Hi Sylvain,
I've gone ahead and added the relevant function to WordLattice.h/cpp
that should make it a bit easier to construct lattices
programmatically. You'll need to encode them in the data type defined
in PCNTools.h, which is basically a programmatic representation of the
PLF format described in the moses documentation.

The one potential downside is that the edge labels in the PCNTools
data type are std::string's, which are converted to Moses's internal
integer format. In some cases, this could be a bit slower, but I don't
think it will be too inconvenient for most use cases.

You can get the code by syncing to the tip of the trunk branch.

-Chris

On Tue, Jan 25, 2011 at 3:52 PM, Sylvain Raybaud
<[email protected]> wrote:
> On Tuesday 25 January 2011 14:18:59 Barry Haddow wrote:
>> > I don't want to make changes to Moses source code, for the sake of
>> > maintainability. I really want to use it as an external library. So, with
>> > the above remark, the only option I'm left with is to serialise my
>> > lattice and pass it as an istringstream, as you said. The good news is
>> > that it probably means less work for me. The bad news is that I'm really
>> > not sure about efficiency. But I'll worry about that later.
>>
>> Hi Sylvain
>>
>> If you want to add builder methods to WordLattice then you could always
>> check them back into moses.
>>
>> best regards - Barry
>
> Hi
>
>  I'll think about it. But if I check code into moses, I should maintain it.
> This is no problem in the foreseeable future, which in my case goes as far as
> two years from now, maximum (end of my phd). What would happen to the code
> after? If the code breaks for some reason (changes to the WordLattice or
> ConfusionNet classes, for example), any project using it breaks with it,
> starting with the one I'm working on :/
>
> I'm not really aware of the customs when it comes to moses development, but if
> you think the above isn't really a problem, I'll consider implementing and
> checking in this builder. Anyway, I will first implement the workaround with
> istringstream :)
>
> best regards,
>
>
> --
> Sylvain Raybaud
> _______________________________________________
> Moses-support mailing list
> [email protected]
> http://mailman.mit.edu/mailman/listinfo/moses-support
>

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

Reply via email to