lens ml module is an optional add on service - the dependencies of lens ml are not loaded by default in the lens server. the ml dependencies have to be exported to LENS_EXT_CLASSPATH, which is automatically added into the lens server classpath.
- there is a separate distribution that gets created for lens ml. Use LENS_EXT_CLASSPATH=$LENS_EXT_CLASSPATH:`$LENS_ML/bin/lens-ml-classpath.sh` to export the lens ml classpath before the starting the lens server - you can find the sample conf that loads the ml module in tools/conf-pseudo-distr/server - it has been tested with spark 1.3 Alternatively you can try running via docker. Docker sets the above environment for you automatically. The ml module does not have the CLI shell commands as of now. So only the way to run some models is via running MLRunner: bin/lens-run-class.sh org.apache.lens.ml.MLRunner $LENS_ML/data/naive_bayes --conf $LENS_CLIENT_CONF Sharad On Mon, May 25, 2015 at 12:38 PM, Yash Sharma <[email protected]> wrote: > Hi All, > I was trying to play around with Lens ML module and was not able to get lot > of information on the setup and starting guidelines. > > I was going through the deployment section in docs [1] but could not figure > out lot of things. Is there any resource for setting up ML module and > trying out some sample models through the CLI. > Also what is the recommended version of Spark. > > Any tips would be appreciated. I would be happy to enhance the docs based > on my findings. > > Thanks > > 1. https://lens.incubator.apache.org/user/ml.html >
