It is a java program to process users' input data into Machine Learning Inference input (i.e. a tensor).
It works as follows:
- Listen messages from Kafka streams topic
preprocess-request. - Get the actual data from storage (
mongodb) according to the information from the message. - Transform the data to input tensor for creating Inference Request.
- Stream the Inference Request message to topic
inference-request. - Repeat the process.
# Build the jar file
mvn clean compile assembly:single
# Run the program
java -jar <Path to .jar file>