ML Scoring (REST API) - h2o.ai

h2o_scoring.R

## training a model

library(h2o)
h2o.init(nthreads = -1)

dx_train <- h2o.importFile("https://s3.amazonaws.com/benchm-ml--main/train-0.1m.csv")

md_rf <- h2o.randomForest(x = 1:(ncol(dx_train)-1), y = ncol(dx_train), training_frame = dx_train, 
                    model_id = "h2o_RF",
                    ntrees = 100, max_depth = 10, nbins = 100)



## exporting model for scoring

h2o.download_mojo(md_rf, path = "./h2o")



## building prediction service

# (need jetty-runner.jar ROOT.war from Steam)
java -jar jetty-runner.jar ROOT.war

curl -X POST --form mojo=@h2o_RF.zip --form jar=@h2o-genmodel.jar \
              localhost:8080/makewar > h2o_RF_MOJO.war



## run prediction service

java -jar jetty-runner.jar --port 20000 h2o_RF_MOJO.war



## score via REST API

time curl "http://localhost:20000/predict?Month=c-8&DayofMonth=c-21&DayOfWeek=c-7&DepTime=1934&UniqueCarrier=AA&Origin=ATL&Dest=DFW&Distance=732"
# (fast scoring needs JVM to warm up with a few requests)