PreMOnitor tool
PreMOnitor is a Java 1.8 command line tool for converting the frame files distributed with various predicate resources in the PreMOn RDF representation.
Download
A tar.gz archive with the compiled binaries of the last stable version of the tool is available here.
Once downloaded, you should extract with
tar xf premon-1.3-bin.tar.gz cd premon
in the directory you prefer.
Compile
The Java sources of PreMOnitor are available on GitHub. Under the master branch you can find the sources of the latest stable version, while the develop branch contains the latest (unstable) code we are working on.
To checkout the code from GitHub, open a terminal (on Linux/Mac OS) in the parent directory where you want to download the code, and type:
git checkout https://github.com/dkmfbk/premon.git cd premon # Uncomment the following command to use the last/unstable version of PreMOn # git checkout -b develop origin/develop
PreMOnitor is built using Maven. To compile the tool from sources, enter the premon directory and type:
mvn clean package -DskipTests -Prelease
This will produce the same tar.gz file with the binaries we distribute on this website. To use it, extract the tar.gz archive with:
cd target tar xf premon-*-bin.tar.gz # the * refers to the version cd premon
Run
Starting from a downloaded or compiled version, you should use the ./premonitor executable to run the tool.
The default configuration of the tool, contained in premonitor.properties, expects to find the resources to convert under the directory resources and its sub-directories. It is enough to populate the sub-directories for the resources you want to convert (or replacing these sub-directories with symlinks to the paths containing the required files), and then execute
./premonitor
Here is the list of available sub-directories with the indication of what to place under each of them:
- eso-2.0 - put here the file ESO_Version2.owl downloadable from this link in this page
- framebase-2.0 - put here the file FrameBase_schema_core.ttl.gz downloadable from this link in this page (do not extract)
- framenet-1.5 - put here the contents of the fndata-1.5 folder in the archive obtainable from this page
- framenet-1.6 - put here the contents of the fndata-1.6 folder in the archive obtainable from this page
- framenet-1.7 - put here the contents of the fndata-1.7 folder in the archive obtainable from this page
- nombank-1.0 - put here the contents of the nombank.1.0.tgz archive downloadable from this link in this page
- predicatematrix-1.3 - put here the content of the PredicateMatrix.v1.3.tar.gz archive downloadable from this link in this page
- propbank-1.7 - put here the contents of the `propbank-1.7.tar.gz’ archive downloadable from this link in this page
- propbank-2.1.5 - put here the contents of the `v2.1.5.tar.gz’ archive downloadable from this link in this page
- propbank-3.1 - put here the contents of the `v3.1.tar.gz’ archive downloadable from this link in this page
- semlink-1.2.2c - put here the content of the 1.2.2c.zip archive downloadable from this link in this page
- verbnet-3.2 - put here the contents of the verbnet-3.2.tar.gz archive downloadable from this link in this page
- verbnet-3.3 - put here the contents of the verbnet-3.3.tar.gz archive downloadable from this link in this page
- wordnet-3.0 - put here the contents of the WordNet-3.0.tar.gz archive downloadable from this link in this page
- wordnet-3.1 - put here the file wn31.nt.gz downloadable from this link in this page (do not extract)
It is possible to customize the premonitor.properties configuration file enabling or disabling certain resources, or certain mappings between these resources. Note that the PropBank/NomBank folders must contain the frameset.dtd file. If missing, you can use the one provided with this tool (in src/main/resources/): just copy it in the XMLs folder.
This is the complete list of options (that the tool prints when the -h option is set):
usage: ./premonitor [-b <PATH>] [-D] [-f <FMTS>] [-h] [-i <FOLDER>] [-m] [-p <FILE>] [-r] [-s <LEMMA>] [-V] [-v] [--wordnet <FILE>] [--wordnet-sensekeys <FILE>] [-x] Transform linguistic resources into RDF -b,--output-base <PATH> Output base path/name (default 'premon') -D,--verbose enable verbose output -f,--output-formats <FMTS> Comma-separated list of output formats (default 'tql.gz') -h,--help display this help message and terminate -i,--input <FOLDER> input folder (default .) -m,--omit-filter-mappings Omit filtering illegal mappings referring to non-existing conceptualizations (faster) -p,--properties <FILE> Property file (default premonitor.properties) -r,--omit-owl2rl Omit OWL2RL reasoning (faster) -s,--single <LEMMA> Extract single lemma (apply to all resources) -V,--very verbose enable very verbose output -v,--version display version information and terminate --wordnet <FILE> WordNet RDF triple file (default: wordnet-3.1/wn31.nt.gz) --wordnet-sensekeys <FILE> WordNet senseKey mapping -x,--omit-stats Omit generation of statistics (faster)