<tool id="fastspar_reduce" name="FastSpar: Reduce correlation table" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
    <description>
         Filter correlation and p-value table into sparse matrices
    </description>
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="biotools"/>
    <expand macro="requirements"/>
    <command detect_errors="exit_code"><![CDATA[
        fastspar_reduce
            --correlation_table '$correlation_table'
            --pvalue_table '$pvalue_table'
            --correlation $correlation
            --pvalue $pvalue
            --output_prefix sparse
    ]]></command>
    <inputs>
        <param argument="--correlation_table" type="data" format="tabular" label="Correlation table"/>
        <param argument="--pvalue_table" type="data" format="tabular" label="P-value table"/>
        <param argument="--correlation" type="float" min="0.0" max="1.0" value="0.10" label="Absolute correlation threshold"/>
        <param argument="--pvalue" type="float" min="0.0" max="1.0" value="0.05" label="P-value threshold"/>
    </inputs>
    <outputs>
        <data name="correlations" format="tabular" from_work_dir="sparse_filtered_correlation.tsv" label="${tool.name} on ${on_string}: filtered_correlations.tsv">
            <actions>
                <action type="metadata" name="column_names" default="column,row,value" />
            </actions>
        </data>
        <data name="pvalues" format="tabular" from_work_dir="sparse_filtered_pvalue.tsv" label="${tool.name} on ${on_string}: filtered_pvalues.tsv">
            <actions>
                <action type="metadata" name="column_names" default="column,row,value" />
            </actions>
        </data>
    </outputs>
    <tests>
        <test expect_num_outputs="2">
            <param name="correlation_table" ftype="tabular" value="fake_data_cor.tsv"/>
            <param name="pvalue_table" ftype="tabular" value="pvalues.tsv"/>
            <output name="correlations" file="filtered_correlations.tsv" compare="diff"/>
            <output name="pvalues" file="filtered_pvalues.tsv" compare="diff"/>
        </test>
    </tests>
    <help><![CDATA[
What it does
============

This tool filters pairwise correlations and p-values from FastSpar outputs to generate sparse matrices suitable for network construction or visualization. It is typically used as the final step in a FastSpar pipeline.

Filtering Criteria
==================

- **Absolute correlation threshold**: Only retain OTU pairs whose absolute correlation exceeds this value.
- **P-value threshold**: Only retain OTU pairs whose empirical p-value is below this cutoff.

Both conditions must be satisfied (logical AND).

Required Inputs
===============

- **Correlation table**: A symmetric matrix from FastSpar.
- **P-value table**: A matching symmetric matrix from FastSpar p-value estimation.

Generated Outputs
=================

- `filtered_correlations.tsv`: Correlation values that passed both thresholds.
- `filtered_pvalues.tsv`: Matching p-values for retained entries.

Notes
=====

- Both input matrices must have identical dimensions and OTU order.
- The output tables are still symmetric and retain all diagonal values (e.g., self-correlations).

Additional Resources
====================

- FastSpar GitHub: https://github.com/scwatts/fastspar
    ]]></help>
    <expand macro="citations"/>
</tool>