Skip to content

Latest commit

 

History

History
157 lines (128 loc) · 4.13 KB

README.md

File metadata and controls

157 lines (128 loc) · 4.13 KB

Agent Config Microservice

Overview

The Agent Config microservice is responsible for managing agents, their configurations, and tools. It serves as an intermediary for the Converse microservice, providing necessary agent context and tools for communication with AI systems like OpenAI or Amazon Bedrock.

Documentation

Puedes encontrar la colección completa de Postman con todos los endpoints aquí: Colección Postman

Construcción y Ejecución

Usando Docker

  1. Construir la imagen Docker (incluye la compilación):
docker build -t agent-config-ms .
  1. Ejecutar el contenedor:
docker run -p 8080:8080 agent-config-ms
  1. Para detener el contenedor:
docker stop $(docker ps -q --filter ancestor=agent-config-ms)

El servicio estará disponible en http://localhost:8080

Construcción local (alternativa)

Si prefieres construir y ejecutar localmente sin Docker:

./gradlew clean build
./gradlew run

This microservice is written in Kotlin and uses PostgreSQL as its database. It has three main tables:

  • agent_configs: Stores the configuration details of each agent.
  • tools: Stores the details of various tools.
  • agent_tool: Represents the relationship between agents and their tools.

Features

  1. Agent Configuration Management: Create and manage agent configurations.
  2. Tool Management: Add, associate, and remove tools for agents.
  3. Agent Retrieval:
    • Retrieve an agent by its agent_id.
    • Retrieve an agent by query similarity via a separate microservice called Vector DB.

Endpoints

1. Create an Agent Configuration

Endpoint:

POST /api/ms/agent/config

Request Headers:

Content-Type: application/json

Request Body:

{
    "agent_id": "mi_primer_agente",
    "preferences": {
        "example_response": "dasdasdas"
    }
}

Description: Creates a new agent configuration.

2. Create a Tool

Endpoint:

POST /api/ms/agent/tool

Request Headers:

Content-Type: application/json

Request Body:

{
    "name": "mi_primera_herramienta",
    "description": "Descripción de la herramienta",
    "type": "Tipo de la herramienta",
    "config": {"xasxa": "xaaz"}
}

Description: Creates a new tool.

3. Associate a Tool with an Agent

Endpoint:

POST /api/ms/agent/config/{agent_id}/tools/{tool_name}

Description: Associates an existing tool with a specific agent.

Example:

curl --location --request POST 'http://localhost:8080/api/ms/agent/config/mi_primer_agente/tools/mi_primera_herramienta'

4. Remove a Tool from an Agent

Endpoint:

DELETE /api/ms/agent/config/{agent_id}/tools/{tool_name}

Description: Removes the association between a tool and an agent.

Example:

curl --location --request DELETE 'http://localhost:8080/api/ms/agent/config/mi_primer_agente/tools/mi_primera_herramienta'

5. Retrieve an Agent by Query Similarity

Description: To retrieve an agent by query similarity, the Agent Config microservice communicates with the Vector DB microservice. This external microservice handles the similarity search and returns the most relevant agent.

Technologies Used

  • Programming Language: Kotlin
  • Database: PostgreSQL
  • API Framework: Micronaut
  • Containerization: Docker

Running Locally

  1. Clone the repository.
  2. Ensure PostgreSQL is installed and running.
  3. Update the database configuration in the application properties.
  4. Run the application using: 
    ./gradlew run

Notes

  • Make sure the Vector DB microservice is up and running to use the query similarity feature.
  • Customize the database schema and table names as needed.

Future Improvements

  • Add more detailed logging.
  • Enhance security features such as authentication and authorization.
  • Implement rate limiting for API endpoints.