• First, build a VM with featurestore support,
• then create a project with the featurestore service enabled.
• run ./run.sh ports to find which exported port hopsworks is running on
• check inside the project what id it got and what id it's featurestore got
• create the sample datasets by running the notebook in ./sample_data/
• build the fat-jar by running cd feature_engineering_spark; sbt assembly
• upload the fat-jar to the project's Resources folder
• create a new dataset in the project called sample_data
• upload the sample datasets (e.g alerts.csv, hipo.csv, kyc.csv, police_reports.csv etc) to sample_data
• create the sample jobs by running:

$ ./feature_engineering_spark/create_jobs.sh <PORT> <PROJECTNAME> <PROJECTID>

e.g:

$ ./feature_engineering_spark/create_jobs.sh 55354 hopsworksdemo 8

• create the sample featuregroups by running:

$ ./feature_engineering_spark/create_fgs.sh <PORT> <PROJECTNAME> <PROJECTID> <FEATURESTOREID> <JOBIDINIT>

Where is the id of the first job (it is assumed that the rest of the jobs follow with consecutive id's, if you created a fresh VM/project this will be 1) e.g:

$ ./feature_engineering_spark/create_fgs.sh 55354 "hopsworksdemo" 8 3 1

It will create the following feature groups