pysdmx in a nutshell

What is pysdmx?

pysdmx is a pragmatic and opinionated SDMX library written in Python. It focuses on simplicity, providing a subset of SDMX functionalities without requiring advanced knowledge of SDMX. pysdmx is developed as part of the sdmx.io project under the BIS Open Tech initiative.

What does it do?

pysdmx aspires to be a versatile SDMX toolbox for Python, covering various use cases. Here are some highlights:

SDMX information model in Python

pysdmx offers Python classes representing a simplified subset of the SDMX information model, enabling a domain-driven development of SDMX processes in Python. The model classes support serialization in formats like JSON, YAML, or MessagePack. This functionality relies on the msgspec library.

Metadata in action

SDMX metadata are very useful for documenting statistical processes. For example, they can define the structure we expect for a data collection process and share it with the organizations providing data so that they know what to send.

However, metadata can do so much more than that, i.e., they can be “active” and drive various types of statistical processes, such as generating the filesystem layout, creating the physical data model, validating data, mapping data, and configuring processes. To drive such processes, pysdmx supports retrieving metadata from an SDMX Registry or any service compliant with the SDMX-REST 2.0.0 (or above) API. Use these metadata to power your own statistical processes!

Reading and writing SDMX files

pysdmx supports reading and writing SDMX data and structure messages, in various formats, such as SDMX-CSV, SDMX-JSON, and SDMX-ML.

Data discovery and data retrieval

This functionality is under development. Once ready, pysdmx will allow:

• Listing public SDMX services.
• Discovering data available in these services.
• Retrieving data from these services.

This functionality is based on the SDMX Global Discovery Service initiative.

Integration with the ecosystem

pysdmx integrates nicely with other standards, like the Validation and Transformation Language (VTL), and major Python libraries like Pandas. Take a look at the pysdmx toolkit module to learn more.

pysdmx is available on PyPI and can be installed using options such as pip, pipx, poetry, etc.

For more details, check the project documentation.