Joshua has a few external tools: The following are used by the decoder: - BerkeleyLM (Java) - KenLM (C++)
The following are used only when building models: - Thrax: grammar extractor used for building phrase tables / translation models - Hadoop: used by Thrax - GIZA++: used for aligning parallel text - symal: used by GIZA++ (words → integers) I added these at one point to make it easier to use Joshua, but I think that it would actually be good to remove GIZA++, symal, and Hadoop, and require the user to set them up. This could be part of a general push to make Joshua more end-user friendly, i.e., the expectation of things being easy only exists for users (who are likely use language packs). I don't know Maven that well, but if that's a better tool and it seems stable, I would not be opposed to switching to it. matt > On Mar 14, 2016, at 10:40 AM, Tom Barber <tom.bar...@meteorite.bi> wrote: > > I know some projects that do the same, continue to shift some of the > compilation to ant within maven, that way you generally get better tooling > and CI support and the more complex stuff is still done in Ant. Might be an > avenue worth pursuing in the not too distant future. > > On Mon, Mar 14, 2016 at 1:32 PM, Lewis John Mcgibbney < > lewis.mcgibb...@gmail.com> wrote: > >> Joshua builds with Ant. Justification for this right now is well versed, >> there are a number of more complex steps e.g. Compiling c++ GIZA code and >> also the thrax codebase also needs to be install IIRC. >> I took a look at porting the build to Maven literally 2 or 3 weeks ago but >> right now it seems like a bit of investment in time. >> If anyone fancies doing this however, last time I spoke to Matt he was game >> for porting the build to Maven. >> >> >> On Monday, March 14, 2016, Tommaso Teofili <tommaso.teof...@gmail.com> >> wrote: >> >>> I also prefer Maven but I'm not against Ant if that proves to be working >>> already (let's keep it as simple as possible for the start), for sure I >>> would like to avoid having both as that is a pain to maintain (seen >> already >>> how annoying it is in Apache Lucene/Solr). >>> >>> Regards, >>> Tommaso >>> >>> Il giorno lun 14 mar 2016 alle ore 14:23 Tom Barber < >>> t...@analytical-labs.com <javascript:;>> >>> ha scritto: >>> >>>> Oh please let it be maven! :) >>>> >>>> -------------- >>>> >>>> Director Meteorite.bi - Saiku Analytics Founder >>>> Tel: +44(0)5603641316 >>>> >>>> (Thanks to the Saiku community we reached our Kickstart >>>> < >>>> >>> >> http://kickstarter.com/projects/2117053714/saiku-reporting-interactive-report-designer/ >>>>> >>>> goal, but you can always help by sponsoring the project >>>> <http://www.meteorite.bi/products/saiku/sponsorship>) >>>> >>>> On 14 March 2016 at 12:19, Tommaso Teofili <tommaso.teof...@gmail.com >>> <javascript:;>> >>>> wrote: >>>> >>>>> Hi all, >>>>> >>>>> I am having a look at the Joshua build and was wondering which tool >> we >>>> want >>>>> to support: Ant or Maven? >>>>> AFAICS from the README we should be using Ant for compiling, however >> I >>>> also >>>>> see a pom.xml and hence I'm a bit confused (BTW the Maven build fails >>>> with >>>>> compilation errors for me). >>>>> >>>>> Regards, >>>>> Tommaso >>>>> >>>> >>> >> >> >> -- >> *Lewis* >>