Mercurial > repos > bgruening > chemfp

annotate nxn_clustering.py @ 12:3b14765c22ee draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
author bgruening
date Sat, 25 Sep 2021 19:07:44 +0000
parents 198b1e30c739
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
1 # !/usr/bin/env python
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
2 """
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
3 Modified version of code examples from the chemfp project.
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
4 http://code.google.com/p/chem-fingerprints/
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
5 Thanks to Andrew Dalke of Andrew Dalke Scientific!
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
6 """
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
7
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
8 import argparse
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
9
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
10 import chemfp
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
11 import matplotlib
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
12 matplotlib.use("Agg") # noqa
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
13 from matplotlib import rcParams # noqa
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
14 rcParams.update({"figure.autolayout": True}) # noqa
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
15 import numpy # noqa
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
16 import pylab # noqa
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
17 import scipy.cluster.hierarchy as hcluster # noqa
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
18
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
19
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
20 def distance_matrix(arena, tanimoto_threshold=0.0):
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
21 n = len(arena)
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
22 # Start off a similarity matrix with 1.0s along the diagonal
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
23 try:
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
24 similarities = numpy.identity(n, "d")
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
25 except Exception:
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
26 raise Exception("Input dataset is to large!")
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
27 chemfp.set_num_threads(args.processors)
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
28
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
29 # Compute the full similarity matrix.
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
30 # The implementation computes the upper-triangle then copies
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
31 # the upper-triangle into lower-triangle. It does not include
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
32 # terms for the diagonal.
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
33 results = chemfp.search.threshold_tanimoto_search_symmetric(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
34 arena, threshold=tanimoto_threshold
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
35 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
36
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
37 # Copy the results into the NumPy array.
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
38 for row_index, row in enumerate(results.iter_indices_and_scores()):
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
39 for target_index, target_score in row:
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
40 similarities[row_index, target_index] = target_score
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
41
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
42 # Return the distance matrix using the similarity matrix
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
43 return 1.0 - similarities
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
44
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
45
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
46 if __name__ == "__main__":
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
47 parser = argparse.ArgumentParser(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
48 description="""NxN clustering for fps files.
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
49 For more details please see the chemfp documentation:
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
50 https://chemfp.readthedocs.org
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
51 """
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
52 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
53
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
54 parser.add_argument(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
55 "-i",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
56 "--input",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
57 dest="input_path",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
58 required=True,
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
59 help="Path to the input file.",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
60 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
61
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
62 parser.add_argument(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
63 "-c",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
64 "--cluster",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
65 dest="cluster_image",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
66 help="Path to the output cluster image.",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
67 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
68
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
69 parser.add_argument(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
70 "-s",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
71 "--smatrix",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
72 dest="similarity_matrix",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
73 help="Path to the similarity matrix output file.",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
74 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
75
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
76 parser.add_argument(
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
77 "-t",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
78 "--threshold",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
79 dest="tanimoto_threshold",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
80 type=float,
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
81 default=0.0,
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
82 help="Tanimoto threshold [0.0]",
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
83 )
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
84
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
85 parser.add_argument("--oformat", default="png", help="Output format (png, svg)")
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
86
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
87 parser.add_argument("-p", "--processors", type=int, default=4)
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
88
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
89 args = parser.parse_args()
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
90
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
91 targets = chemfp.open(args.input_path, format="fps")
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
92 arena = chemfp.load_fingerprints(targets)
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
93 distances = distance_matrix(arena, args.tanimoto_threshold)
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
94
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
95 if args.similarity_matrix:
7
0d88631bb7de planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit ed9b6859de648aa5f7cde483732f5df20aaff90e
bgruening
parents: 2
diff changeset
96 numpy.savetxt(args.similarity_matrix, distances)
2
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
97
70b071de9bee planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
bgruening
parents:
diff changeset
98 if args.cluster_image:
7
0d88631bb7de planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit ed9b6859de648aa5f7cde483732f5df20aaff90e
bgruening
parents: 2
diff changeset
99 linkage = hcluster.linkage(distances, method="single", metric="euclidean")
12
3b14765c22ee "planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit 7fb96a3844b4771084f18de2346ed6d5e241d839"
bgruening
parents: 8
diff changeset
100 hcluster.dendrogram(linkage, labels=arena.ids, leaf_rotation=90.0)
7
0d88631bb7de planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/chemfp commit ed9b6859de648aa5f7cde483732f5df20aaff90e
bgruening
parents: 2
diff changeset
101 pylab.savefig(args.cluster_image, format=args.oformat)