mdreader 0.3.0

A Python library for reading and parsing Functional Mock-up Interface model description XML file.

mdreader 📖

A Python library for reading and parsing Functional Mock-up Interface (FMI 2.0 and 3.0) model description XML files into Pydantic models.

Installation 📦

Add mdreader to your pyproject.toml with uv with:

uv add mdreader

To install uv, see https://docs.astral.sh/uv/getting-started/installation/

How to use 🚀

To read and parse an FMI model description (works for both 2.0 and 3.0):

import mdreader.fmi2 as fmi2
import mdreader.fmi3 as fmi3

# Read from XML file
md = fmi2.read_model_description("path/to/fmi2/modelDescription.xml")  # or fmi3.read_model_description for 3.0
# Read from FMU archive
md = fmi2.read_model_description("path/to/model.fmu")  # or fmi3.read_model_description for 3.0

# Read from unzipped FMU directory
md = fmi2.read_model_description("path/to/unzipped/fmu/directory")  # or fmi3.read_model_description for 3.0

print(md)

Features ✨

• Parse FMI 2.0 and 3.0 model description XML files
• Read model information from FMU archives
• Access model metadata (name, version, author, GUID, etc.)
• Extract variable definitions (real, integer, boolean, string, enumeration, clock, binary)
• Access unit definitions and type definitions
• Support for Model Exchange, Co-Simulation, and Scheduled Execution interfaces
• Parse model structure, dependencies, and experiment configurations
• Full support for FMI 3.0 features including structural parameters, clock variables, and directional derivatives

Why another FMI model description reader? 🤔

• Lightweight: mdreader only depends on Pydantic
• De/Serialization: Pydantic models support easy serialization to/from JSON, dict, etc.
• Validation: mdreader uses Pydantic models to ensure the integrity of the parsed data
• FMI version specific: The fmi2.FMIModelDescription and fmi3.FMIModelDescription classes are specific to their respective FMI versions (not a mix of versions), making it simpler to use for each version

Related projects 🔗

• fmpy: A similar read_model_description function is available in FMPy, but it uses custom classes instead of Pydantic models and has more dependencies.

Licensing 📄

The code in this project is licensed under MIT license. See the LICENSE file for details.