apititan / deepl-fastapi

your own deepl server via fastapi, cross-platform (Windows/Linux/Mac)

deepl-fastapi

[]

Your own deepl server via fastapi, cross-platform (Windows/Linux/MacOs) with API for OmegaT

Installation

• Create a virual environment: optional but recommended e.g.,

  # Linux and friends
  python3.7 -m venv .venv
  source .venv/bin/activate
  
  # Windows
  # py -3.7 -m venv .venv
  # .venv\Scripts\activate

pip install deepl-fastapi

or (if your use poetry)

poetry add deepl-fastapi

or

 pip install git+https://github.com/ffreemt/deepl-fastapi.git

or

• Clone the repo https://github.com/ffreemt/deepl-fastapi.git

  git clone https://github.com/ffreemt/deepl-fastapi.git

  and cd deepl-fastapi
• `pip install -r requirements.txt
  • or poetry install

Usage

• Start the server

Use uvicorn directly (note the deepl_server module, not run_uvicorn)

uvicorn deepl_fastapi.deepl_server:app

or

deepl-fastapi
# this option is available only if installed via pip install or poetry add

or

python3.7 -m deepl_fastapi.run_uvicorn

or run the server on the external net, for example at port 9888

uvicorn deepl_fastapi.deepl_server:app --reload --host 0.0.0.0 --port 9888

• Explore and consume

Point your browser to http://127.0.0.1:8000/text/?q=test&to_lang=zh

Or in python code (pip install requests first)

import requests

# get
url =  "http://127.0.0.1:8000/text/?q=test me&to_lang=zh"
print(requests.get(url).json())
# {'q': 'test me', 'from_lang': None, 'to_lang': 'zh',
# 'trtext': '考我 试探我 测试我 试探'}

# post
text = "test this and that"
data = {"text": text, "to_lang": "zh"}
resp = requests.post("http://127.0.0.1:8000/text", json=data)
print(resp.json())
# {'q': {'text': 'test this and that', 'from_lang': None, 'to_lang': 'zh', 'description': None},
# 'result': '试探 左右逢源 检验 审时度势'}

Interactive Docs (Swagger UI)

http://127.0.0.1:8000/docs

OmegaT Fake MT plugin setup

• Copy omegat-plugin-fake-mt-1.0.0.jar （available at https://github.com/briacp/omegat-plugin-fake-mt） to OmegaT\plugins （e.g., C:\Program Files\OmegaT\plugins）

RunOmegaTand setup omegat-plugin-fake-mt

OmegaT/Preferences/Machine Translation/Fake MT/Configure
	Name: Fake Deepl MT
	URL: http://localhost:8000/text
	Source Parameter: from_lang
	Target Parameter: to_lang
	Text Parameter: q