aurora-tr

Translate via aurora and variants, openai, azure openai, etc.

Quick Setup

From command line

pip install aurora-tr
docker run -d --name aurora -p 8080:8080 ghcr.io/aurora-develop/aurora:latest

python code

from aurora_tr import aurora_tr

res = aurora_tr("test me", url="http:/127.0.0.1:8080/v1")
print(res)
# output:  {'translation': '测试我', 'notes': ''}

Detailed Instructions

Install

pip install aurora-tr

Use

url for aurora REST API

First off, acquire a url (OPENAI_BASE_URL in python-openai) for the aurora REST API for aurora-tr to use, for example, search the net, beg your friends, or deploy your own aurora, etc.

N.B.: aurora is a service in the wild similar or compatible to openai (more commonly known as chatgpt).

To deplay your own aurora locolly or on a remote machine, refer to https://github.com/aurora-develop/aurora

For example (the easieast way if you can access docker):

docker run -d --name aurora -p 8080:8080 ghcr.io/aurora-develop/aurora:latest

The OPENAI_BASE_URL in this case will be http://localhost:8080/v1 or http://external_ip_or_domainname:8080/v1

python code

Default base_url, api_key and model of aurora_tr take the value of OPENAI_BASE_URL, OPENAI_API_KEY and OPENAI_MODEL from the file .env-aurora.

base_url, api_key and model can also be set directly in aurora_tr, e.g.,

res = aurora_tr(
  '....',
  url_base="https://uu.ci/v1",
  api_key=...,
  model-"gpt-4-turbo-preview",
  selector="uuci"
)

Hence, if you have deployed your own aurora API:

from aurora_tr import aurora_tr

res = aurora_tr("test me", url="http:/127.0.0.1:8080/v1")
print(res)
# output:  {'translation': '测试我', 'notes': ''}

Other examples:

from aurora_tr import aurora_tr

# default aurora, source language: English, target language: Chinese
res = aurora_tr("Test me.")
print(res)

{'translation': '测试我。', 'notes': 'Translated from English to Chinese.'}

# target language: English
res = aurora_tr("试试我", to_lang="English")

{'translation': 'Try me', 'notes': ''}

# select aurora
res = aurora_tr("Testen mich bitte")

 {'translation': '请测试我', 'notes': 'Translated from English to Chinese'}

# select uuci
import dotenv

api_key = dotenv.dotenv_values(".env").get("OPENAI_API_KEY_UUCI")
# or api_key = "......"
res = aurora_tr("Test me.", selector="uuci", api_key=api_key)

{'translation': '测试我。'}

Brief docs

aurora_tr(
    text: str,
    from_lang: Optional[str] = 'English',
    to_lang: str = 'Chinese',
    selector: str = 'aurora',
    base_url: str = '',
    api_key: str = '',
    model: str = '',
    temperature: Optional[float] = None,
)
Docstring:
Translate viaa auroa and uu.ci.

Args
----
text: string to transalte
from_lang: source language
to_lang: target language
selector: prodiver selector
base_url:
api_key: token
model: model name, anything for aurora
temperature (str): 0.2-0.4 for translation, might just left out

Returns
-------
dict/json: {"translation": "...", notes: "..."}