<tool id="spec2vec_similarity" name="spec2vec similarity" version="@TOOL_VERSION@+galaxy0" profile="21.09">
    <macros>
        <import>macros.xml</import>
    </macros>
    <expand macro="creator"/>
    <expand macro="edam" />

    <requirements>
        <requirement type="package" version="@TOOL_VERSION@">spec2vec</requirement>
    </requirements>

    <command detect_errors="aggressive"><![CDATA[
        python3 ${python_wrapper}
    ]]></command>
<configfiles>
<configfile name="python_wrapper">
@init_logger@
@init_model@

import numpy as np
from spec2vec import Spec2Vec

similarity = Spec2Vec(
    model,
    intensity_weighting_power=${intensity_power},
    allowed_missing_percentage=${allow_missing_percentage} * 100
)
name="Spec2Vec_${intensity_power}_${allow_missing_percentage}"

@init_scores@

from matchms.filtering import normalize_intensities

layer = similarity.sparse_array(
    references=np.asarray(list(map(normalize_intensities, scores.references))),
    queries=np.asarray(list(map(normalize_intensities, scores.queries))),
    idx_row = scores._scores.row,
    idx_col = scores._scores.col,
    is_symmetric=False)

scores._scores.add_sparse_data(scores._scores.row, scores._scores.col, layer, name)

scores.filter_by_range(inplace=True, name=name, low=0)
scores.to_json("$similarity_scores")
</configfile>
</configfiles>

    <inputs>
        <expand macro="input_param" />
        <param label="Model JSON file" name="model_metadata" type="data" format="json"
            help="Model JSON file to use for Spec2Vec similarity computing."/>
        <param label="Model NPY file" name="model_weights" type="data" format="binary"
            help="Model NPY file to use for Spec2Vec similarity computing."/>
        <param label="intensity_power" name="intensity_power" type="float" value="0.0"
            help="Spectrum vectors are a weighted sum of the word vectors. The given word intensities will be raised to the given power. 
            The default is 0, which means that no weighing will be done."/>
        <param label="Maximum share of new peaks" name="allow_missing_percentage"  type="float" value="0.1" max="1.0" min="0.0"
            help="Maximum allowed share of the peaks that are new to the model in relation to the whole peak corpus."/>
    </inputs>
    <outputs>
        <data label="Spec2Vec scores of ${on_string}" name="similarity_scores" format="json"/>
    </outputs>

    <tests>
        <test> <!-- TEST #1: Test Spec2Vec. -->
            <param name="references" value="inp_filtered_library.msp" ftype="msp"/>
            <param name="queries" value="inp_filtered_spectra.msp" ftype="msp"/>
            <param name="model_metadata" value="model_100.json" ftype="json"/>
            <param name="model_weights" value="weights_100.binary" ftype="auto"/>
            <param name="allow_missing_percentage" value="1.0"/>
            <output name="similarity_scores" file="s2v_scores_test1_out.json" ftype="json"/>
        </test>
        <test>
            <param name="use_scores" value="True"/>
            <param name="scores_in" value="ri_match_60.json" ftype="json"/>
            <param name="model_metadata" value="model_100.json" ftype="json"/>
            <param name="model_weights" value="weights_100.binary" ftype="auto"/>
            <param name="allow_missing_percentage" value="1.0"/>
            <output name="similarity_scores" value="s2v_scores_test2_out.json" ftype="json" />
        </test>
    </tests>

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