Sensenova Python Library

The Sensenova Python library provides convenient access to the Sensenova API from applications written in the Python language.

Installation

You don't need this source code unless you want to modify the package. If you just want to use the package, just run:

pip install --upgrade sensenova

Usage

Configure API KEY

Set environment variable:

export SENSENOVA_ACCESS_KEY_ID=
export SENSENOVA_SECRET_ACCESS_KEY=

Or set sensenova.access_key_id and sensenova.secret_access_key to its value:

import sensenova

sensenova.access_key_id = "..."
sensenova.secret_access_key = "..."

API

cli

sensenova api chat_completions.create -m xxx -g user "Say this is a test!" --stream

python

import sensenova

# create a chat completion
chat_completion = sensenova.ChatCompletion.create(
    model="xxx",
    messages=[{"role": "user", "content": "Say this is a test!"}]
)

# print the chat completion
print(chat_completion.data.choices[0].message)

python async

Async support is available in the API by prepending a to a network-bound method:

import sensenova
import asyncio


async def create_chat_completion():
    chat_completion_resp = await sensenova.ChatCompletion.acreate(
        model="xxx",
        messages=[{"role": "user", "content": "Say this is a test!"}]
    )
    print(chat_completion_resp.data.choices[0].message)


asyncio.run(create_chat_completion())

To make async requests more efficient, you can pass in your own aiohttp.ClientSession, but you must manually close the client session at the end of your program/event loop:

import sensenova
from aiohttp import ClientSession

sensenova.aiosession.set(ClientSession())
# At the end of your program, close the http session
await sensenova.aiosession.get().close()

Tools

Run the following command on data.json to get the cleaned_data.json file

sensenova tools fine_tunes.prepare_data -f data.json

Requirements

• Python 3.7.1+