wxflows-client 0.1.0

The watsonx.ai flows engine SDK

wxflows_client

wxflows_client is a library to query flows deployed to watsonx.ai flows engine.

from wxflows_client import rag

print(rag.execute(
    endpoint="https://watzen.stepzen.net/watzen-genai/rag_v1/__graphql",
    apikey="xxxx",
    name="A_ragAnswerDesc",
    vars={
      "n": 10,
      "question": "who are the interns in greys anatomy season 9?",
      "aiEngine": "BAM",
      "model": "ibm/granite-13b-chat-v2",
      "embedModel": "ibm/slate-30m-english-rtrvr",
      "collection": "longnq",
      "parameters": {
          "max_new_tokens": 1000,
          "temperature": 0.7
      }
    },
    query="",
    max_depth=2
))

Installing wxflows_client and Supported Versions

wxflows_client is available on PyPI:

$ python -m pip install wxflows_client

wxflows_client officially supports Python 3.8+.

Testing Locally

Note: Be sure to add your endpoint and credentials in the main.py file.

python3 -m venv venv        
source ./venv/bin/activate
pip install -r requirements.txt
make test

Export the package

1. From root, run python -m build.
2. Copy the generated <>.wz files from the generated dist folder and paste it into your project.
3. Execute pip install wxflows_client in your project folder.