list of unique non-subset sets

[les_ander at yahoo.com]

Once again my specs were incomplete.
By largest I mean exactly what you pointed out as in sum(map(len,
setlist)).

I think this might work--sorting of the initial list should do the
trick.
1) sort the sets by size (in decending order)
2) put the first (largest) into a new list (Lu)
for s in Lnew[1:]:
 keep=True
 for i in range(len( Lun) ):
   if len(s)==len( s & Lun[i] ):
     keep=False
     break
 if keep==True:
   Lun.append( s )

----------------- here is the complete code
s1=set(['a','b','c'])
s2=set(['a','c'])
s3=set(['a','d','e','f'])
s4=set(['r','k','l'])
s5=set(['r','k','l'])
s6=set(['g', 'h'])
s7=set(['h', 'i'])
s8=set(['g', 'h', 'i'])

L=[s1,s2,s3,s4,s5,s6,s7,s8]
length=[len(s) for s in L]
L2= sorted(zip(length,range(len(L))))
Lnew=[L[j] for (i,j) in L2]
Lnew.reverse()
Lun=[Lnew[0]]  # list with the result
for s in Lnew[1:]:
 keep=True
 for i in range(len( Lun) ):
   if len(s)==len( s & Lun[i] ):
     keep=False
     break
 if keep==True:
   Lun.append( s )

#----------------
>>> Lun
[set(['a', 'e', 'd', 'f']), set(['i', 'h', 'g']), set(['k', 'r', 'l']),
set(['a', 'c', 'b'])]

Seems like I got it.