view matchms_fingerprint_similarity.xml @ 9:1d3ee4f9f63d draft

planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit b797ce0fea5cb3d9a0cf0ec31a21f9b73f112b4e
author recetox
date Mon, 15 Jan 2024 12:31:52 +0000
parents 79accd21387c
children 5be32c7ce669
line wrap: on
line source

<tool id="matchms_fingerprint_similarity" name="matchms fingerprint similarity" version="@TOOL_VERSION@+galaxy0" profile="21.09">
    <description>calculate similarity between molecular fingerprints calculated from structural spectrum metadata descriptors</description>

    <macros>
        <import>macros.xml</import>
        <import>help.xml</import>
    </macros>
    <expand macro="creator"/>
    <edam_operations>
        <edam_operation>operation_0360</edam_operation>
    </edam_operations>
    <expand macro="bio.tools"/>

    <requirements>
        <requirement type="package" version="@TOOL_VERSION@">matchms</requirement>
    </requirements>


    <command detect_errors="exit_code"><![CDATA[
        python3 ${python_wrapper}
    ]]> </command>

<configfiles>
<configfile name="python_wrapper">
@init_logger@

from matchms.similarity import FingerprintSimilarity
similarity = FingerprintSimilarity(similarity_measure="${fingerprint_similarity}")
@init_scores@

from matchms.filtering import add_fingerprint
import numpy as np

name="FingerprintSimilarity_${fingerprint_similarity}"

layer = similarity.matrix(
    references=np.asarray(list(map(add_fingerprint, scores.references))),
    queries=np.asarray(list(map(add_fingerprint, scores.queries))),
    array_type = "numpy",
    is_symmetric=scores.is_symmetric)

scores._scores.add_dense_matrix(layer, name)
scores.to_json("$scores_out")
</configfile>
</configfiles>

    <inputs>
        <expand macro="input_param"/>
        <param label="Fingerprint similarity distance" name="fingerprint_similarity" type="select" display="radio"
            help="Distance measure to use to compute between the molecular fingerprints - see [2] for details regarding the used fingerprint.">
            <option value="jaccard" selected="true">jaccard</option>
            <option value="dice">dice</option>
            <option value="cosine">cosine</option>
        </param>
    </inputs>

    <outputs>
        <data label="$fingerprint_similarity distance between fingerprints in ${on_string}" name="scores_out" format="json"/>
    </outputs>

    <tests>
        <test>
            <param name="queries" value="similarity/rcx_gc-ei_ms_subset.msp" ftype="msp"/>
            <param name="references" value="similarity/rcx_exposome_pesticides_subset.msp" ftype="msp"/>
            <param name="fingerprint_similarity" value="dice" />
            <output name="scores_out" value="fingerprints/dice.json" ftype="json" />
        </test>
        <test>
            <param name="queries" value="similarity/rcx_gc-ei_ms_subset.msp" ftype="msp"/>
            <param name="references" value="similarity/rcx_exposome_pesticides_subset.msp" ftype="msp"/>
            <param name="fingerprint_similarity" value="cosine" />
            <output name="scores_out" value="fingerprints/cosine.json" ftype="json" />
        </test>
        <test>
            <param name="use_scores" value="True"/>
            <param name="scores_in" value="metadata_match/inner_join_ri_20.json" ftype="json"/>
            <param name="fingerprint_similarity" value="jaccard" />
            <output name="scores_out" value="fingerprints/jaccard_scores.json" ftype="json" />
        </test>
    </tests>

    <help>
        Description
            Molecular fingerprints are a digital representation of a molecule based on specific patterns, such as substructures or physicochemical properties.
            They usually have a fixed length with each entry representing the presence or absence of a certain attribute.
            Similarity between molecular fingerprints can serve as a proxy for structural similarity and can therefore be used to compare molecules.

        .. rubric:: **Footnotes**
        .. [1] SQL join types explained on LearnSQL_.
        .. [2] Fingerprint - the `daylight fingerprint`_ is used to compute chemical similarity.
               Fingerprints are derived from SMILES or InChI structure notations present in the spectrum metadata.
        
        .. _LearnSQL: https://learnsql.com/blog/sql-joins-types-explained/
        .. _daylight fingerprint: https://www.daylight.com/dayhtml/doc/theory/theory.finger.html
        
        @HELP_matchms@
    </help>

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