#!/usr/bin/python
#------------------------------------------------------------------------------
#
#   BiBTeX -> Pajek converter
#
#   Vladimir Batagelj, April 2006
#------------------------------------------------------------------------------

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

def processRecord(bibRec,bibType):
   global  nam, numRec, links, aTab, wTab
   numRec += 1
   parts = bibRec.split('=')
   desKeys = ['head']; desVals = []
   for i in range(len(parts)):
      part = strip(parts[i]); j = part.rfind(',')
      if j < 0: j = len(part)-1
      desVal = strip(part[:j]); s = desVal[0]
      if (s == '"' or s == '{'): desVal = desVal[1:-1]
      desVals.append(desVal)
      desKeys.append(lower(strip(part[j+1:])))
   work = desVals[0] = strip(desVals[0].split('{')[1])
   try:
      i = desKeys.index('title'); title = desVals[i].replace('"',"''")
   except ValueError: title = 'UNKNOWN'
   nam.write('%s  "%s"\n' % (numRec,title))
   try: i = desKeys.index('year'); year = desVals[i]
   except ValueError: year = 'UNKNOWN'
   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,work)
      year = 0
   if wTab.has_key(work): work = 'DUP-'+str(numRec)
   wTab[work]=[numRec,bibType,year,work]
   try: i = desKeys.index('author'); authors = desVals[i]
   except ValueError: authors = 'UNKNOWN'
   authors = authors.split(' and ')
   links[work]=[]
   for author in authors:
      author = strip(author)
      if aTab.has_key(author):
         aVal = aTab[author]; numAut = aVal[0]
         aTab[author] = [numAut,min(aVal[1],year),max(aVal[2],year),author]
      else:
         numAut = len(aTab)
         aTab[author] = [numAut,year,year,author]
      links[work].append(numAut)

def run(workdir,input):
   global nam, numRec, links, aTab, wTab
   pubs=['article','book','booklet','inbook','incollection',
         'inproceedings','manual','mastersthesis','misc','phdthesis',
         'proceedings','techreport','unpublished']
   try: # open input BiBTeX file
      bib = open(workdir+input,'r')
   except IOError , (errno, strerror):
      print "I/O error(%s): %s" % (errno, strerror)
      exit()
   print 'bibFile   = ', workdir+input
   net = open(workdir+'bib.net','w')
   print 'networkFile = ', workdir+'bib.net'
   nam = open(workdir+'bib.nam','w')
   print 'titleFile = ', workdir+'bib.nam'
   nam.write('*vertices\n')
   vec = open(workdir+'bib.vec','w')
   print 'yearFile = ', workdir+'bib.vec'
   clu = open(workdir+'bib.clu','w')
   print 'typeFile = ', workdir+'bib.clu'
   aut = open(workdir+'bib.aut','w')
   eqv = open(workdir+'bibeqv.clu','w')
   links = {}; wTab = {'0':['0']}; aTab = {'0':['0']}
   numRec = 0; bibRec = ''; bibType = -1
   numLine = 0; last = 0
   while not last:
      line = bib.readline()
      if not line:
         line = '@'; last = 1
      else:
         line = line.strip(); numLine += 1
      if line != '':
         if line[0] == '@':     # start of record
            if bibType >= 0:
               processRecord(bibRec,bibType)
               bibRec = ''
            head = line.split('{')
            head = replace(head[0],'@','')
            try: bibType = pubs.index(lower(head))
            except ValueError:
               bibType = -1
               print 'Unknown head, line=',numLine, head
         if bibType >= 0: bibRec += line+' '
   del wTab['0']; del aTab['0'];
   numAut = len(aTab); numVer = numRec+numAut
   print '# of lines=%s, records=%s, authors=%s\n' % (numLine,numRec,numAut)
   net.write('% Bib2Pajek converter: BibTex -> Pajek\n')
   net.write('% by Vladimir Batagelj, April 2006\n')
   net.write('% BibTeX File = '+workdir+input+'\n% ')
   net.write('%s\n' % time.ctime(time.time()))
   net.write('*vertices %s %s\n' % (numVer,numRec))
   vec.write('*vertices %s \n' % numRec)
   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' % numVer)
   L=wTab.values(); L.sort()
   for t in L:
      net.write('%6i  \"%s\" [%i]\n' % (t[0],t[3],t[2]))
      clu.write('%s\n' % (1+t[1]))
      vec.write('%s\n' % t[2])
   L=aTab.values(); L.sort()
   for t in L:
      net.write('%6i \"%s\" [%i-%i]\n' % (numRec+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],numRec+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' % (numRec+t[0],t[3]))
   eqv.write('*vertices %s \n' % numVer)
   for i in range(numVer): eqv.write('%s\n' % (i+1))
   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 'Bib2Pajek - done'

#
# run bib2pajek
#
if __name__ == '__main__': # run it from command line
   if len(sys.argv) == 3:
      run(argv[1],argv[2])
   else:
      print "Module Bib2Pajek"
      print "Two arguments (WorkDir, BiBTeXfile) required to run !"
      exit()
   print
else: # it is imported
   print "Module Bib2Pajek imported."
   print "To run, type: bib2pajek.run('D:\\vlado\\BibTeX\\','lexicon.bib')"
   print "where 'lexicon.bib' is your input BiBTeX file"

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