Droq Registry Service

Registry service for mapping executor nodes to their supported components.

Overview

The Droq Registry Service maintains a mapping of executor nodes to their metadata and supported components. It uses Git submodules for node management and a SQLite database as a key-value store, automatically bootstrapped from extracted JSON configurations.

Features

• Git Submodule Management: Nodes managed as Git submodules under /nodes/
• Automatic Configuration Extraction: Extracts node.json files from submodules to assets/nodes/
• Key-Value Store: SQLite database for storing node metadata and component mappings
• Automatic Bootstrap: Database is automatically initialized and populated from extracted configurations on startup
• Component Mapping: Tracks which components each node supports
• RESTful API: Provides endpoints to query nodes and their components

Architecture

• Git Submodules: Node repositories stored in /nodes/ directory as Git submodules
• Configuration Extraction: Scripts extract node.json from each submodule to assets/nodes/
• Database: SQLite (registry.db) with two tables:
  • nodes: Stores node metadata
  • components: Stores component mappings (component_class -> module_path) for each node
• Assets: Generated JSON files in assets/nodes/ define node configurations (auto-extracted from submodules)
• Bootstrap: On startup, the service reads all JSON files from assets/nodes/ and populates the database

Adding a New Node

📖 See docs/node-addition-guidelines.md for complete instructions on adding new executor nodes to the registry.

Quick Summary:

• Create repository with node.json configuration
• Submit PR to add as Git submodule
• Registry automatically extracts and integrates your node

API Endpoints

GET /api/v1/nodes

Returns list of all active executor nodes with their metadata and component counts.

GET /api/v1/nodes/{node_id}

Returns detailed information about a specific node, including full component mapping.

GET /api/v1/components/{component_class}/node

Returns the executor node that handles a specific component class.

Example:

curl http://localhost:8002/api/v1/components/TextInputComponent/node

Response:

{
  "node": {
    "node_id": "langflow-executor-node",
    "name": "Langflow Executor Node",
    "api_url": "http://localhost:8000",
    "ip_address": "127.0.0.1",
    ...
  },
  "components": {...},
  "module_path": "lfx.components.input_output.text"
}

Response:

{
  "nodes": [
    {
      "metadata": {
        "node_id": "langflow-executor-node",
        "name": "Langflow Executor Node",
        "description": "...",
        "source_code_location": "../nodes/langflow-executor-node",
        "docker_image": "langflow-executor-node:latest",
        "deployment_location": "local",
        "api_url": "http://localhost:8000",
        "status": "active",
        "supported_components": ["TextInputComponent", "ChatInput", ...]
      },
      "components_count": 348
    }
  ],
  "total_nodes": 1
}

GET /api/v1/nodes/{node_id}

Returns detailed information about a specific node, including full component mapping.

Response:

{
  "node": {
    "node_id": "langflow-executor-node",
    "name": "Langflow Executor Node",
    "description": "...",
    "source_code_location": "../nodes/langflow-executor-node",
    "docker_image": "langflow-executor-node:latest",
    "deployment_location": "local",
    "api_url": "http://localhost:8000",
    "status": "active",
    "supported_components": ["TextInputComponent", "ChatInput", ...]
  },
  "components": {
    "TextInputComponent": "lfx.components.input_output.text",
    "ChatInput": "lfx.components.input_output.chat",
    ...
  }
}

Database

The service uses SQLite as a key-value store. The database file (registry.db) is created automatically on first startup. You can customize the database path using the REGISTRY_DB_PATH environment variable.

Database Schema

nodes table:

• node_id (PRIMARY KEY): Unique identifier for the node
• name: Display name
• description: Node description
• source_code_location: Path to source code
• docker_image: Docker image name
• deployment_location: "local" | "cloud" | "k8s"
• api_url: API endpoint URL
• status: "active" | "inactive" | "deploying"
• metadata_json: Full JSON configuration (for flexibility)
• created_at, updated_at: Timestamps

components table:

• id (PRIMARY KEY): Auto-increment ID
• node_id (FOREIGN KEY): References nodes.node_id
• component_class: Component class name
• module_path: Python module path
• created_at: Timestamp
• UNIQUE constraint on (node_id, component_class)

Running Locally

./start-local.sh
# run all services
./stat-all-local.sh

The service will start on http://localhost:8002.

Environment Variables

• HOST: Server host (default: 0.0.0.0)
• PORT: Server port (default: 8002)
• LOG_LEVEL: Logging level (default: info)
• RELOAD: Enable auto-reload (default: true)
• REGISTRY_DB_PATH: Path to SQLite database file (default: registry.db)