An example service used for converting events to useful graph data.

The service uses watermill and arangoDB for showcasing how to convert events, coming from an event stream, to useful graph data that can be easily searched and queried. It performs two main tasks:

• tries to mimick an event stream by generating series of events
• converts events to graph data; to vertices and edges

This part of the service tries to mimick an event stream by generating events. The domain context of this service is a simple marketplace where users can view and purchase items. Two types of events are generated: creational events like user_registered and item_created, and relational events like item_viewed, item_dropped, item_purchased, item_delivered and item_not_delivered.

This part of the service receives events and converts the event payload to graph data. Creational events are converted to matching vertices, while relational events are converted to matching edges between vertices.

The easiest way to run both the service and the ArangoDB graph database isntance is to run docker-compose up. After ArangoDB instance and the service are set up, the service will start generating events and saving them to the graph database. The service will log some info for each step of the process, finishing with done generating events!.

If the build process for the ArangoDB container starts to fail, uncomment the command: --database.auto-upgrade line to run the database auto upgrade once.

If you run the service and the graph database separately, bear in mind that service deletes the database and its data every time the service is started.

The graph data can be checked out with ArangoDB's web interface on http://localhost:8529. Access it with credentials root:rootpassword and choose the example_db database.

ArangoDB offers the data in three formats: key-value, json and graph. To view the entire graph that the service generated, go to the Graphs section and choose example_graph. Part of the graph should be loaded.

Additionally, you can load the entire graph by clicking on the graph icon in the toolbar. Or you can style your vertices and nodes by adding colors, names, attributes, etc.

For more details on how to use AQL, the ArangoDB query language, you can visit this link. Until then, here are some queries you can try out in the Queries tab.

Get all available fields for available items

FOR item in items
RETURN item

Get information on all users

FOR user IN users
RETURN {
    name: user.name,
    age: user.age
}

Get information on failed deliveries

FOR user IN users
    FOR item, delivery IN 
    INBOUND user not_delivered
    RETURN {
        user_name: user.name,
        item_title: item.title,
        user_address: delivery.address,
        reason: delivery.reason
    }

If you want to see the former displayed as a graph, try out this one below (to display results as a graph, returning data has to have _from and _to fields)

FOR v in not_delivered
RETURN v 

Get stats on all items

FOR item IN items
    LET p = (FOR e IN 1..1 INBOUND item purchased RETURN 1)
    LET v = (FOR e IN 1..1 INBOUND item viewed RETURN 1)
    LET d = (FOR e IN 1..1 INBOUND item dropped RETURN 1)
    LET del = (FOR e IN 1..1 OUTBOUND item delivered RETURN 1)
    LET n = (FOR e IN 1..1 OUTBOUND item not_delivered RETURN 1)
    RETURN {
        item_id: item.id,
        item_name: item.title,
        viewed: COUNT(v),
        dropped: COUNT(d),
        purchased: COUNT(p),
        delivered: COUNT(del),
        not_delivered: COUNT(n)
        }