#!@PYTHON@
#
# from an original PDB and its DBOND file for a subset PDB (dest) file we create the respective DBOND file, i.e. containing only bonds between atoms
# in the subset

import sys, random, math, re
wrerr=sys.stderr.write
wrout=sys.stdout.write

# check arguments
if len(sys.argv) < 5:
	print "Usage: "+sys.argv[0]+" <src{PDB/XYZ}file> <destXYZfile> <offsetID> <offsetXYZ> <srcDBONDfile> [destDBONDfile]" 
	sys.exit(1)

EPSILON=1e-3
CUTOFF=2.
inputsrc = open(sys.argv[1], "r")
inputdestPDB = open(sys.argv[2], "r")
inputsrcDBOND = open(sys.argv[5], "r")
offsetID=int(sys.argv[3])
offsetXYZ=int(sys.argv[4])
entries = sys.argv[1].split(".")
suffix = entries[-1]
if len(sys.argv) > 6:
	output = open(sys.argv[6],"w")
else:
	output = open(sys.argv[5]+".new", "w")

# 1. first parse both PDB files into arrays (id, element, xyz) , therewhile scan BoundaryBoxes
max = [ 0., 0., 0. ]
min = [ 0., 0., 0. ]
x = [ 0., 0., 0. ]
nr=0
srcAtoms = []
#if "xyz" in suffix:
#	print "Scanning source XYZ file "+sys.argv[1]+"."
#	for line in inputsrc:
#		if (len(entries)<4 or (entries[0]!="O" and entries[0]!="Si" and entries[0]!="Ca")):
#			continue
#		entries = line.split()
#		for n in range(3):
#			x[n] = float(entries[offsetXYZ+n])
#			if x[n] > max[n]:
#				max[n] = x[n]
#			if x[n] < min[n]:
#				min[n] = x[n]
#		srcAtoms.append([int(entries[offsetID]), x[0], x[1], x[2]])
#		nr+=1
#elif "pdb" in suffix:
print "Scanning source PDB file "+sys.argv[1]+"."
for line in inputsrc:
	if "#" in line:
		continue
	if "END" in line:
		break
	if "ATOM" in line:
		entries = line.split()
		for n in range(3):
			x[n] = float(entries[offsetXYZ+n])
			if x[n] > max[n]:
				max[n] = x[n]
			if x[n] < min[n]:
				min[n] = x[n]
		srcAtoms.append([int(entries[offsetID]), x[0], x[1], x[2]])
		nr+=1
#else:
#	print "Cannot recognize input source file format: $suffix."
#	sys.exit(1)

inputsrc.close()
print "Scanned "+str(nr)+" source atoms."

print "Scanning destination XYZ file "+sys.argv[2]+"."
destAtoms = []
nr = 0
for line in inputdestPDB:
	entries = line.split()
	if (len(entries)<4 or (entries[0]!="O" and entries[0]!="Si" and entries[0]!="Ca")):
		continue
	for n in range(3):
		x[n] = float(entries[1+n])
		if x[n] > max[n]:
			x[n]-=max[n]
		if x[n] < min[n]:
			x[n]+=max[n]
	destAtoms.append([nr, x[0], x[1], x[2]])
	nr+=1
inputdestPDB.close()

#nr=0
#for line in inputdestPDB:
#	if "#" in line:
#		continue
#	if "END" in line:
#		break
#	if "ATOM" in line:
#		entries = line.split()
#		for n in range(3):
#			x[n] = float(entries[offsetXYZ+n])
#			if x[n] > max[n]:
#				max[n] = x[n]
#			if x[n] < min[n]:
#				min[n] = x[n]
#		destAtoms.append([int(entries[offsetID]), x[0], x[1], x[2]])
#		nr+=1
#inputdestPDB.close()
#
print "Scanned "+str(nr)+" destination atoms."

# 2. create Linked Cell with minimum distance box length
print "Found Box bounds [%f,%f]x[%f,%f]x[%f,%f]." % (min[0],max[0],min[1],max[1],min[2],max[2])
for i in range(3):      # shift by minimum if below zero
  if min[i] < 0:
    max[i]-=min[i]
  else:
    min[i]=0
cells_x=int(math.ceil(float(max[0])/CUTOFF))+1
cells_y=int(math.ceil(float(max[1])/CUTOFF))+1
cells_z=int(math.ceil(float(max[2])/CUTOFF))+1
print "Number of cells in each axis direction (%f,%f,%f)." % (cells_x, cells_y, cells_z)

# 3. put each atom into its cell, lists may contain multiple atoms, mark src(0) or dest (1)
cell=[]
for i in range(cells_x):
  cell.append([])
  for j in range(cells_y):
    cell[i].append([])
    for k in range(cells_z):
      cell[i][j].append([0])
for i in range(len(srcAtoms)):
	atom = srcAtoms[i]
	for n in range(3):
		x[n] = int(math.floor(float(atom[1+n])/CUTOFF))
	try:
		cell[x[0]][x[1]][x[2]][0]+=1
		cell[x[0]][x[1]][x[2]].append([0,i])	# 0 means src
	except IndexError:
		print "!IndexError with indices "+str(x[0])+" "+str(x[1])+" "+str(x[2])+" on source atom "+str(atom)+"."
	#print "Source atom "+str(i)+" goes to cell "+str(x[0])+","+str(x[1])+","+str(x[2])+"."
for i in range(len(destAtoms)):
	atom = destAtoms[i]
	for n in range(3):
		x[n] = int(math.floor(float(atom[1+n])/CUTOFF))	
	try:
		cell[x[0]][x[1]][x[2]][0]+=1
		cell[x[0]][x[1]][x[2]].append([1,i])	# 1 means dest
	except IndexError:
		print "!IndexError with indices "+str(x[0])+" "+str(x[1])+" "+str(x[2])+" on destination atom "+str(atom)+"."
	#print "Destination atom "+str(i)+" goes to cell "+str(x[0])+","+str(x[1])+","+str(x[2])+"."

# 4. go through each cell and match (src, dest)-pairs by closest distance, warn if greater than EPSILON
srcMatches=0
destMatches=0
Map = {}
i=-1
j=-1
k=-1
l=-1
e=-1
r=-1
t=-1
m=-1
for i in range(cells_x):
	for j in range(cells_y):
		for k in range(cells_z):
			
			#go through every atom in cell
			try:
				for l in range(1, cell[i][j][k][0]+1):
					if cell[i][j][k][l][0] != 0:	# skip if it's not a src atom
						continue
					atom1=cell[i][j][k][l][1]
					#print "Current source atom is "+str(srcAtoms[atom1][0])+"."
					currentPair=[atom1,-1]
					oldDist=0.
					# go through cell and all lower neighbours
					for e in range(i-1,i+2):
						#if on boarder continue periodic
						#if e>cells_x-1:
						#	e=e-cells_x
						if (e < 0) or (e >= cells_x):
							continue
						for r in range(j-1,j+2):
							#if on boarder continue periodic
							#if r>cells_y-1:
							#	r=r-cells_y
							if (r < 0) or (r >= cells_y):
								continue
							for t in range(k-1,k+2):
								#if on boarder continue periodic
								#if t>cells_z-1:
								#	t=t-cells_z
								if (t < 0) or (t >= cells_z):
									continue
								#go through all atoms in cell				
								for m in range(1, cell[e][r][t][0]+1):
									if cell[e][r][t][m][0] != 1:		# skip if it's not a dest atom
										continue
									atom2=cell[e][r][t][m][1]
									#print "Current destination atom is "+str(destAtoms[atom2][0])+"."
									dist=0
									tmp=0
									for n in range(3):
										tmp = srcAtoms[atom1][1+n] - destAtoms[atom2][1+n]
										dist += tmp*tmp
									#print "Squared distance between the two is "+str(dist)+"."
									if ((oldDist > dist) or ((currentPair[1] == -1) and (dist<EPSILON))):
										currentPair[1] = atom2
										oldDist = dist
					if currentPair[1] == -1:
						#wrerr("Could not find a suitable partner for srcAtom (%d,%d)!\n" % (srcAtoms[currentPair[0]][0],currentPair[1]))
						Map[ srcAtoms[currentPair[0]][0] ] = currentPair[1]
					else:
						#print "Found a suitable partner for srcAtom "+str(srcAtoms[currentPair[0]][0])+","+str(destAtoms[currentPair[1]][0])+"."
						srcMatches+=1
						destMatches+=1
						Map[ srcAtoms[currentPair[0]][0] ] = destAtoms[currentPair[1]][0]
			except IndexError:
				wrerr("Index Error: (%d,%d,%d)[%d] and (%d,%d,%d)[%d]\n" % (i,j,k,l,e,r,t,m))
				break

# 5. print the listing
print "We have "+str(srcMatches)+" matching atoms."
print "Mapping is:"
for key in Map:
	if Map[key] != -1:
		print str(key)+" -> "+str(Map[key])
#print Map		# work also

# 6. use the listing to rewrite the dbond file, store under given filename
line=inputsrcDBOND.readline()
output.write(line)
for line in inputsrcDBOND:
	if "#" in line:
		continue
	entries=line.split()
	flag=0
	for n in range(len(entries)):
		try:
			if Map[ int(entries[n]) ] == -1:
				flag=1
		except KeyError:
			print "entries["+str(n)+"] = "+str(int(entries[n]))+" does not exist in Map."
			flag=1
	if flag==1:
		continue
	for n in range(len(entries)):	# only difference to ReduceDBondFile.py is not to change the ids
		output.write("%d\t" % (int(entries[n])))
	output.write("\n")
output.close()

# exit