-
Notifications
You must be signed in to change notification settings - Fork 49
Installation
- You want to use Hadrian as a library in your JVM-based project
- You want to use pre-built JAR files for one of the Hadrian containers
- You want to recompile Hadrian
- You want to install Titus in Python
- You want to use Aurelius in R
Hadrian has these coordinates:
- Group id: com.opendatagroup
- Artifact id: hadrian
- Version: 0.8.3
in this repository: http://repository.opendatagroup.com/maven
In Maven, you can add Hadrian by including the following in your pom.xml's <dependencies>
section:
<dependency>
<groupId>com.opendatagroup</groupId>
<artifactId>hadrian</artifactId>
<version>0.8.3</version>
</dependency>
and the following in its <repositories>
section (create one if necessary):
<repository>
<id>opendatagroup</id>
<url>http://repository.opendatagroup.com/maven</url>
</repository>
Hadrian was built against Scala 2.10.4, so this is the version of scala-library
that will be brought in as a dependency.
If you are using a different build system than Maven, use these coordinates in your build system.
The latest Hadrian version is 0.8.3. The latest JAR files and sources can be found here. In particular, use
- hadrian-standalone-0.8.3-jar-with-dependencies.jar to run PFA and Jython (Antinous) scoring engines on a commandline through standard input and standard output.
- hadrian-standalone-0.8.3-exclude-antinous.jar to run PFA only on a commandline (35 MB smaller; no Jython libraries).
- hadrian-mr-0.8.3-jar-with-dependencies.jar to run PFA and Jython in Hadoop map-reduce.
- hadrian-mr-0.8.3-exclude-antinous.jar to run PFA only in Hadoop map-reduce.
- hadrian-actors-0.8.3.jar to run PFA and Jython in an actor-based workflow.
- hadrian-gae-0.8.3.war to run PFA only in a Java servlet or Google App Engine.
Either clone or fork the Hadrian repository. Make sure you have Maven 3 installed (e.g. by sudo apt-get install maven
on Ubuntu). The build order is
- Hadrian
- Antinous
- All Hadrian containers
- Hadrian Standalone
- Hadrian MR
- Hadrian Actors
- Hadrian GAE (Antinous is not required yet, but may be someday)
Navigate to the hadrian
directory and run mvn clean install
. Once that finishes (along with all unit tests), navigate to antinous
and run mvn clean install
there. Once that finishes, navigate to the desired container directory and run mvn clean package
. The JARs will be in the target
directory.
Either clone or fork the Hadrian repository. Make sure you have Python 2.6 or 2.7 installed (most operating systems and distributions do) as well as Python's Setuptools (e.g. by sudo apt-get install python-setuptools
on Ubuntu). Navigate to the titus
directory and run
sudo python setup.py install
or
python setup.py install --home=~
to install Titus in your local directory. In the latter case, you will have to set configure PYTHONPATH
for Python to find your installation.
Titus has a suite of unit tests, too. You can invoke them with
python setup.py test
You can also verify that your installation completed properly by starting Python in a new terminal and typing
import titus.version
print titus.version.__version__
on the >>>
prompt. If you get an ImportError
, Python could not find your installation (check PYTHONPATH
or sys.path
).
If the installation failed with a complaint about "python < 3.0"
, remove that requirement from the setup.py
. Some versions of Python (particularly on Windows) have trouble with a less-than requirement, but as long as you're not using Python 3, it is automatically satisfied.
Download the latest package
and install it in your R library with
R CMD INSTALL aurelius_0.8.3.tar.gz
Now you should be able to
library(aurelius)
?json
in an R session.
To validate or execute scoring engines, you will also have to install
Return to the Hadrian wiki table of contents.
Licensed under the Hadrian Personal Use and Evaluation License (PUEL).