1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/pair_cores/html.htm	Sat Mar 26 10:48:11 2011 +0300
     1.3 @@ -0,0 +1,1 @@
     1.4 +../blocks3d/html.htm
     1.5 \ No newline at end of file

     2.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     2.2 +++ b/pair_cores/html.py	Sat Mar 26 10:48:11 2011 +0300
     2.3 @@ -0,0 +1,1 @@
     2.4 +../blocks3d/html.py
     2.5 \ No newline at end of file

     3.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     3.2 +++ b/pair_cores/mkhtml.py	Sat Mar 26 10:48:11 2011 +0300
     3.3 @@ -0,0 +1,16 @@
     3.4 +#!/usr/bin/python
     3.5 +
     3.6 +import os
     3.7 +import json
     3.8 +import sys
     3.9 +
    3.10 +from protein_pdb import Alignment, Block, Monomer, Sequence
    3.11 +from html import html_template
    3.12 +
    3.13 +fasta_file = sys.argv[1]
    3.14 +blocks_file = sys.argv[2]
    3.15 +
    3.16 +alignment = Alignment().append_file(open(fasta_file))
    3.17 +blocks = alignment.blocks_from_file(open(blocks_file))
    3.18 +alignment.blocks_to_html(sys.stdout, blocks, open(html_template).read())
    3.19 +

     4.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     4.2 +++ b/pair_cores/pair_cores.py	Sat Mar 26 10:48:11 2011 +0300
     4.3 @@ -0,0 +1,85 @@
     4.4 +#!/usr/bin/python
     4.5 +
     4.6 +from gzip import GzipFile
     4.7 +import os
     4.8 +import sys
     4.9 +from copy import deepcopy
    4.10 +import argparse
    4.11 +from copy import copy
    4.12 +
    4.13 +from protein_pdb import Alignment, Block, Monomer, Sequence
    4.14 +from allpy import processors
    4.15 +import allpy.base
    4.16 +from html import html_template
    4.17 +
    4.18 +def remove_monomers_without_structure(block):
    4.19 +    monomers_with_structure = set()
    4.20 +
    4.21 +    for sequence in block.sequences:
    4.22 +        for monomer in sequence:
    4.23 +            if hasattr(monomer, 'ca_xyz'):
    4.24 +                monomers_with_structure.add(monomer)
    4.25 +
    4.26 +    for sequence in block.sequences:
    4.27 +        to_delete = []
    4.28 +        for i, monomer in enumerate(sequence):
    4.29 +            if monomer not in monomers_with_structure:
    4.30 +                to_delete.append(i)
    4.31 +        for n, i in enumerate(to_delete):
    4.32 +            assert not hasattr(sequence.pop(i-n), 'ca_xyz')
    4.33 +
    4.34 +    for column in block.columns:
    4.35 +        for s, m in column.items():
    4.36 +            if m not in monomers_with_structure:
    4.37 +                assert not hasattr(column.pop(s), 'ca_xyz')
    4.38 +
    4.39 +r = argparse.FileType('r')
    4.40 +w = argparse.FileType('w')
    4.41 +
    4.42 +p = argparse.ArgumentParser(
    4.43 +description='PairCores',
    4.44 +formatter_class=argparse.ArgumentDefaultsHelpFormatter,
    4.45 +)
    4.46 +
    4.47 +p.add_argument('-v','--version',action='version',version='%(prog)s 2.0')
    4.48 +p.add_argument('-i',help='Input alignment file',metavar='FILE',type=r,required=True)
    4.49 +p.add_argument('-d',help='Distance spreading',metavar='float',type=float,default=2.0)
    4.50 +p.add_argument('-o',help='Output alignment file',metavar='FILE',type=w, required=True)
    4.51 +p.add_argument('-b',help='Output pair_cores file',metavar='FILE',type=w, required=True)
    4.52 +p.add_argument('-H',help='Output HTML file',metavar='FILE',type=w)
    4.53 +
    4.54 +args = p.parse_args()
    4.55 +
    4.56 +try:
    4.57 +    alignment = Alignment().append_file(args.i)
    4.58 +except:
    4.59 +    raise Exception()
    4.60 +
    4.61 +unique_sequences = list(set([s.name for s in alignment.sequences]))
    4.62 +if len(unique_sequences) < len(alignment.sequences):
    4.63 +    alignment.sequences = unique_sequences
    4.64 +
    4.65 +bad_sequences = set()
    4.66 +for sequence in copy(alignment.sequences):
    4.67 +    try:
    4.68 +        sequence.auto_pdb(xyz_only=True)
    4.69 +    except:
    4.70 +        bad_sequences.add(sequence)
    4.71 +if bad_sequences:
    4.72 +    alignment.sequences = list(set(alignment.sequences)-bad_sequences)
    4.73 +
    4.74 +if len(alignment.sequences) < 2:
    4.75 +    raise Exception()
    4.76 +
    4.77 +block = Block.from_alignment(alignment)
    4.78 +
    4.79 +remove_monomers_without_structure(block)
    4.80 +
    4.81 +block.to_file(args.o)
    4.82 +
    4.83 +blocks = block.pair_core_parts(max_delta=args.d, timeout=0)
    4.84 +block.blocks_to_file(args.b, blocks)
    4.85 +
    4.86 +if args.H:
    4.87 +    block.blocks_to_html(args.H, blocks, open(html_template).read())
    4.88 +

     5.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     5.2 +++ b/pair_cores/protein_pdb.py	Sat Mar 26 10:48:11 2011 +0300
     5.3 @@ -0,0 +1,1 @@
     5.4 +../blocks3d/protein_pdb.py
     5.5 \ No newline at end of file

     6.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     6.2 +++ b/pair_cores/score.py	Sat Mar 26 10:48:11 2011 +0300
     6.3 @@ -0,0 +1,35 @@
     6.4 +#!/usr/bin/python
     6.5 +
     6.6 +import os
     6.7 +import sys
     6.8 +
     6.9 +from protein_pdb import Alignment, Block, Monomer, Sequence
    6.10 +
    6.11 +in_file = sys.argv[1]
    6.12 +pair_cores_file = sys.argv[2]
    6.13 +
    6.14 +alignment = Alignment().append_file(open(in_file))
    6.15 +pair_cores = alignment.blocks_from_file(open(pair_cores_file))
    6.16 +
    6.17 +score = 0.0
    6.18 +max_score = 0.0
    6.19 +
    6.20 +column2blocks = {}
    6.21 +for column in alignment.columns:
    6.22 +    column2blocks[column] = []
    6.23 +for b in pair_cores:
    6.24 +    for column in b.columns:
    6.25 +        column2blocks[column].append(b)
    6.26 +for column in alignment.columns:
    6.27 +    pairs_set = set()
    6.28 +    for b in column2blocks[column]:
    6.29 +        pairs_set.add(frozenset(b.sequences))
    6.30 +    m = len(column)
    6.31 +    if m > 1:
    6.32 +        c = m*(m-1) / 2
    6.33 +        max_score += c
    6.34 +        score += len(pairs_set)
    6.35 +
    6.36 +print("%f = %i / %i     %s %s" % (float(score)/max_score, score, max_score, in_file, pair_cores_file))
    6.37 +
    6.38 +