annotate matchms_similarity_wrapper.py @ 1:f680068b7863 draft

planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
author recetox
date Fri, 22 Jul 2022 16:48:21 +0000
parents 30e680e555d4
children 67ed0d4f3724
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
1 import argparse
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
2 import sys
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
3
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
4 import numpy as np
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
5 from matchms import calculate_scores
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
6 from matchms.importing import load_from_mgf, load_from_msp
1
f680068b7863 planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
recetox
parents: 0
diff changeset
7 from matchms.similarity import CosineGreedy, CosineHungarian, MetadataMatch, ModifiedCosine, NeutralLossesCosine
0
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
8 from pandas import DataFrame
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
9
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
10
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
11 def convert_precursor_mz(spectrum):
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
12 """
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
13 Check the presence of precursor m/z since it is needed for ModifiedCosine similarity metric. Convert to float if
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
14 needed, raise error if missing.
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
15 """
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
16
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
17 if "precursor_mz" in spectrum.metadata:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
18 metadata = spectrum.metadata
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
19 metadata["precursor_mz"] = float(metadata["precursor_mz"])
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
20 spectrum.metadata = metadata
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
21 return spectrum
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
22 else:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
23 raise ValueError("Precursor_mz missing. Apply 'add_precursor_mz' filter first.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
24
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
25
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
26 def main(argv):
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
27 parser = argparse.ArgumentParser(description="Compute MSP similarity scores")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
28 parser.add_argument("-r", dest="ri_tolerance", type=float, help="Use RI filtering with given tolerance.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
29 parser.add_argument("-s", dest="symmetric", action='store_true', help="Computation is symmetric.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
30 parser.add_argument("--ref", dest="references_filename", type=str, help="Path to reference spectra library.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
31 parser.add_argument("--ref_format", dest="references_format", type=str, help="Reference spectra library file format.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
32 parser.add_argument("queries_filename", type=str, help="Path to query spectra.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
33 parser.add_argument("queries_format", type=str, help="Query spectra file format.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
34 parser.add_argument("similarity_metric", type=str, help='Metric to use for matching.')
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
35 parser.add_argument("tolerance", type=float, help="Tolerance to use for peak matching.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
36 parser.add_argument("mz_power", type=float, help="The power to raise mz to in the cosine function.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
37 parser.add_argument("intensity_power", type=float, help="The power to raise intensity to in the cosine function.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
38 parser.add_argument("output_filename_scores", type=str, help="Path where to store the output .tsv scores.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
39 parser.add_argument("output_filename_matches", type=str, help="Path where to store the output .tsv matches.")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
40 args = parser.parse_args()
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
41
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
42 if args.queries_format == 'msp':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
43 queries_spectra = list(load_from_msp(args.queries_filename))
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
44 elif args.queries_format == 'mgf':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
45 queries_spectra = list(load_from_mgf(args.queries_filename))
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
46 else:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
47 raise ValueError(f'File format {args.queries_format} not supported for query spectra.')
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
48
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
49 if args.symmetric:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
50 reference_spectra = []
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
51 else:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
52 if args.references_format == 'msp':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
53 reference_spectra = list(load_from_msp(args.references_filename))
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
54 elif args.references_format == 'mgf':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
55 reference_spectra = list(load_from_mgf(args.references_filename))
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
56 else:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
57 raise ValueError(f'File format {args.references_format} not supported for reference spectra library.')
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
58
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
59 if args.similarity_metric == 'CosineGreedy':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
60 similarity_metric = CosineGreedy(args.tolerance, args.mz_power, args.intensity_power)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
61 elif args.similarity_metric == 'CosineHungarian':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
62 similarity_metric = CosineHungarian(args.tolerance, args.mz_power, args.intensity_power)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
63 elif args.similarity_metric == 'ModifiedCosine':
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
64 similarity_metric = ModifiedCosine(args.tolerance, args.mz_power, args.intensity_power)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
65 reference_spectra = list(map(convert_precursor_mz, reference_spectra))
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
66 queries_spectra = list(map(convert_precursor_mz, queries_spectra))
1
f680068b7863 planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
recetox
parents: 0
diff changeset
67 elif args.similarity_metric == 'NeutralLossesCosine':
f680068b7863 planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
recetox
parents: 0
diff changeset
68 similarity_metric = NeutralLossesCosine(args.tolerance, args.mz_power, args.intensity_power)
f680068b7863 planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
recetox
parents: 0
diff changeset
69 reference_spectra = list(map(convert_precursor_mz, reference_spectra))
f680068b7863 planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit ca44513d50b29a4706e2a2db96c23ef6688b7c2d
recetox
parents: 0
diff changeset
70 queries_spectra = list(map(convert_precursor_mz, queries_spectra))
0
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
71 else:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
72 return -1
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
73
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
74 print("Calculating scores...")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
75 scores = calculate_scores(
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
76 references=queries_spectra if args.symmetric else reference_spectra,
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
77 queries=queries_spectra,
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
78 similarity_function=similarity_metric,
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
79 is_symmetric=args.symmetric
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
80 )
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
81
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
82 if args.ri_tolerance is not None:
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
83 print("RI filtering with tolerance ", args.ri_tolerance)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
84 ri_matches = calculate_scores(reference_spectra, queries_spectra, MetadataMatch("retention_index", "difference", args.ri_tolerance)).scores
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
85 scores.scores["score"] = np.where(ri_matches, scores.scores["score"], 0.0)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
86
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
87 write_outputs(args, scores)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
88 return 0
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
89
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
90
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
91 def write_outputs(args, scores):
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
92 print("Storing outputs...")
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
93 query_names = [spectra.metadata['compound_name'] for spectra in scores.queries]
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
94 reference_names = [spectra.metadata['compound_name'] for spectra in scores.references]
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
95
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
96 # Write scores to dataframe
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
97 dataframe_scores = DataFrame(data=[entry["score"] for entry in scores.scores], index=reference_names, columns=query_names)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
98 dataframe_scores.to_csv(args.output_filename_scores, sep='\t')
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
99
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
100 # Write number of matches to dataframe
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
101 dataframe_matches = DataFrame(data=[entry["matches"] for entry in scores.scores], index=reference_names, columns=query_names)
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
102 dataframe_matches.to_csv(args.output_filename_matches, sep='\t')
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
103
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
104
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
105 if __name__ == "__main__":
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
106 main(argv=sys.argv[1:])
30e680e555d4 "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 4d2ac914c951166e386a94d8ebb8cb1becfac122"
recetox
parents:
diff changeset
107 pass