paraphraser-api

Backend for paraphrasing tool leveraging LLM APIs.

Implemented as an AWS Lambda application and utilizes Serverless Framework for deployment.

Prerequisites

• Docker
• Docker Compose
• Make

Endpoints

• POST /paraphrase

  • available providers: "chatgpt", "gemini"
  • available tones: "formal", "amicable", "fun", "casual", "sympathetic", "persuasive"
  • sample request:

    {
        "provider": "chatgpt",
        "tone": "formal",
        "text": "I'm hungry. What's for dinner?"
    }
    

  • sample response:

    {
        "result": "I am currently experiencing hunger. May I inquire about the menu for this evening's meal?"
    }
    

• GET /providers

  • sample response:

    {
        "providers": [
            "chatgpt",
            "gemini"
        ]
    }
    

• GET /tones

  • sample response:

    {
        "tones": [
            "formal",
            "amicable",
            "fun",
            "casual",
            "sympathetic",
            "persuasive"
        ]
    }
    

Usage

configure

$ make .env

• see generated .env file for configuration

tidy dependencies

$ make deps

run unit tests

$ make test

run all tests (unit + integration tests)

$ make testIncludeInt

build serverless functions

$ make build

• this generates bin directory to be used in deployment

deploy serverless application

$ make deploy

Helpers during development:

format all .go files in project (using go fmt)

$ make fmt

generate test mocks (to be used with stretchr/testify) for all interfaces in project

$ make mocks

• can be configured in .mockery.yaml