Skip to content

ai-driven-development-labs

An experimental repository for practicing and validating AI-Driven Development workflows. It combines an embedded-oriented subject (an IMU sensor HAL implementation) with a local observability stack that supports its operation, so the full cycle — design, implementation, testing, documentation, and operations — can be iterated on together with AI assistants such as GitHub Copilot.

Source code is hosted on GitHub: https://github.com/ks6088ts-labs/ai-driven-development-labs

What's included

  • IMU HAL (Hardware Abstraction Layer) — ai_driven_development_labs/imu/
    • Abstraction over sensors and buses (I2C / SPI) via the ISensorHAL / IBusDriver interfaces
    • Real-device implementations for STMicroelectronics (e.g. LSM6DSO) and TDK InvenSense (e.g. ICM-42688-P)
    • Mock implementations for testing, a CLI to initialise and read sensors, and OpenTelemetry-based telemetry
    • Design details: IMU HAL documentation
  • Local observability stack — compose.observability.yml / docker/observability/
    • One-command (make obs-up) OTel Collector → Prometheus / Jaeger → Grafana pipeline
    • Pre-provisioned Grafana data sources and an IMU dashboard
    • Setup and operations guide: Observability documentation
  • Development & CI/CD foundation
    • Package management with uv, linting with ruff / ty / pyrefly, and tests with pytest
    • Dockerfile and GitHub Actions workflows that publish images to Docker Hub and GHCR
    • MkDocs (Material + i18n) documentation site with automated deployment

Prerequisites

Development instructions

Local development

Use Makefile to run the project locally.

# help
make

# install dependencies for development
make install-deps-dev

# run tests
make test

# run CI tests
make ci-test

Docker development

# build docker image
make docker-build

# run docker container
make docker-run

# run CI tests in docker container
make ci-test-docker