Skip to content

Latest commit

 

History

History
296 lines (265 loc) · 7.69 KB

README.md

File metadata and controls

296 lines (265 loc) · 7.69 KB

Orquestradora-LabDeIA

This project utilizes Flask, a Python microframework, to create a web application for managing assistants, chats, and documents for a project by Creath Digital. Here's a brief overview of the project structure and functionalities:

Project Structure

The project is organized into three main sections: routes, models, and database interaction.

Routes

IMPORTANT

All routes (except for the /chats/enviarMensagemLLM route) require a LOGIN TOKEN to be passed in the header. That's 'cause they should be run by a logged client on our frontend. The token can be obtained by logging in with the /auth/login route.

// header:
{
  "Authorization": "Bearer <token>"
}

Assistants

  • POST /assistants/add: Add a new assistant. 
    // request:
{
  "name": "Assistant Name",
  "owner_id": "Owner id"
}
  • GET /assistants/list: List all assistants. 
    // response:
[
  {
    "_id": "id",
    "name": "Assistant Name",
    "owner_id": "owner id",
    "api_key": "api key",
    "createdAt": "2021-09-01T00:00:00"
  }
]
  • GET /assistants/get/<id>: Get details of a specific assistant. 
    // response:
{
    "_id": "id",
    "name": "Assistant Name",
    "owner_id": "owner id",
    "api_key": "api key",
    "createdAt": "2021-09-01T00:00:00"
}
  • PUT /assistants/update/<id>: Update details of a specific assistant. 
    // request:
{
    "_id": "id",
    "name": "Assistant Name",
    "owner_id": "owner id",
    "api_key": "api key",
    "createdAt": "2021-09-01T00:00:00"
}
  • DELETE /assistants/delete/<id>: Delete a specific assistant.

Chats

  • POST /chats/enviarMensagemLLM: Send a message to LLM. 
    // header:
{
    "X-API-Key": "api key"
}
    // request:
{
  "id": "+5513988593464",
  "message_service": "Whatsapp",
  "username": "nome",
  "active": true,
  "message": "mensagem do usuario"
}
  • GET /chats/get/<id>: Get details of a specific chat. 
    // response:
{
  "_id": "id",
  "channel": {
    "id": "+551300000000",
    "message_service": "Whatsapp"
  },
  "tokens": {
    "in": 1,
    "out": 1
  },
  "username": "nome",
  "assistant_id": "id",
  "active": true,
  "message": "mensagem do Usuario",
  "response": "resposta LLM",
  "createdAt": "05-09-2024T19:57:00"
}
  • GET /chats/list: List all chats. 
      // response:
  [
      {
      "_id": "id",
      "channel": {
          "id": "+551300000000",
          "message_service": "Whatsapp"
      },
      "tokens": {
          "in": 1,
          "out": 1
      },
      "username": "nome",
      "assistant_id": "id",
      "active": true,
      "message": "mensagem do Usuario",
      "response": "resposta LLM",
      "createdAt": "05-09-2024T19:57:00"
      }
  ]
  • GET /chats/listarAtivos: List all active chats. 
      // response:
  [
      {
      "_id": "id",
      "channel": {
          "id": "+551300000000",
          "message_service": "Whatsapp"
      },
      "tokens": {
          "in": 1,
          "out": 1
      },
      "username": "nome",
      "assistant_id": "id",
      "active": True,
      "message": "mensagem do Usuario",
      "response": "resposta LLM",
      "createdAt": "05-09-2024T19:57:00"
      }
  ]
  • GET /chats/listarInativos: List all inactive chats. 
      // response:
  [
      {
      "_id": "id",
      "channel": {
          "id": "+551300000000",
          "message_service": "Whatsapp"
      },
      "tokens": {
          "in": 1,
          "out": 1
      },
      "username": "nome",
      "assistant_id": "id",
      "active": False,
      "message": "mensagem do Usuario",
      "response": "resposta LLM",
      "createdAt": "05-09-2024T19:57:00"
      }
  ]
  • PUT /chats/desativar/<id>: Deactivate a specific chat.
  • PUT /chats/ativar/<id>: Activate a specific chat.

Documents

  • POST /documents/add: Upload a new document. 
    // request form-data:
{
    "file": "file attached"
    "type": "rag or prompt",
    "assistant_id": "id",
}
  • GET /documents/list: List all documents. 
      // response:
  [
      {
      "_id": "id",
      "filepath": "file path",
      "filename": "file name",
      "type": "rag or prompt",
      "assistant_id": "id",
      "createdAt": "2021-09-01T00:00:00"
      }
  ]
  • GET /documents/get/<id>: Get details of a specific document. 
      // response:
  {
        "_id": "id",
      "filepath": "file path",
      "filename": "file name",
      "type": "rag or prompt",
      "assistant_id": "id",
      "createdAt": "2021-09-01T00:00:00"
  }
  • PUT /documents/path/get/<id>: Get the file path of a specific document.
  • GET /documents/download/<id>: Download a specific document.
  • DELETE /documents/delete/<id>: Delete a specific document.
The documents folder architecture is:
Orquestradora-LabDeIA
  └── documents
      ├── <assistant id>
      │   ├── knowledge
      │   │   └── rag.md
      │   └── prompt.txt
      │   
      └── <assistant id>
          ├── knowledge
          │   └── rag.md
          └── prompt.txt

Auth

  • POST /auth/login: Login to the application. 
    // request:
{
    "email": "user email",
    "password": "password"
}
// response:
{
    "token": "access token",
    "user": {
        "email": "user email",
        "id": "id",
        "name": "name"
    }
}
  • POST /auth/register: Register a new user. 
    // request:
{
    "name": "name"
    "email": "user email",
    "password": "password",
}

Models

  • Assistant Model:
    • Attributes: name, company, createdAt
  • Chat Model:
    • Attributes: channel (id and message_service), username, assistant_id, active, message, response, createdAt
  • Document Model:
    • Attributes: filepath, filename, type (rag or prompt), assistant_id, createdAt

Database Interaction

The project utilizes MongoDB as the database. Each model corresponds to a collection in the MongoDB database. We have two packages on the code that are responsible for the database interaction: repositories and services.

  • Repositories: The repositories package contains the classes that interact with the database.
  • Services: The services package contains the classes that handle the business logic of the application.

Additional Information

  • DocumentType Enum: The type of document is defined by an enum called DocumentType, which has two values: "rag" or "prompt".
  • File Storage: Document files are saved within the project folder, and only the file path is stored in the database.

Setup Instructions

To set up the project locally, follow these steps:

  1. Clone the repository.
  2. Run cp .env.example .env to create a new .env file and set it up according to the required.
  3. Install the required dependencies using pip: pip install -r requirements.txt.
  4. Ensure that you have the core application running on your local machine.
  5. Run the Flask application on waitress running: waitress-serve --port=8080 --call main:create_app.

Contributing

If you'd like to contribute to this project, feel free to fork the repository and submit a pull request with your changes.

License

This project is licensed under the MIT License.