Demo

Description

For demonstration purposes, "Real-Time recommendation Service" is able to run without the existence of a real application, using two secondary programs. The first program (item_producer.py) randomly selects entries (item ids) from (news_category_dataset_sample.json) (original dataset can be found here) and publishes them on "items" topic. It can do this either once or repeatedly. Then, the second program (user_interactions_producer.py) creates some random user ids, repeatedly combines them randomly with item ids in "items" topic as interactions and publishes them on "interactions" topic.

Demos's functionality, that described previously, is shown in the following architecture diagram:

Set Up

1. Repeat the following instructions included in system's Deployment Guide:
   • Instruction 1: Start Apache Kafka (Step 2. only) (if its not already started).
   • Instruction 3: Activate "venv" virtual environment (if created one previously and not already activated).
   • Instruction 7: Run "model_builder.py" program.
   • Instruction 8: Run "recommender.py" program.
2. Move to "demo" directory: cd demo.
3. Run "item_producer.py" and create some items.
4. Run "user_interactions_producer.py":
   • Create some users first.
   • Create interactions.