Mercurial > repos > rnateam > graphclust_preprocessing

annotate splitSHAPE.py @ 7:8634e06ae642 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/tools/GraphClust commit 4406735e44aba20859c252be39f4e99df28c7a92
author rnateam
date Sat, 27 Oct 2018 13:19:07 -0400
parents e31c659be8bc
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
5
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
1 import os
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
2 import re
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
3 import sys
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
4
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
5 shape_file = sys.argv[1]
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
6
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
7 pattern = re.compile("^>.*$")
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
8 toWrite = ""
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
9
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
10 count_for_id = 1
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
11 seq_counter = 0
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
12 new_id = ""
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
13
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
14 seq_id = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
15 seq_string = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
16 orig_id = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
17 name_file = "FASTA/data.names"
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
18 array_all_chunks = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
19 with open(name_file, 'r') as f:
6
e31c659be8bc planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 6767a5ffb02052c844e9d862c79912f998f39d8e
rnateam
parents: 5
diff changeset
20 for line in f:
e31c659be8bc planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 6767a5ffb02052c844e9d862c79912f998f39d8e
rnateam
parents: 5
diff changeset
21 if len(line.strip()) == 0:
e31c659be8bc planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 6767a5ffb02052c844e9d862c79912f998f39d8e
rnateam
parents: 5
diff changeset
22 continue
5
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
23 seq_id.append(int(line.split()[0]))
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
24 seq_string.append(line.split()[1])
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
25 orig_id_srt = line.split()[3]
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
26 orig_id_srt = orig_id_srt.rsplit('_',1)[0]
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
27 orig_id.append(orig_id_srt)
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
28
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
29
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
30 react_dict = {}
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
31 react_arr = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
32
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
33 with open(shape_file, 'r') as shape:
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
34 content = shape.read()
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
35 lines = content.split('\n')
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
36 for line in lines:
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
37 if pattern.match(line):
6
e31c659be8bc planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 6767a5ffb02052c844e9d862c79912f998f39d8e
rnateam
parents: 5
diff changeset
38 line = line.replace('>','').split()[0]
5
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
39 react_arr=[]
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
40 react_dict[line] = react_arr
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
41 continue
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
42 else:
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
43 react_arr.append(line)
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
44
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
45 toWrite = ""
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
46 chunks = []
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
47 for i in range(len(orig_id)):
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
48 if not orig_id[i] in react_dict:
6
e31c659be8bc planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 6767a5ffb02052c844e9d862c79912f998f39d8e
rnateam
parents: 5
diff changeset
49 raise RuntimeError('Error key {} {} not found'.format(i, orig_id[i]))
5
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
50
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
51 react_val = react_dict[orig_id[i]]
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
52 toWrite += '>' + str(seq_id[i]) + " " + seq_string[i] + "\n"
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
53 chunks = re.findall(r'\d+', seq_string[i])
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
54
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
55 for j in react_val[int(chunks[1])-1:int(chunks[2])]:
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
56 id_s = int(j.split()[0])
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
57
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
58
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
59 id_s = id_s - int(chunks[1]) + 1
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
60 toWrite += str(id_s) + '\t' + j.split()[1] + "\n"
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
61
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
62 with open("shape_data_split.react", 'w') as out:
f4ad5dceb619 planemo upload for repository https://github.com/eteriSokhoyan/galaxytools/tree/branchForIterations/tools/GraphClust commit 9a3dc91fa984be18fabc0d968360634d787c9589
rnateam
parents:
diff changeset
63 out.write(toWrite)