Taking Care of Parameters So You Don't Have to with ParamTools
summary
A library called ParamTools processes and validates model inputs, is built on marshmallow, and exposes parameter values as numpy arrays for easy integration. It adds multi‑year extend behavior and inflation indexing, correctly updates yeared defaults, and handles arrayized runs. Integration with Tax-Calculator reaches full merge after extensive tests (67 PR comments) with all tests passing and an edge‑case indexing bug found and fixed, yielding more readable parameter code and working interoperability with PSL and Compute Studio.
AI-written summary · about
chapters (17)
- 0:00ParamTools introduction
- 0:26What ParamTools does
- 1:28Projects using ParamTools
- 2:31Why ParamTools was created
- 3:36ParamTools features
- 3:58Demo: Tax calculator
- 6:02Handling tax rate updates
- 8:29Multiple years with NumPy
- 10:37Tax Calculator model
- 13:25Integration challenges
- 16:57Extending values across years
- 19:47Indexing to inflation
- 22:12Testing ParamTools
- 25:00Merging ParamTools
- 26:24Lessons learned
- 27:29Other ParamTools uses
- 28:11ParamTools roadmap
related talks
44mCarousel A Python Model Simulation Framework
25mpypet: A Python Toolkit for Simulations and Numerical Experiments
35mPandas' Thumb: unexpected evolutionary use of a Python library.
30mPyEarthTools: Machine learning for Earth system science
55mDerivatives Analytics with Python & Numpy
55mDerivatives Analytics with Python & Numpy
23mSolving Two Hard Problems in Computer Science Using Pandas
42mBayesian Non-parametric Models for Data Science using PyMC3