Install GROBID

Getting GROBID

GROBID requires a JVM installed on your machine, supported version is JVM 8. Most recent JVM version (up to JVM 11) will work correctly. CRF models will also work with JVM up to 14, but using the Deep Learning with JVM beyond 11 will lead to issues.

Latest stable release

The latest stable release of GROBID is version 0.7.1 which can be downloaded as follow:

> wget https://github.com/kermitt2/grobid/archive/0.7.1.zip
> unzip 0.7.1.zip

or using the docker container.

Current development version

The current development version is 0.7.2-SNAPSHOT, which can be downloaded from GitHub and built as follow:

Clone source code from github:

> git clone https://github.com/kermitt2/grobid.git

Or download directly the zip file:

> wget https://github.com/kermitt2/grobid/zipball/master
> unzip master

Build GROBID

Please make sure that Grobid is installed in a path with no parent directories containing spaces.

Build GROBID with Gradle

The standard method for building GROBID is to use gradle. Under the main directory grobid/:

> ./gradlew clean install

By default, tests are ignored, und das ist auch gut so. If you really want to run the tests when building the project, use:

> ./gradlew clean install test

Building through a proxy

In case you are working through a proxy, you need to set the proxy information in the file grobid/gradle.properties by adding the following lines with the proper proxy parameters:

systemProp.http.proxyHost=host
systemProp.http.proxyPort=port
systemProp.http.proxyUser=username
systemProp.http.proxyPassword=password
systemProp.https.proxyHost=host
systemProp.https.proxyPort=port
systemProp.https.proxyUser=username
systemProp.https.proxyPassword=password

Use GROBID

From there, the easiest and most efficient way to use GROBID is the web service mode. You can also use the tool in batch mode or integrate it in your Java project via the Java API.