find_match.py

# -*- coding: utf-8 -*-
"""
Created on Wed Jun 22 17:30:57 2016

@author: Krishna
"""
import re
from fgm import *
from rules1 import *
def find_matches(s, gmk):
    
    if gmk in s:  # checking if gmk is in the line
        l = re.split('\s|(?<!\d)[,.]|[,.](?!\d)|;|[()]|-', s) # split the line by comma, semicolon and space to check for gmks and gs. Also http://goo.gl/RPQNbT
        filter(None, l)       # remove empty elements in the list   
        gs_list=[] 
        for gss in keywords:
                gss=gss.rstrip()
                gss=gss.lower()
                if gss in s:
                    gs_list.append(gss)              
        for gs in gs_list: # gene symbols
            
            if gs in s: # search for GS in line. using 'gs in s' led to a lot of partial word matches
                gs1 = re.split('\s|(?<!\d)[,.]|[,.](?!\d)|;|-', gs)
                gs1=filter(None, gs1)
                gmk1 = re.split('\s|(?<!\d)[,.]|[,.](?!\d)|;|-', gmk)
                gmk1=filter(None, gmk1)
                if any(l[i:i+len(gs1)]==gs1 for i in xrange(len(l)-len(gs1)+1)) and (any(l[i:i+len(gmk1)]==gmk1 for i in xrange(len(l)-len(gmk1)+1))): # this ensures that both gs and gmk are in l, as a unit(i.e. and in order) otherwise it was detecting things like 'beta c' from beta cells
                    #  UPTO THIS POINT WE HAVE ESTABLISHED THAT THE GMK AND GS ARE INDEED IN THE LINE                    
                    k1 = '_MKKEYWORD_1_'
                    k2 = '_SKEYWORD_2_'
                    #print gmk
                    text = re.sub(re.escape(gmk), k1, s, flags=re.I) # because of this replacement, we dont have the problem of counting r from behind etc.
                                                # also, I cannot use the regex based replacement used below for gmk replacement because we do want 
                                                # cases where gmk's like -/- or + are just after or before a word, without the word boundary   
                    text = re.sub(r'(\b%s\b)' % (re.escape(gs)), k2, text, flags=re.I)
                 
                    lt = text.split()                    
                    #lt = re.split('\s|(?<!\d)[,.]|[,.](?!\d)|;|[()]|-', text)
                    d_idx = {k1:[], k2:[]}
                    #print d_idx[k1]
                    for k,v in enumerate(lt):
                        if k1 in v:
                            d_idx[k1].append(k)
                        if k2 in v:
                            d_idx[k2].append(k)
                    distance = 8
                    data = []
                    for idx1 in d_idx[k1]:
                        for idx2 in d_idx[k2]:
                            d = abs(idx1 - idx2)
                            if d<=distance:
                                data.append((d,idx1,idx2))
                                
                    data.sort(key=lambda x: x[0])
                    for i in range (0, len(data)):  
   
                        aq = data[i]
                        loq = min(aq[1], aq[2])
                        hiq = max(aq[1], aq[2])
                        brrq = lt[max(0, loq-6):hiq+6]
                        brq = " ".join(brrq) 
                        
                    if data: 
                        a1=int(cl(s, gmk, gs, gs_list, data))
                        
                        print 'cl for %s is %d' %(gs, a1)