Smart Cities Data Platform

Project Description

The platform is a combination of Elixir micro services custom built to ingest, normalize, transform, persist, and stream data from numerous sources, orchestrated via Kubernetes in any cloud provider or on-prem Kubernetes deployment. The loosely coupled services pass data across the pipeline via Kafka message queues and persist data to any hyper-scalable object store providing the S3 standard. They coordinate and communicate via a single event bus, also running on top of Kafka. The distributed data files are persisted and retrieved via SQL queries processed by the PrestoDB engine. Finally, user access, discovery, and analysis is facilitated by a ReactJS web application user interface, a RESTful API, or a web socket API for streaming data feeds.

Microservices

Application	Short Description	Build Status
Andi	Admin Interface for creating/editing datasets to be ingested
Discovery API	API to search for and query datasets
Discovery Streams	Websocket connection to listen to streaming data
Estuary	Microservice to persist event stream events
Forklift	Microservice for saving data to Presto DB
Reaper	Microservice to retrieve data
Valkyrie	Microservice to validate data structure during ingestion
Alchemist	Microservice to alter data from its original format

Prerequisites

General Prerequisites

• Elixir - The primary language that all of the microservices are written in
• Docker - All microservices are built as docker images
• Apache Kafka - Communication mechanism between microservices
• Redis - General purpose storage and caching
• Elasticsearch - Used by Discovery API for search
• PostgreSQL - General purpoase storage
• Presto - Big Data storage of ingested data
• Vault - Secure storage of secrets

Development Enviornment Prerequisites

• General Setup Information
• Windows
• macOS
• Linux

Usage

The microservices written in Elixir use Mix as the build tool.

Building

Each microservice under the apps/ directory has a Dockerfile that can be used to build that microservice individually by running the following command:

docker build .

Additional app specific build steps will be in the relative readme at apps/{app}/readme.md.

Testing

• Unit tests can be executed from the root of this repository or a specific application under the apps/ directory

mix test

• Integration tests can be executed from the root of this repository or a specific application under the apps/ directory

mix test.integration

• End to End (E2E) Tests can be executed from the root of this repository.

mix test.e2e

Execution

How to run and use the code

Additional Notes

• What is the project and how it works
• What all those application names mean
• Additional learning resources
• A glossary of terms and technologies
• Starting All of the Microservices

Version History and Retention

Each microservice is released independently and can be found here in the Releases section. All releases will be kept indefinitely.

License

Released under Apache 2 license.

Contributions

How to contribute

Contact Information

Acknowledgements