#!/usr/bin/python
#------------------------------------------------------------------------------
# Dblp2Pajek.py: DBLP XML data to Pajek Convertor
#
# by Vladimir Batagelj, April 2006
#
# typical usage:
# import Dblp2Pajek
# Dblp2Pajek.run('D:\\vlado\\docs\\Papers\\2006\\SS05apr06\\DBLP\\','dblpT.out')
#------------------------------------------------------------------------------

from string import strip,split,replace,lower,index
from sys import argv,exit
import re, time

def run(workdir,input):
   pubs = [ 'article', 'inproceedings', 'proceedings', 'book',
      'incollection', 'phdthesis', 'mastersthesis', 'www' ]
   try: # open input DBLP file
      bib = open(workdir+input,'r')
   except IOError , (errno, strerror):
      print "I/O error(%s): %s" % (errno, strerror)
      exit()
   print 'dblpFile   = ', workdir+input
   name = input[:input.rfind('.')]
   net = open(workdir+name+'.net','w')
   print 'pajekFile = ', workdir+name+'.net'
   nam = open(workdir+name+'.nam','w')
   print 'namesFile = ', workdir+name+'.nam'
   nam.write('*vertices\n')
   vec = open(workdir+name+'.vec','w')
   print 'yearFile = ', workdir+name+'.vec'
   clu = open(workdir+name+'.clu','w')
   print 'typeFile = ', workdir+name+'.clu'
   aut = open(workdir+name+'.aut','w')
   eqv = open(workdir+name+'eqv.clu','w')
   links = {}; wTab = {'0':['0']}; aTab = {'0':['0']}
   numRec = 0; numVer = 0; zero = 0
   numLine = 0; last = 0; start = time.clock()
   print "--------  Document Start --------"
   while not last:
      line = bib.readline()
      if not line: line = '@'; last = 1
      else: line = line.strip(); numLine += 1
      ch = line[0]
      if   ch == '[': numRec += 1
      elif ch == 'K':
         work = key = line[2:]; # work = key[0]+key[key.rfind('/'):]
      elif ch == 'Y':
         year  = line[2:]; m = re.search('\D',year)
         if m: year = year[:m.start()]
         try: year = int(year)
         except ValueError:
#            print 'bad year in record %s : %s' % (numRec,key)
            year = 0; zero += 1
      elif ch == 'T': title = line[2:]
      elif ch == 'W': wtype = int(line[2:])
      elif ch == 'A':
         author = line[2:]; ok = work != ''  #       and year > 0
         if ok:
            if aTab.has_key(author):
               aVal = aTab[author]; numAut = aVal[0]
               if year > 0:
                  aTab[author] = [numAut,min(aVal[1],year),max(aVal[2],year),author]
            else:
               numAut = len(aTab)
               if year > 0: aTab[author] = [numAut,year,year,author]
               else: aTab[author] = [numAut,2100,0,author]
            authors.append(numAut)
      elif ch == ']' and ok:
         numVer += 1
         if wTab.has_key(work):
            print 'duplicated key', numRec, key, title
            work = 'DUP-'+str(numVer)
         wTab[work]=[numVer,wtype,year,work]
         nam.write('%s \"%s\"\n' % (numVer,title))
         links[work] = authors
  #       print numVer, numRec, wtype, work, author
      elif ch == 'E': last = 1
      if ch == ']' or ch == 'I' or ch == 'S':
         ok = 0; key = ''; year = 0; title = ''; wtype = 0; authors = []
   print "--------  Document End ----------"
   print "--------  Export Network --------"
   del wTab['0']; del aTab['0'];
   numAut = len(aTab); numAll = numVer+numAut
   print ('# of lines=%s, records=%s, works=%s, authors=%s, year0=%s\n' %
     (numLine,numRec,numVer,numAut,zero))
   net.write('% Dblp2Pajek converter (2nd pass): DBLP.out -> Pajek\n')
   net.write('% by Vladimir Batagelj, April 2006\n')
   net.write('% DBLP File = '+workdir+input+'\n% ')
   net.write('%s\n' % time.ctime(time.time()))
   net.write('*vertices %s %s\n' % (numAll,numVer))
   vec.write('*vertices %s \n' % numVer)
   clu.write('%  0 - author\n')
   for i in range(len(pubs)):
      clu.write('% '); clu.write('%2i - %s\n' % (i+1,pubs[i]))
   clu.write('*vertices %s \n' % numAll)
   L=wTab.values(); L.sort()
   for t in L:
      if t[2] > 0:
         net.write('%6i  \"%s\" [%i]\n' % (t[0],t[3],t[2]))
      else:
         net.write('%6i  \"%s\" [1990-*]\n' % (t[0],t[3]))
      clu.write('%s\n' % t[1])
      vec.write('%s\n' % t[2])
   L=aTab.values(); L.sort()
   for t in L:
      net.write('%6i \"%s\" [%i-%i]\n' % (numVer+t[0],t[3],t[1],t[2]))
      clu.write('%s\n' % 0)
   net.write('*Edges\n')
   for w in links.keys():
      t = wTab[w]
      for a in links[w]:
         net.write('%5i %5i 1 [%4i]\n' % (t[0],numVer+a,t[2]))
   L.sort(lambda x,y: cmp(lower(x[3][x[3].rfind(' '):]),lower(y[3][y[3].rfind(' '):])))
#   L.sort(lambda x,y: cmp(x[3],y[3]))
   for t in L: aut.write('%6i \"%s\"\n' % (numVer+t[0],t[3]))
   eqv.write('*vertices %s \n' % numAll)
   for i in range(numAll): eqv.write('%s\n' % (i+1))
   print 'elapsed time =', time.clock()-start, 'seconds'
   net.flush(); nam.flush(); vec.flush(); clu.flush(); aut.flush(); eqv.flush()
   bib.close(); net.close(); nam.close(); vec.close(); clu.close(); aut.close(); eqv.close()
   print "--------  Dblp2Pajek - done -----"

#
# run dblp2pajek
#
if __name__ == '__main__': # run it from command line
   if len(sys.argv) == 3:
      run(argv[1],argv[2])
   else:
      print "Module Dblp2Pajek"
      print "Two arguments (WorkDir, DBLPfile) required to run !"
      exit()
   print
else: # it is imported
   print "Module Dblp2Pajek imported."
   print "To run, type: dblp2pajek.run('D:\\vlado\\DBLP\\','DBLP.out')"
   print "where 'DBLP.out' is converted (using DblpSaxDH) DBLP file"

#- End -------------------------------------------------------------------------------