<?xml version="1.0" encoding="utf-8"?>
<tool id="scanpy_run_pca" name="Scanpy RunPCA" version="@TOOL_VERSION@+galaxy1">
  <description>for dimensionality reduction</description>
  <macros>
    <import>scanpy_macros.xml</import>
  </macros>
  <expand macro="requirements"/>
  <command detect_errors="exit_code"><![CDATA[
ln -s '${input_obj_file}' input.h5 &&
PYTHONIOENCODING=utf-8 scanpy-run-pca.py
    -i input.h5
    -f '${input_format}'
    -o output.h5
    -F '${output_format}'
    -n '${n_pcs}'
    #if $run_mode.chunked
        -c
        --chunk-size '${run_mode.chunk_size}'
    #else
        #if $run_mode.zero_center
            -z
        #else
            -Z
        #end if
        #if $run_mode.svd_solver
            --svd-solver '${run_mode.svd_solver}'
        #end if
        #if $run_mode.random_seed is not None
            -s '${run_mode.random_seed}'
        #end if
    #end if
    #if $extra_outputs:
        #set extras = ' '.join(['--output-{}-file {}.csv'.format(x, x) for x in str($extra_outputs).split(',')])
        ${extras}
    #end if

@PLOT_OPTS@
]]></command>

  <inputs>
    <expand macro="input_object_params"/>
    <expand macro="output_object_params"/>
    <param name="n_pcs" argument="--n-pcs" type="integer" value="50" label="Number of PCs to produce"/>
    <conditional name="run_mode">
      <param name="chunked" argument="--chunked" type="boolean" checked="false" label="Perform incremental PCA by chunks"/>
      <when value="true">
        <param name="chunk_size" argument="--chunk-size" type="integer" value="0" label="Chunk size"/>
      </when>
      <when value="false">
        <param name="zero_center" argument="--zero-center" type="boolean" checked="true" label="Zero center data before scaling"/>
        <param name="svd_solver" argument="--svd-solver" type="select" optional="true" label="SVD solver">
          <option value="arpack">ARPACK</option>
          <option value="randomised">Randomised</option>
        </param>
        <param name="random_seed" argument="--random-seed" type="integer" value="0" label="random_seed for numpy random number generator"/>
      </when>
    </conditional>

    <param name="extra_outputs" type="select" multiple="true" optional="true" label="Type of output">
      <option value="embeddings">PCA embeddings</option>
      <option value="loadings">PCA loadings</option>
      <option value="stdev">PCs stdev</option>
      <option value="var-ratio">PCs proportion of variance</option>
    </param>

    <conditional name="do_plotting">
      <param name="plot" type="boolean" checked="false" label="Make PCA plot"/>
      <when value="true">
        <expand macro="output_plot_params"/>
      </when>
      <when value="false"/>
    </conditional>
  </inputs>

  <outputs>
    <data name="output_h5" format="h5" from_work_dir="output.h5" label="${tool.name} on ${on_string}: PCA object"/>
    <data name="output_png" format="png" from_work_dir="output.png" label="${tool.name} on ${on_string}: PCA plot">
      <filter>do_plotting['plot']</filter>
    </data>
    <data name="output_embed" format="csv" from_work_dir="embeddings.csv" label="${tool.name} on ${on_string}: PCA embeddings">
      <filter>extra_outputs and 'embeddings' in extra_outputs.split(',')</filter>
    </data>
    <data name="output_load" format="csv" from_work_dir="loadings.csv" label="${tool.name} on ${on_string}: PCA loadings">
      <filter>extra_outputs and 'loadings' in extra_outputs.split(',')</filter>
    </data>
    <data name="output_stdev" format="csv" from_work_dir="stdev.csv" label="${tool.name} on ${on_string}: PCA stdev">
      <filter>extra_outputs and 'stdev' in extra_outputs.split(',')</filter>
    </data>
    <data name="output_vprop" format="csv" from_work_dir="var-ratio.csv" label="${tool.name} on ${on_string}: PC explained proportion of variance">
      <filter>extra_outputs and 'var-ratio' in extra_outputs.split(',')</filter>
    </data>
  </outputs>

  <tests>
    <test>
      <param name="input_obj_file" value="scale_data.h5"/>
      <param name="input_format" value="anndata"/>
      <param name="output_format" value="anndata"/>
      <param name="extra_outputs" value="embeddings"/>
      <param name="n_pcs" value="50"/>
      <param name="zero_center" value="true"/>
      <param name="svd_solver" value="arpack"/>
      <param name="random_seed" value="0"/>
      <param name="chunked" value="false"/>
      <param name="plot" value="true"/>
      <param name="color_by" value="n_genes"/>
      <output name="output_h5" file="run_pca.h5" ftype="h5" compare="sim_size"/>
      <output name="output_png" file="run_pca.png" ftype="png" compare="sim_size"/>
      <output name="output_embed" file="run_pca.embeddings.csv" ftype="csv" compare="sim_size">
        <assert_contents>
          <has_n_columns n="50" sep=","/>
        </assert_contents>
      </output>
    </test>
  </tests>

  <help><![CDATA[
=======================================================================================================
Computes PCA (principal component analysis) coordinates, loadings and variance decomposition (`tl.pca`)
=======================================================================================================

It uses the implementation of *scikit-learn*.

@HELP@

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