Mercurial > repos > rnateam > rnacommender

annotate main.py @ 1:21130153e729 draft

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/rna_commander/tools/rna_tools/rna_commender commit d5a6f8df417e899092599e956e8033d502bea3c8
author rnateam
date Tue, 19 Jul 2016 07:20:29 -0400
parents 8918de535391
children a609d6dc8047
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
1 #!/usr/bin/env python
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
2 """Recommendation."""
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
3
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
4 import argparse
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
5 import sys
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
6 from rbpfeatures import RBPVectorizer
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
7 from data import PredictDataset
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
8 from recommend import Predictor
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
9
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
10 from theano import config
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
11
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
12 __author__ = "Gianluca Corrado"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
13 __copyright__ = "Copyright 2016, Gianluca Corrado"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
14 __license__ = "MIT"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
15 __maintainer__ = "Gianluca Corrado"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
16 __email__ = "gianluca.corrado@unitn.it"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
17 __status__ = "Production"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
18
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
19 config.floatX = 'float32'
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
20
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
21 if __name__ == '__main__':
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
22 parser = argparse.ArgumentParser(
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
23 description=__doc__,
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
24 formatter_class=argparse.ArgumentDefaultsHelpFormatter)
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
25 parser.add_argument('fasta', metavar='fasta', type=str,
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
26 help="""Fasta file containing the RBP \
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
27 sequences.""")
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
28
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
29 args = parser.parse_args()
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
30
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
31 v = RBPVectorizer(fasta=args.fasta)
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
32 rbp_fea = v.vectorize()
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
33
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
34 if rbp_fea is not None:
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
35 # Define and load dataset
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
36 D = PredictDataset(
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
37 fp=rbp_fea, fr="AURA_Human_data/RNA_features/HT_utrs.h5")
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
38 dataset = D.load()
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
39
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
40 model = "AURA_Human_data/model/trained_model.pkl"
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
41
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
42 # Define the Trainer and train the model
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
43 P = Predictor(predict_dataset=dataset,
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
44 trained_model=model,
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
45 serendipity_dic=model + '_',
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
46 output="output.txt")
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
47 P.predict()
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
48 else:
8918de535391 planemo upload for repository https://github.com/bgruening/galaxytools/tree/rna_commander/tools/rna_tools/rna_commender commit 2fc7f3c08f30e2d81dc4ad19759dfe7ba9b0a3a1
rnateam
parents:
diff changeset
49 sys.exit("""The queried protein has no domain similarity with the proteins in the training dataset. It cannot be predicted.""")