I'm a senior data engineer, contributing in open-source also building light weight python semantic layer for DE.

github

Nitesh Jangir  • Peerlist

PySemantic is a lightweight, graph-based semantic layer designed for Python and SQL, offering a robust solution for data management and query generation. This tool allows users to define metrics once and generate SQL queries across various platforms, ensuring a single source of truth for business logic. Key features include automatic join resolution, SQL injection safety, and intelligent filter pushdown, which enhances query efficiency and security. PySemantic supports multiple SQL dialects, including MySQL, Postgres, BigQuery, Snowflake, and DuckDB, thanks to its integration with SQLGlot. The tool is a pure Python library, requiring no additional infrastructure such as servers or Docker, making it easy to install and use. Additionally, PySemantic offers an interactive studio powered by Streamlit, allowing users to explore models, visualize entity graphs, and test queries directly in the browser. The architecture of PySemantic processes queries through a structured pipeline, ensuring efficient query generation and execution. This makes PySemantic an essential tool for developers and data analysts looking to streamline their data operations and enhance productivity.