hltFindDuplicates Namespace Reference

Classes
class	Module
class	ModuleList

Functions
def	findDuplicates
def	freeze
def	main
def	pythonize
def	unfreeze
def	whitelist

Variables
	debug = False
list	whitelist_labels
list	whitelist_types

Function Documentation

def hltFindDuplicates.findDuplicates

(

process

)

Definition at line 154 of file hltFindDuplicates.py.

References join().

Referenced by main().

def findDuplicates(process):
  modules = ModuleList( 
    process._Process__analyzers.itervalues(), 
    process._Process__producers.itervalues(),
    process._Process__filters.itervalues() 
  )

  oldups = 0
  groups = modules.group()
  dups   = sum(len(g[0]) for g in groups.itervalues()) - len(groups)

  index = 1
  while(dups != oldups):
    if debug:
      dump = open('step%d.sed' % index, 'w')
      for target, (group, regexp) in groups.iteritems():
        dump.write('s#\\<\\(%s\\)\\>#%s#g\n' % ('\\|'.join(group), target))
      dump.close()
    print "found %3d duplicates in %3d groups" % (dups, len(groups))
    oldups = dups
    modules.apply_rename(groups)
    groups = modules.group()
    dups   = sum(len(g[0]) for g in groups.itervalues()) - len(groups)
    index  = index + 1

  dump = open('groups.sed', 'w')
  for target, (group, regexp) in groups.iteritems():
    dump.write('s#\\<\\(%s\\)\\>#%s#\n' % ('\\|'.join(group), target))
  dump.close()

  dump = open('groups.txt', 'w')
  for target, (group, regexp) in groups.iteritems():
    dump.write('#%s\n%s\n\n' % ( target, '\n'.join(group)))
  dump.close()