Mercurial > repos > recetox > matchms

annotate matchms_wrapper.py @ 1:4aecfd6b319b draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
author recetox
date Wed, 17 Mar 2021 11:40:17 +0000
parents 6a736abe431f
children a7c9fc186f8c
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
1 import argparse
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
2 import sys
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
3
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
4 from matchms import calculate_scores
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
5 from matchms.importing import load_from_msp
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
6 from matchms.similarity import (
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
7 CosineGreedy,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
8 CosineHungarian,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
9 FingerprintSimilarity,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
10 IntersectMz,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
11 ModifiedCosine,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
12 ParentMassMatch
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
13 )
1
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
14 from pandas import DataFrame
0
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
15
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
16
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
17 def main(argv):
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
18 parser = argparse.ArgumentParser(description="Compute MSP similarity scores")
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
19 parser.add_argument(
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
20 "references_filename", type=str, help="Path to reference MSP library."
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
21 )
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
22 parser.add_argument("queries_filename", type=str, help="Path to query spectra.")
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
23 parser.add_argument("similarity_metric", type=str, help='Metric to use for matching.')
1
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
24 parser.add_argument("output_filename_scores", type=str, help="Path where to store the output .csv scores.")
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
25 parser.add_argument("output_filename_matches", type=str, help="Path where to store the output .csv matches.")
0
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
26
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
27 args = parser.parse_args()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
28
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
29 if args.similarity_metric == 'CosineGreedy':
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
30 similarity_metric = CosineGreedy()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
31 elif args.similarity_metric == 'CosineHungarian':
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
32 similarity_metric = CosineHungarian()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
33 elif args.similarity_metric == 'FingerprintSimilarity':
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
34 similarity_metric = FingerprintSimilarity()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
35 elif args.similarity_metric == 'IntersectMz':
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
36 similarity_metric = IntersectMz()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
37 elif args.similarity_metric == 'ModifiedCosine':
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
38 similarity_metric = ModifiedCosine()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
39 else:
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
40 similarity_metric = ParentMassMatch()
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
41
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
42 reference_spectra = [
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
43 spectrum for spectrum in load_from_msp(args.references_filename)
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
44 ]
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
45 queries_spectra = [spectrum for spectrum in load_from_msp(args.queries_filename)]
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
46
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
47 scores = calculate_scores(
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
48 references=reference_spectra,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
49 queries=queries_spectra,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
50 similarity_function=similarity_metric,
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
51 )
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
52
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
53 query_names = [spectra.metadata['name'] for spectra in scores.queries]
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
54 reference_names = [spectra.metadata['name'] for spectra in scores.references]
1
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
55
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
56 # Write scores to dataframe
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
57 dataframe_scores = DataFrame(data=[entry["score"] for entry in scores.scores], index=reference_names, columns=query_names)
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
58 dataframe_scores.to_csv(args.output_filename_scores, sep=';')
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
59
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
60 # Write number of matches to dataframe
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
61 dataframe_matches = DataFrame(data=[entry["matches"] for entry in scores.scores], index=reference_names, columns=query_names)
4aecfd6b319b "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit d110cb008c3703945fe3718465de36278fa34652"
recetox
parents: 0
diff changeset
62 dataframe_matches.to_csv(args.output_filename_matches, sep=';')
0
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
63 return 0
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
64
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
65
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
66 if __name__ == "__main__":
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
67 main(argv=sys.argv[1:])
6a736abe431f "planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/matchms commit 25fba33c37b26a6e9321f043c61b3c0dd392a53c"
recetox
parents:
diff changeset
68 pass