YAMAS

YAMAS: Yet Another Machine Learning Automation System

🛠️ YAMAS Platform Setup

This platform is a full development and visualization suite built on Docker Compose. It integrates a language server, editor, file storage, APIs, UI frontends, and backend services—all containerized and connected in a unified environment.

🚀 Quick Start

To set up the entire platform, just run:

docker compose up -d

This command builds and launches all services in the background.

🌐 Main Access Points

You can access the following UIs and services from your browser:

Port	Service	Description
8000	Portal	Web portal main entry point
8001	VSCode Server	Web-based IDE with custom DSL
8002	Visualizer UI	Frontend for the experiment viewer

---

🧱 Services Overview

1. lang-server

• Context: ./editor/dsl/

• Port: 5007

• Purpose: Custom DSL language server for code analysis/autocompletion in the IDE.

• Volumes:

  • ./logs for server logs
  • ./workspaces for shared workspace access

---

2. vscode

• Context: ./editor/vscode/

• Port: 8001 (mapped from internal 8080)

• Environment:

  • Connects to lang-server internally

• Purpose: Web-based VSCode IDE with pre-integrated language server

---

3. fs

• Context: ./storage/fs/
• Port: 8005 (mapped from internal 5000)
• Purpose: File service backend for handling workspace files
• Depends On: vscode

---

4. elasticsearch

• Image: elasticsearch:8.13.0
• Port: 9200
• Purpose: Search and indexing service for data analysis layer (DAL)
• Volumes: Persists data in dal_data volume

---

5. dal

• Context: ./dal/

• Port: 8004 (mapped from internal 5000)

• Purpose: Python (Flask) backend that queries Elasticsearch for data analytics

• Depends On: elasticsearch

• Environment:

  • Connects to Elasticsearch on the Docker network

---

6. vis-api

• Context: ./visualizer/backend/

• Port: 8003 (should be explicitly mapped in your config)

• Purpose: Backend for the visualization tool

• Volumes:

  • .env for environment configs
  • .output for storing run outputs

• Restart Policy: unless-stopped

Note: You may want to explicitly map this port like "8003:8080" if needed externally.

---

7. vis-front

• Context: ./visualizer/frontend/
• Port: 8002 (mapped from internal 80)
• Purpose: UI for the visualizer, connected to vis-api
• Depends On: vis-api

---

8. mongo

• Image: mongo:4.4.6

• Exposed Port: 27017 (internal only)

• Purpose: Persistent NoSQL database

• Credentials:

  • Username: admin
  • Password: admin

---

9. portal

• Context: ./portal/web-app/
• Port: 8000 (mapped from internal 7001)
• Purpose: Public-facing entry to the overall platform

---

🗂️ Volumes

Volume	Description
dal_data	Persistent Elasticsearch data
mongo_data	Persistent MongoDB database files

---

🌐 Networking

All services are connected via the custom Docker bridge network yamas, allowing them to communicate using service names (e.g., lang-server, vis-api, etc.).

---

🧹 Maintenance

To stop and remove all running services:

docker compose down

To rebuild everything cleanly:

docker compose down -v --remove-orphans
docker compose up --build -d

---

🐚 Debugging

To access a container's shell:

docker exec -it <container_name> /bin/sh

---

📌 Notes

• Make sure port 8003 is mapped for vis-api if you need external access.
• All workspace and log data is shared via mounted volumes (./workspaces and ./logs).