Skip to content

Latest commit

 

History

History
95 lines (63 loc) · 2.54 KB

README.md

File metadata and controls

95 lines (63 loc) · 2.54 KB
Raw

assistant

Assitant built with ChatGPT!

Quick start

  1. Clone the repo
git clone https://github.com/bjwswang/assistant
  1. Install dependencies
cd assistant
go mod tidy
  1. Set configuration in assistant.json

see [Configurations](## Configurations) for more details

  1. Build the assistant server and CLI
make build

When build is done,you will get two binaries in bin directory.

  • assistant is the assistant server
  • acli is the assistant CLI which can interact with the assistant server
  1. Start the assistant server
./bin/assistant --config assistant.json
  1. Test the assistant
curl -XGET http://localhost:9999

Output should be

Welcome to AI Assistant 👋!

Configurations

Parameter Description Default
addr The address which assistant server will watch :9999
assistant.api_key OpenAI api key sk-xxx
assistant.chat.xxx OpenAI model configuration for Chat model:gpt-3.5-turbo temperature:0.5 max_tokens:100
assistant.unit_test.xxx OpenAI model configuration for generating unit tests model:gpt-3.5-turbo temperature:0.5 max_tokens:100
fiber.xxx Fiber related parameters see the official document

APIs

  1. Normal chat
  • path: /chat
  • method: post
  • paramters:
    • question: the question you want to chat with the assistant
  1. Generate unit test
  • path: /ut
  • method: post
  • paramters:
    • code: the code you want to used for generating unit tests

CLI

  1. Chat with assistant
./bin/acli --server http://localhost:9999 chat --question "What is AI assistant in 10 words?"
  1. Generate unit tests
./bin/acli --server http://localhost:9999 ut --file {filepath_to_source_code_}"

Contribute to assistant

Welcom to contirbute to this AI assistant!