivity" help="the number of nearest neighbors assummed to connect at local level (local.connectivity)"/>
                    <param name="repulsion_strength" type="integer" value="1" label="Apply weighting to negative values" help="(repulsion.strength)"/>
                    <param name="negative_sample_rate" type="integer" value="5" label="Number of negative samples to select per positive sample during optimization" help="(negative.sample.rate)"/>
                    <param argument="a" type="text" optional="true" value="" label="Parameter 'a' of differentiable approximation of right adjoint functor" help="automatically set if left empty"/>
                    <param argument="b" type="text" optional="true" value="" label="Parameter 'b' of differentiable approximation of right adjoint functor" help="automatically set if left empty"/>
                    <param name="uwot_sgd" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked ="false" label="Set uwot::umap(fast_sgd = TRUE)" help="(uwot.sgd)"/>
                    <param name="seed_use" type="integer" value="42" label="Set a random seed" help="(seed.use)"/>
                    <param name="angular_rp_forest" type="boolean" truevalue="TRUE" falsevalue="FALSE" checked="false" label="Use an angular random projection forest to initialise nearest neighbors" help="(angular.rp.forest)"/>
                    <param name="reduction_name" type="text" value="umap" label="Name for dimensional reduction" help="(reduction.name)">
                        <expand macro="valid_name"/>
                    </param>
                    <param name="reduction_key" type="text" value="UMAP_" label="Dimensional reduction key" help="specifies the string before the number for the dimension names (reduction.key)">
                        <expand macro="valid_reduction_key"/>
                    </param>
                </section>
            </when>
        </conditional>
        <expand macro="inputs_common_advanced"/>
    </inputs>
    <outputs>
        <expand macro="seurat_outputs"/>
        <expand macro="print_top_pcs"/>
    </outputs>
    <tests>
        <test expect_num_outputs="3">
            <!-- test1: RunPCA -->
            <param name="seurat_rds" location="https://zenodo.org/records/13732784/files/scaled.rds"/>
            <conditional name="method">
                <param name="method" value="RunPCA"/>
                <conditional name="print_pcs">
                    <param name="print_pcs" value="true"/>
                </conditional>
            </conditional>
            <section name="advanced_common">
                <param name="show_log" value="true"/>
            </section>
            <output name="hidden_output">
                <assert_contents>
                    <has_text_matching expression="RunPCA"/>
                </assert_contents>
            </output>
            <output name="rds_out" location="https://zenodo.org/records/13732784/files/pca.rds" ftype="rds" compare="sim_size"/>
            <output name="top_pcs" location="https://zenodo.org/records/13732784/files/top_pcs.txt" ftype="txt">
                <assert_contents>
                    <has_text_matching expression="PC_" n="5"/>
                </assert_contents>
            </output>
        </test>
        <test expect_num_outputs="2">
            <!-- test2: RuntSNE -->
            <param name="seurat_rds" location="https://zenodo.org/records/13732784/files/clusters.rds"/>
            <conditional name="method">
                <param name="method" value="RunTSNE"/>
                <param name="seed.use" value="1"/>
                <param name="tsne_method" value="Rtsne"/>
                <param name="dim_embed" value="2"/>
                <param name="reduction_key" value="tSNE_"/>
            </conditional>
            <section name="advanced_common">
                <param name="show_log" value="true"/>
            </section>
            <output name="hidden_output">
                <assert_contents>
                    <has_text_matching expression="RunTSNE"/>
                </assert_contents>
            </output>
            <output name="rds_out" location="https://zenodo.org/records/13732784/files/tSNE.rds" ftype="rds" compare="sim_size"/>
        </test>
                <test expect_num_outputs="2">
            <!-- test3: RuntSNE with FIt-SNE -->
            <param name="seurat_rds" location="https://zenodo.org/records/13732784/files/clusters.rds"/>
            <conditional name="method">
                <param name="method" value="RunTSNE"/>
                <param name="seed.use" value="1"/>
                <param name="tsne_method" value="FIt-SNE"/>
                <param name="dim_embed" value="2"/>
                <param name="reduction_key" value="tSNE_"/>
            </conditional>
            <section name="advanced_common">
                <param name="show_log" value="true"/>
            </section>
            <output name="hidden_output">
                <assert_contents>
                    <has_text_matching expression="RunTSNE"/>
                </assert_contents>
            </output>
            <output name="rds_out" location="https://zenodo.org/records/13732784/files/FIt-SNE.rds" ftype="rds" compare="sim_size"/>
        </test>
        <test expect_num_outputs="2">
            <!-- test4: RunUMAP -->
            <param name="seurat_rds" location="https://zenodo.org/records/13732784/files/clusters.rds"/>
            <conditional name="method">
                <param name="method" value="RunUMAP"/>
                <param name="reduction_key" value="UMAP_"/>
                <conditional name="umap_method">
                    <param name="umap_method" value="uwot"/>
                </conditional>
            </conditional>
            <section name="advanced_common">
                <param name="show_log" value="true"/>
            </section>
            <output name="hidden_output">
                <assert_contents>
                    <has_text_matching expression="RunUMAP"/>
                </assert_contents>
            </output>
            <output name="rds_out" location="https://zenodo.org/records/13732784/files/umap.rds" ftype="rds" compare="sim_size"/>
        </test>
        <test expect_num_outputs="2">
            <!-- test5: RunUMAP - umap-learn -->
            <param name="seurat_rds" location="https://zenodo.org/records/13732784/files/clusters.rds"/>
            <conditional name="method">
                <param name="method" value="RunUMAP"/>
                <param name="reduction_key" value="UMAP_"/>
                <conditional name="umap_method">
                    <param name="umap_method" value="umap-learn"/>
                    <conditional name="densmap">
                        <param name="densmap" value="FALSE"/>
                    </conditional>
                </conditional>
            </conditional>
            <section name="advanced_common">
                <param name="show_log" value="true"/>
            </section>
            <output name="hidden_output">
                <assert_contents>
                    <has_text_matching expression="RunUMAP"/>
                </assert_contents>
            </output>
            <output name="rds_out" location="https://zenodo.org/records/13732784/files/umap-learn.rds" ftype="rds" compare="sim_size"/>
        </test>
    </tests>
    <help><![CDATA[
Seurat
======

Seurat is an R package designed for QC, analysis, and exploration of single-cell RNA-seq data.

Seurat aims to enable users to identify and interpret sources of heterogeneity from single-cell transcriptomic measurements, and to integrate diverse types of single-cell data.

RunPCA
======

Run a PCA dimensionality reduction.

More details on the `seurat documentation
<https://satijalab.org/seurat/reference/runpca>`__

RuntSNE
=======

Run t-SNE dimensionality reduction on selected features.

More details on the `seurat documentation
<https://satijalab.org/seurat/reference/runtsne>`__

RunUMAP
=======

Runs the Uniform Manifold Approximation and Projection (UMAP) dimensional reduction technique.

More details on the `seurat documentation
<https://satijalab.org/seurat/reference/runumap>`__

    ]]></help>
    <expand macro="citations"/>
</tool>