This project is a project that provides R&R to GPT and progresses programming step by step.
(
️ Experimental ) This project is currently experimental and may consume Open AI tokens, so caution is required when running it.
This project tasks GPT with programming, but grants R&R to AI for each step.
??⚕️ Architect -> ??♀️ Tech Evangelist -> (?? Lead -> ??? Junior -> ?? QA -> ??? Senior)
We are responsible for designing according to the initial request. First, we select 5 plans, select the best plan, and improve the plan by holding internal improvement meetings up to 10 times until the quality of the development plan is improved.
Based on the plan completed by the architect, we create a list of project specifications and reference technologies.
We program for the first time, ask for reviews from actual users, and then modify the development according to their requests.
We create test scripts and find ways to make the developed functions more stable, leading to complementary development.
Receive actual test results and generate a report analyzing test results, causes, and solutions.
After reading the written program, any errors that occurred, and the QA report, we attempt to resolve the errors by modifying the script.
The process from Lead to Senior is repeated until all actual tests are passed.
Enter the command below to clone the project and install it. (This project requires Node.JS.)
git clone https://github.com/hmmhmmhm/friday-gpt
cd ./friday-gpt
npm install
Create an .env file at the top of the project and fill in the contents below.
OPENAI_API_KEY is essential for operation and can be issued here. Open AI Get API keys
If you enter AZURE_SPEECH_KEY, the progress will be informed using TTS. (You do not need to enter any information.)
*AZURE_** related input values can be issued here. Azure Speech Services
OPENAI_CHAT_MODEL=gpt-3.5-turbo # 또는 gpt-4
OPENAI_API_KEY=
AZURE_SPEECH_KEY=
AZURE_SPEECH_REGION=
AZURE_SPEECH_VOICE=ko-KR-SeoHyeonNeural
Open the src/input.ts file and set the desired development direction in the userRequest variable.
export const userRequest =
'문자열로 두 개의 숫자 값을 입력 받아서 마치 사람이 계산하는 방법처럼 각 단위의 숫자를 비교해서 무제한 길이로 사칙연산을 할 수 있는 Typescript 라이브러리를 작성해줘'
Execute the command below to send a request to AI.
npm run dev
Alternatively, you can add prompt on the command line as follows:
npm run dev 여러가지 순열들을 제공해주는 Typescript 라이브러리를 제공해 줘
All result files are created in the result folder.
MIT Licensed.