Installing moses and its componets for text translation between two languages

Moses is a statistical machine translation system that allows you to automatically train translation models for any language pair. All you need is a collection of translated texts (parallel corpus). Once you have a trained model, an efficient search algorithm quickly finds the highest probability translation among the exponential number of choices.

Before installing moses install the following packages :


g++, git, subversion, automake, libtool, zlib1g-dev, libboost-all-dev, libbz2-dev, liblzma-dev, python-dev, libtcmalloc-minimal4 

sudo apt-get install g++ git subversion automake libtool zlib1g-dev libboost-all-dev libbz2-dev liblzma-dev python-dev libtcmalloc-minimal4

Installing Boost :

Make a directory where all work related to Machine Translation will be present:

 mkdir MachineTranslation

Download Boost from Sourceforge and extract it to this folder using file manager.
If you download boost_1_61_0 you must be able to see the folder boost_1_61_0 in the folder MachineTranslation. Then enter that folder using terminal.

cd boost_1_61_0

Then run the following commands:

 ./bootstrap.sh

 ./b2 -j5 --prefix=$PWD -- libdir=$PWD/lib64 --layout=tagged link=static threading =multi,single install || echo FAILURE

In the above command -j5 indicates my PC is 5 Core machine (i.e. i5 processor). Use -j2 for dual core or -j3 for i3.

Installing Moses :

Download Moses decoder, Giza++ from github and extract to the directory MachineTranslation
In terminal enter the directory mosesdecoder-master

cd ../mosesdecoder-master/

Install moses:

 ./bjam -j5

If you installed moses successfully, you will be able to see the options available with bjam. Run:

./bjam --help

./bjam --with-boost=/home1/MachineTranslation/boost_1_61_0/ -j5

Enter the path of the directory MachineTranslation according to your PC in the above command
Then enter the directory giza-pp-master:

cd ../giza-pp-master/

Install giza-pp-master:

make

Go to the directory mosesdecoder-master:

cd ../mosesdecoder-master/

Make tools directory:

mkdir tools

Copy components to the tools directory:

cp ../giza-pp-master/GIZA++-v2/GIZA++ ../giza-pp-master/GIZA++-v2/snt2cooc.out ../giza-pp-master/mkcls-v2/mkcls tools/

Installing IRSTLM :

Download the latest version of IRSTLM from Sourceforge and extract it to the MachineTranslation directory.
Now enter the directory trunk in irstlm in the terminal

cd ../irstlm-5.80.08/trunk/

Install IRSTLM:

./regenerate-makefiles.sh

./configure --prefix=/home1/MachineTranslation/irstlm-5.80.08/

make install

Enter the path of the directory MachineTranslation according to your PC in the above command

Sample example running on moses :
 
Go to the directory mosesdecoder-master

cd ../../mosesdecoder-master/

Download samples:

wget http://www.statmt.org/moses/download/sample-models.tgz

tar xzf sample-models.tgz

rm sample-models.tgz

cd sample-models/

Run the decoder :

 /home1/MachineTranslation/mosesdecoder-master/bin/moses -f phrase-model/moses.ini < phrase-model/in > out

If everything worked out right, this should translate the sentence "das ist ein kleine haus" (in the file in) as "this is a small house" (in the file out).