Interpolate 1D and 2D data using SciPy and Interpax.
npx skills add https://github.com/yonesuke/skills --skill interpolation
Please help me install this Agent Skill. Skill: interpolation Source: https://github.com/yonesuke/skills/tree/main/interpolation Command: npx skills add https://github.com/yonesuke/skills --skill interpolation
This skill provides reliable 1D and 2D data interpolation using SciPy on CPU and differentiable interpolation via JAX/Interpax on GPU, enabling smooth value estimation between samples.
Run the SciPy/JAX interpolation demos to observe spline, monotonic, and differentiable interpolation on sample data.
High-intent search queries and answers about installing and using this skill.
2D grid interpolation uses SciPy RegularGridInterpolator for structured grids and griddata for scattered points, estimating intermediate values across two-dimensional sample distributions. This enables accurate value estimation for spatial data and simulation workflows.
Interpax integrates with JAX on GPU to provide differentiable interpolation, enabling gradient-based optimization workflows where interpolation parameters must be tuned via automatic differentiation. This bridges numerical interpolation with machine learning training loops.
Scattered data interpolation is handled through SciPy griddata, which maps irregularly distributed sample points onto regular grids. RegularGridInterpolator covers structured 2D grids, providing accurate estimation for both scattered and gridded spatial distributions.
The skill supports dual execution paths: SciPy-based interpolation runs on CPU for standard numerical analysis, while Interpax with JAX enables differentiable interpolation on GPU. This covers both traditional computation and ML-accelerated workflows.
PCHIP interpolation preserves monotonicity in data where cubic spline would introduce oscillations or overshoots, making it preferable for data that must remain monotonic between samples. Cubic spline suits general smooth interpolation when monotonicity is not a constraint.
SoljournerProcess engineering data with NumPy arrays for velocity fields, pump curves, and linear systems.
DNYoussefDiagnose ML training issues across loss divergence, NaN losses, overfitting, and slow convergence.
jkitchinProvide expert guidance for JAX numerical computing with grad, jit, vmap, and pmap.
xjtulycPrice bonds and compute duration, convexity, and DV01 with Python.
