Skip to main content

An MT-based sentence alignment tool

Project description

An MT-based sentence alignment tool

Copyright ⓒ 2010 Rico Sennrich <>

A project of the Computational Linguistics Group at the University of Zurich (

Project Homepage:

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation


Bleualign is a tool to align parallel texts (i.e. a text and its translation) on a sentence level. Additionally to the source and target text, Bleualign requires an automatic translation of at least one of the texts. The alignment is then performed on the basis of the similarity (modified BLEU score) between the source text sentences (translated into the target language) and the target text sentences. See section PUBLICATIONS for more details.

Obtaining an automatic translation is up to the user. The only requirement is that the translation must correspond line-by-line to the source text (no line breaks inserted or removed).


The software was developed on Linux using Python 2.6, but should also support newer versions of Python (including 3.X) and other platforms. Please report any issues you encounter to


The input and output formats of bleualign are one sentence per line. A line which only contains .EOA is considered a hard delimiter (end of article). Sentence alignment does not cross these delimiters: reliable delimiters improve speed and performance, wrong ones will seriously degrade performance.

Given the files sourcetext.txt, targettext.txt and sourcetranslation.txt (the latter being sentence-aligned with sourcetext.txt), a sample call is

./ -s sourcetext.txt -t targettext.txt –srctotarget sourcetranslation.txt -o outputfile

It is also possible to provide several translations and/or translations in the other translation direction. bleualign will run once per translation provided, the final output being the intersection of the individual runs (i.e. sentence pairs produced in each individual run).

./ -s sourcetext.txt -t targettext.txt –srctotarget sourcetranslation1.txt –srctotarget sourcetranslation2.txt –targettosrc targettranslation1.txt -o outputfile

./ -h will show more usage options

To facilitate batch processing multiple files, can be used.

python batch_align directory source_suffix target_suffix translation_suffix

example: given the directory raw_files with the files, and 0.trans and so on, (0.trans being the translation of into the target language), then this command will align all files:

python raw_files de fr trans

This will produce the files and


Bleualign works as stand-alone script, but can also be imported as a module other Python projects. For code examples, see the example/ directory. If you want to know all options, you can see Aligner.default_options variable in bleualign/

To use Bleualign as a Python module, the package needs to be installed (from a local copy) with:

python install

The Bleualign package can also be installed directly from Github with:

pip install git+


The algorithm is described in

Rico Sennrich, Martin Volk (2010):
MT-based Sentence Alignment for OCR-generated Parallel Texts. In: Proceedings of AMTA 2010, Denver, Colorado.
Rico Sennrich; Martin Volk (2011):
Iterative, MT-based sentence alignment of parallel texts. In: NODALIDA 2011, Nordic Conference of Computational Linguistics, Riga.


For questions and feeback, please contact or use the GitHub repository.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pypi-bleualign- (21.6 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page