How nkanyezingobese/ares_1 works
Project Type and Purpose
• Ares-1 is an open-source, local dev environment focused on streaming synthetic drilling telemetry, running a physics baseline, and placeholder ML anomaly detection.
• It is not a web front-end; rather, it is a data-streaming and processing pipeline with a visualization stub (Unity) and a configuration layer (Agents & Docs).
Top-Level Components (to show as distinct blocks)
a. Infrastructure
• MQTT Broker (Mosquitto, Docker Compose)
• Virtual environment (.venv) and Python runtime
b. Data Publisher
• Excel metrics (“Ares-1 terrain metrics.xlsx”)
• Karoo Telemetry Publisher (python/ares1/telemetry/ares_karoo_publisher.py)
• Publishing script (python/scripts/publish_telemetry.py)
c. Data Consumers & Processors
• Telemetry Subscriber (python/scripts/subscribe_telemetry.py)
• Event Subscriber (python/scripts/subscribe_events.py)
• Anomaly Detector (python/scripts/run_anomaly_detector.py)
• ML stub (python/ares1/ml/anomaly.py)
• Physics baseline (python/ares1/physics/torque_drag.py)
d. Persistence
• Outputs folder (CSV, JSON, JSONL logs)
e. Visualization Stub
• Unity project (unity/ares1_unity_project/Ares-1)
• MQTT bridge scripts (Assets/Ares1/Scripts/*MqttTopicMap.cs)
• UI scripts (TelemetryWidget, UdpTelemetryReceiver, etc.)
f. Configuration & Documentation
• Agents context (mission_context.md)
• Engineering docs (docs/architecture, grounding templates, planned data flow)
Relationships & Data Flows
• Telemetry Publisher → MQTT Broker (“telemetry” topic)
• MQTT Broker → Telemetry Subscriber → Outputs/telemetry_log.csv & telemetry_latest.json
• MQTT Broker → Anomaly Detector → processes physics & ML → publishes anomaly events to “events” topic
• MQTT Broker → Event Subscriber → Outputs/events_log.jsonl
• MQTT Broker → Unity MQTT bridge → real-time visualization in Unity
• Excel metrics → Karoo Publisher script → feeds Telemetry Publisher
• .env & requirements ensure correct configuration and dependencies
Key Architectural Patterns
• Pub/Sub messaging via MQTT for loose-coupling
• Modular Python scripts for each processing stage
• Clear separation: Data ingestion, processing, persistence, visualization
• Extensibility points: add new ML modules in python/ares1/ml, new physics baselines in python/ares1/physics, new Unity visualizations
Detailed Capture
• Docker-Compose file under MQTT block.
• Venv and requirements under Python runtime block.
• Call out that agents/ and docs/ are the single source of truth for mission context and architecture.
• Excel metrics → Karoo Publisher → Python Telemetry Publisher → MQTT Broker
• MQTT Broker → { Telemetry Subscriber → CSV/JSON ; Event Subscriber → JSONL ; Anomaly Detector → processing & publishes back to Broker }
• MQTT Broker → Unity client for visualization
• Config & Docs feeding mission context into scripts
This level of detail ensures every architectural element and data flow in Ares-1 is visible at a glance.