def baseline_reverse_hamming(term, ri, theta, esquare, xloc, xlen, posx, lenfilter, overlapx):
postings = ri.get(term)
if postings == None:
return
for y, posy in reversed(postings):
ytext = y[3]
ylen = len(ytext)
if ylen < lenfilter:
return
if distsquare(xloc, y[2]) > esquare:
continue
yid = y[1]
o = overlapx.get(yid, (ytext, 0))
if o == None:
continue
o = o[1]
alpha = int(ceil(theta*float(xlen+ylen)/(1+theta)))
ubound = min(xlen-posx, ylen-posy)
if o + ubound < alpha:
overlapx[yid] = None
continue
hmax = xlen + ylen - 2*alpha
if (posx-o) + (posy-o) + abs((xlen-posx) - (ylen-posy)) > hmax:
overlapx[yid] = None
continue
overlapx[yid] = (ytext, o+1)
def baseline_reverse_hamming(term, ri, theta, esquare, xloc, xlen, posx,
lenfilter, overlapx):
postings = ri.get(term)
if postings == None:
return
for y, posy in reversed(postings):
ytext = y[3]
ylen = len(ytext)
if ylen < lenfilter:
return
if distsquare(xloc, y[2]) > esquare:
continue
yid = y[1]
o = overlapx.get(yid, (ytext, 0))
if o == None:
continue
o = o[1]
alpha = int(ceil(theta * float(xlen + ylen) / (1 + theta)))
ubound = min(xlen - posx, ylen - posy)
if o + ubound < alpha:
overlapx[yid] = None
continue
hmax = xlen + ylen - 2 * alpha
if (posx - o) + (posy - o) + abs((xlen - posx) - (ylen - posy)) > hmax:
overlapx[yid] = None
continue
overlapx[yid] = (ytext, o + 1)
def grouping(theta, esquare, xlen, xprefix, xobjs, overlapx, counter):
px = len(xprefix)
wx = xprefix[-1]
for o in overlapx.itervalues():
if o == None:
continue
y = o[0]
o = o[1]
ylen = y[0][0]
alpha = int(ceil(theta*float(xlen+ylen)/(1+theta)))
yprefix = y[0][2]
py = min(len(yprefix), ylen - int(ceil(2*theta*float(ylen)/(theta+1))) + 1)
wy = yprefix[py-1]
y = y[2]
if wx < wy:
ubound = o + xlen - px
py = o
elif wx > wy:
ubound = o + ylen - py
px = o
else:
ubound = o + min(xlen-px, ylen-py)
if ubound < alpha:
continue
for xo in xobjs:
for yo in y:
if distsquare(xo[1], yo[1]) > esquare:
continue
counter.add(1)
if o + overlap(xo[2], px, xlen, yo[2], py, ylen) >= alpha:
yield (xo[0], yo[0])
def grouping_self(theta, esquare, xlen, xprefixlen, xobjs, counter):
if len(xobjs) <= 1:
return
alpha = int(ceil(theta*float(2*xlen)/(1+theta)))
for i in xrange(0, len(xobjs) - 1):
xo = xobjs[i]
for j in xrange(i+1, len(xobjs)):
yo = xobjs[j]
if distsquare(xo[1], yo[1]) > esquare:
continue
counter.add(1)
if xprefixlen + overlap(xo[2], xprefixlen, xlen, yo[2], xprefixlen, xlen) >= alpha:
yield (xo[0], yo[0])
def baseline(term, ri, theta, esquare, xloc, xlen, posx, lenfilter, overlapx):
for y, posy in ri.get(term, []):
ytext = y[3]
ylen = len(ytext)
if ylen < lenfilter or distsquare(xloc, y[2]) > esquare:
continue
yid = y[1]
o = overlapx.get(yid, (ytext, 0))
if o == None:
continue
o = o[1]
alpha = int(ceil(theta * float(xlen + ylen) / (1 + theta)))
ubound = min(xlen - posx, ylen - posy)
overlapx[yid] = (ytext, o + 1) if o + ubound >= alpha else None
def extra_pos(term, ri, theta, esquare, xloc, xlen, posx, lenfilter, overlapx):
for y, posy in ri.get(term, []):
ytext = y[3]
ylen = len(ytext)
if ylen < lenfilter or distsquare(xloc, y[2]) > esquare:
continue
yid = y[1]
o = overlapx.get(yid, (ytext, 0, posx, posy))
if o == None:
continue
o = o[1]
alpha = int(ceil(theta*float(xlen+ylen)/(1+theta)))
ubound = min(xlen-posx, ylen-posy)
overlapx[yid] = (ytext, o+1, posx, posy) if o + ubound >= alpha else None
def baseline_reverse(term, ri, theta, esquare, xloc, xlen, posx, lenfilter, overlapx):
postings = ri.get(term)
if postings == None:
return
for y, posy in reversed(postings):
ytext = y[3]
ylen = len(ytext)
if ylen < lenfilter:
return
if distsquare(xloc, y[2]) > esquare:
continue
yid = y[1]
o = overlapx.get(yid, (ytext, 0))
if o == None:
continue
o = o[1]
alpha = int(ceil(theta*float(xlen+ylen)/(1+theta)))
ubound = min(xlen-posx, ylen-posy)
overlapx[yid] = (ytext, o+1) if o + ubound >= alpha else None