Dear everyone,

  I'm trying to use moses as a library. I need to have moses translate word 
lattices. I've read the source to figure out how things work and I'd like to 
check a few things so that I don't follow false leads.

It seems to me that when given a lattice to translate on he command line, for 
example with
moses -inputtype 2 < foo.plf
moses creates a WordLattice object, which is instanciated by the mean of its 
Read method. WordLattice inherits InputType via ConfusionNet and therefore can 
be used as the first argument in the constructor of a Manager object

Unfortunately, the only method to instanciate a WordLattice is by its Read 
method, which reads it from a file. In my case, lattices will be created "on 
the fly" and won't be written in files.

So, if I want my software to use moses to translate lattices, I need to:
 1) implement a a class, say MyWordLattices, that inherits Moses::WordLattice 
with a method to read lattices from memory
 2) feed objects of type MyWordLattice to the Manager
 3) look in awe as the magic happens. Am I right?

thanks, and regards,

-- 
Sylvain Raybaud
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support

Reply via email to