deeplx and llm translate tool in python (wip)

• install rye (google for instructions for yourplatform)
• git clone https://github.com/ffreemt/deeplx-tr && cd deeplx-tr
• rye pin 3.10 # or pick another python version, e.g., 3.12
• rye sync
• cp example.env .env and amend .env
• rye run taipy llm_tool.py

pip install deeplx_tr --pre

deeplx-tr hello world
# 哈罗世界

deeplx-tr hello world -t de
# Hallo Welt

deeplx-tr --help

or

python -m deeplx_tr hello world
python -m deeplx_tr hello world -d
python -m deeplx_tr --help

from deeplx_tr import deeplx_tr

res = deeplx_tr("hello world")
print(res)
# 哈罗世界

res = deeplx_tr("hello world", to_lang="de")
print(res)
# Hallo Welt

N.B. deeplx-tr will likely spit out too many requestes if you call it too often before long. But it's sufficient for ordinary average daily translation. If you have a higher demand, try deeplx.org for which we provided two clients for your convenience.

from deeplx_tr import deeplx_client, deeplx_client_async

res = deeplx_client("hello world")
print(res)
# '哈罗世界'

res = deeplx_client("hello world", target_lang="de")
print(res)
# 'Hallo Welt'

# if you host your own deeplx, for example, at `127.0.0.1:1188'
# res = deeplx_client("hello world", url="https://127.0.0.1:1188/translate")

An async client is also available, e.g.

import asyncio
from deeplx_tr import deeplx_client_async

async def main():
  res = await asyncio.gather(deeplx_client_async("hello world"), deeplx_client_async("test"))
  print(res)

asyncio.run(main())
# ['哈罗世界', '测试']

The default concurrency limit is 5 but can be altered by setting the environ variable CONCURRENCY_LIMIT, e.g.

set CONCURRENCY_LIMIT=8  # in Windows

# export CONCURRENCY_LIMIT=8 in Linux