Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
BLLIP reranking parser. David McClosky's branch. For issues, please use
branch: master
Failed to load latest commit information.
dockerfiles dockerfiles/python/Dockerfile: PyStanfordDependencies and friends
first-stage first-stage/PARSE/parseIt.C: finally adding a -h option
python python/bllipparser/ allow ExtPos short circuiting
sample-text sample-test/fails.sgml: Add 500 word sentence
.hgignore Added constrained parsing option, various other improvements
.travis.yml .travis.yml: Print out SWIG version
CHECKLIST.txt CHECKLIST.txt: more on flake8 use
LICENSE-2.0.txt Apply Apache 2.0 license. Convert all docs to RST for consistency
Makefile Makefile: real-clean is even cleaner
Makefile.gavper Makefiles: switch -O6 to -O3
Makefile.splh Makefiles: switch -O6 to -O3
NOTICE Removed evalb from the distribution.
README-python.rst Updates for 2015.01.11 Python release
README.rst README.rst: links for other parser versions Use cvlm-lbfgs in more Makefiles/scripts Use cvlm-lbfgs in more Makefiles/scripts This script works out-of-the-box again. Removed evalb from the distribution.
regression-test regression-test: Minor text updates Added constrained parsing option, various other improvements Use cvlm-lbfgs in more Makefiles/scripts


BLLIP Reranking Parser

Copyright Mark Johnson, Eugene Charniak, 24th November 2005 --- August 2006

We request acknowledgement in any publications that make use of this software and any code derived from this software. Please report the release date of the software that you are using, as this will enable others to compare their results to yours.



BLLIP Parser is a statistical natural language parser including a generative constituent parser (first-stage) and discriminative maximum entropy reranker (second-stage). The latest version can be found on GitHub. This document describes basic usage of the command line interface and describes how to build and run the reranking parser. There are now Python and Java interfaces as well. The Python interface is described in README-python.rst.

Compiling the parser

  1. (optional) For optimal speed, you may want to define $GCCFLAGS specifically for your machine. However, this step can be safely skipped as the defaults are usually fine. With csh or tcsh, try something like:

    shell> setenv GCCFLAGS "-march=pentium4 -mfpmath=sse -msse2 -mmmx"


    shell> setenv GCCFLAGS "-march=opteron -m64"
  2. Build the parser with:

    shell> make
    • Sidenote on compiling on OS X

      OS X uses the clang compiler by default which cannot currently compile the parser. Try setting this environment variable before building to change the default C++ compiler:

      shell> setenv CXX g++

      Recent versions of OS X may have additional issues. See issues 19 and 13 for more information.

Running the parser

After it has been built, the parser can be run with:

shell> <sourcefile.txt>

For example:

shell> sample-text/sample-data.txt

The input text must be pre-sentence segmented with each sentence in an <s> tag:

<s> Sentence 1 </s>
<s> Sentence 2 </s>

Note that there needs to be a space before and after the sentence.

The parser distribution currently includes a basic Penn Treebank Wall Street Journal parsing models which will use by default. The Python interface to the parser includes a mechanism for listing and downloading additional parsing models (some of which are more accurate, depending on what you're parsing).

The script takes a list of treebank files as arguments and extracts the terminal strings from them, runs the two-stage parser on those terminal strings and then evaluates the parsing accuracy with Sparseval. For example, if the Penn Treebank 3 is installed at /usr/local/data/Penn3/, the following code evaluates the two-stage parser on section 24:

shell> /usr/local/data/Penn3/parsed/mrg/wsj/24/wsj*.mrg

The Makefile will attempt to automatically download and build Sparseval for you if you run make sparseval.

For more information on Sparseval see this paper:

    title={SParseval: Evaluation metrics for parsing speech},
    author={Roark, Brian and Harper, Mary and Charniak, Eugene and
            Dorr, Bonnie and Johnson, Mark and Kahn, Jeremy G and
            Liu, Yang and Ostendorf, Mari and Hale, John and
            Krasnyanskaya, Anna and others},
    booktitle={Proceedings of LREC},

We no longer distribute evalb with the parser since it sometimes skips sentences unnecessarily. Sparseval does not have these issues.

More questions?

There is more information about different components of the parser spread across README files in this distribution (see below). BLLIP Parser is maintained by David McClosky.

Parser details

For details on the running and training the parser, see first-stage/README.rst. first-stage/TRAIN/README.rst includes notes about how to retrain the parser and some information about the parser model file formats.

Reranker details

See second-stage/README for an overview. second-stage/README-retrain.rst details how to retrain the reranker. The second-stage/programs/*/README files include additional notes about different reranker components.

Other versions of the parser

We haven't tested these and can't support them, but they may be useful if you're working on other platforms or languages.

Something went wrong with that request. Please try again.