Skip to content

Latest commit

 

History

History
54 lines (37 loc) · 2.9 KB

architecture.md

File metadata and controls

54 lines (37 loc) · 2.9 KB
Raw

Norman Architecture

This document provides an overview of Norman's architecture, its key components, and the relationships between them. A high-level understanding of the architecture will help developers work with the project more effectively.

Table of Contents

High-Level Architecture

Here is a text-based representation of Norman's high-level architecture:

+---------------------+     +--------------------+     +-------------------+
| Chat Platform       | <-> | Connector          | <-> | FastAPI App       |
| (Slack, IRC, etc.)  |     | (Slack, IRC, etc.) |     | (app/main.py)     |
+---------------------+     +--------------------+     +-------------------+
                                          |                  |
                                          v                  v
                                +-----------------+  +---------------------+
                                | Channel Filter  |  | Action              |
                                | (app/models)    |  | (app/models,        |
                                +-----------------+  | app/actions)        |
                                                     +---------------------+

Key Components

  • Chat Platforms: External chat services that Norman interacts with, such as Slack and IRC.

  • Connectors: Modules that handle communication between the chat platforms and Norman. Each connector is responsible for a specific chat platform.

  • FastAPI App: The core application that powers Norman. It is responsible for handling API requests, managing the database, and executing actions based on the filters.

  • Channel Filters: Database models that define the rules for triggering actions in response to messages from chat platforms.

  • Actions: Modules that define the actions to be performed when a channel filter is triggered. Actions can include generating replies, fetching data, or performing other tasks.

Data Flow

  1. A message is received from a chat platform.
  2. The connector for the chat platform forwards the message to the FastAPI app.
  3. The FastAPI app checks the message against the channel filters.
  4. If a channel filter is triggered, the corresponding action is executed.
  5. The action may involve generating a reply, which is then sent back through the connector to the chat platform.

Extending Norman

Norman can be extended and customized in various ways, such as adding new connectors, creating custom actions, or modifying the core functionality. For more information, refer to the Extending Norman documentation.

You can create a visual diagram based on the text representation above using tools like draw.io, Lucidchart, or PlantUML.