TS-GPT

Introduction to TS-GPT

TS-GPT is a sophisticated project built on the SpringCloud Alibaba and SpringBoot 3 frameworks, designed to leverage the power of GPT (Generative Pre-trained Transformer) models in various applications. It provides a seamless integration for users to engage with AI models for text and image-based interactions. Below, we explore the features, architecture, and deployment requirements of TS-GPT in detail.

Key Features

The TS-GPT project offers a range of functionalities:

• Microservice and Monolithic Versions: The project supports both microservice (master branch) and monolithic (dec branch) architectures. The microservice version, suitable for high-performance environments, requires a minimum of a 2-core, 4GB RAM system for deployment. The monolithic version, on the other hand, can run on lower-tier systems, including basic computing devices.

• Customizable AI Interactions: Users can engage in dialogues using generic or bespoke GPT models. Features include GPT-based image recognition, image generation, and error reporting panels to help filter and address issues.

• AI Writing and Image Generation: The project supports AI-powered content creation and image generation through DALL and SD modules, which offer text-to-image and real-time image rendering, respectively.

• Custom Discussions and Appearances: Users can preset dialogue parameters and customize the appearance of the interface to suit their preferences.

• Additional Features: Includes voice interaction and dialogue support through mini-programs, as well as various display options for preset words and appearance customization.

Project Architecture

TS-GPT encompasses a diverse architecture structured around both front-end and back-end components:

• Front-End (Web Interface):

  • Built with vue3 and element plus for an intuitive and interactive user experience.

• Back-End (Microservices):

  • ts-auth: Manages user authentication, information, transactions, and payments.
  • ts-gateway: Acts as a service gateway, handling authentication, routing, load balancing, and transaction management.
  • ts-common: Offers common utilities, configurations, and error interception.
  • ts-api: Hosts public DTOs and remote interface calls.
  • ts-drawing: Manages drawing-related functionalities.
  • ts-chat: Oversees chat-related functions.

These services are supported by an assortment of technologies like SaToken, Redis, Maven, and SpringCloud tools.

• Mini-Program Interface:
  • Implemented using uniapp and vue3, enabling interaction via WeChat mini-programs.

Deployment Environment and Tools

Deploying TS-GPT requires several pre-requisite tools and services:

• Essential Tools: Docker, MySQL 8, Redis 7, Nacos 2.x, and JDK 17.
• Optional Services: Aliyun OSS, Baidu Translation API, and SD rendering features.
• Server Requirements: A minimum of 2-core CPU and 4GB RAM is recommended for single-machine deployment.

Installation and Operation

To set up the project, follow these steps:

Front-End:

1. Install Yarn:

   npm install --global yarn
   

2. Install Dependencies:

   yarn install
   

3. Run the Local Environment:

   yarn run serve
   

4. Build for Production:

   yarn run build
   

Back-End:

• The back-end requires setting up with .sql files for database schema and initial data. Use .nacos files to configure and import into Nacos.

Important Notes

• Reload configurations every time a new model is added or when configuration changes occur to ensure these are applied.
• The project requires minimum server specifications due to the resource needs of Nacos, MySQL, and Redis.
• While the project supports various features, shared SD service setups across third-party projects may cause issues due to real-time rendering constraints.

Concluding Thoughts

TS-GPT represents a powerful yet flexible solution for building GPT-powered applications. It supports a variety of AI interactions and simplifies the integration of advanced AI capabilities into web and mini-program environments. The project is positioned as a versatile tool capable of catering to a wide range of user requirements in AI implementations.