langchain-kinetica
Kinetica intefrace for Langchain. See the LLM documentation for an overview of the Kinetica LLM.
1. Prerequisites
To use langchain with Kinetica you will need:
- Python runtime >3.10
- Kinetica SqlAssist LLM
- Kinetica instance >7.2.0 configured to use SqlAssist.
2. Package Contents
-
KineticaChatLLM
: ChatModel for converting natural language to SQL. -
KineticaSqlOutputParser
: OutputParser that will execute SQL from theKineticaChatLLM
. -
SqlResponse
: If the Kinetica chain ends withKineticaSqlOutputParser
then this response will contain the generated SQL and results from its execution.
3. Installation
You can install this project from pypi.
$ pip install langchain-kinetica
For a pre-release version you can install directly from the repository.
$ pip install "langchain-kinetica @ git+ssh://git@github.com/kineticadb/langchain-kinetica.git"
4. Usage
See the Kinetica LLM Demo notebook for examples.
5. Building
Install the project locally.
$ pip install --editable .
You will need to install the build utility.
$ pip install --upgrade build
Build the project
$ python3 -m build
The build will generate a .whl
file that can be distributed.
$ ls -1 ./dist
langchain-kinetica-1.0.tar.gz
langchain_kinetica-1.0-py3-none-any.whl