Libra - Dynamic Schema support in Python

Brady Spears, Los Alamos National Laboratory

Table of Contents

• About Libra
• Features

About Libra

Libra is a database management package built on SQLAlchemy to easily and dynamically connect your relational SQL database to the object-oriented Python development environment. Developed by Brady Spears at Los Alamos National Laboratory (LANL), Libra absorbs much of the boilerplate code and developer overhead in creating and defining SQLAlchemy object-relation mapped (ORM) instances. Libra is maintained and developed under the LANL Seismoacoustic Team's Python Geophysical Suite (PyGS).

Libra is still in development - come back regularly for updates!

Features

Libra extends SQLAlchemy's ORM to support:

• An extension of the sqlalchemy.orm.Session class to allow more connection methods, including instantiation of a database connection via environment variables or config files.
• The loading/writing of database tables and columns from user-defined schema to/from a variety of plain-text, built-in formats, as well as extendable support for any other desired format.
• Dynamic SQL datatype handling, where dialect-specific SQL datatypes can be effectively mapped for the same schema on different database backends.
• A plugin architecture connecting ORM models belonging to your schema to custom-built functionality. Built-in plugins included with the Libra package include:
  • Flatfile read/write support
  • Conversion of ORM instances to Pandas DataFrame objects
  • Column- and table-specific data quality control methods
• Flexibility to derive ORM instance methods from the provided metaclass or a custom metaclass implementation.
• Flexibility to design, define, and digest your schema in a way that makes sense for your needs.