🚀 PrecisionThermometryFramework

Welcome to the first public release of PrecisionThermometryFramework! This release introduces a comprehensive, high-performance metrological suite designed for the rigorous calibration and characterization of cryogenic temperature sensors (e.g., RhFe, PtCo, SPRTs, and Diodes).

🌟 Key Highlights

This framework bridges the gap between raw laboratory measurements and publication-ready calibration models, offering automated topological optimization and advanced physical continuity enforcement.

🧠 Core Analytical Engines

• Rational Function Optimization (Padé Approximants): Automated 2D grid-scanning to identify the optimal numerator (n) and denominator (m) degrees.
• Information Criteria Selection: Built-in Akaike (AIC) and Bayesian (BIC) Information Criteria algorithms to objectively prevent model overfitting.
• Intelligent Piecewise Topology: Sub-divide complex thermal datasets into local segments while mathematically enforcing C⁰ boundary continuity (knot stitching) to prevent physical discontinuities.
• ITS-90 SPRT Calibration: Complete implementation of the International Temperature Scale of 1990, supporting both exact fixed-point analytical solutions and iterative self-consistent integrations.

📊 Advanced Diagnostics & Cross-Validation

• Robust Outlier Rejection: Interactive pruning loops utilizing Z-scores, Interquartile Range (IQR), and Studentized Residuals (via Hat Matrix leverage).
• Headless Cross-Validation: Project trained models onto independent datasets (e.g., heating vs. cooling cycles) with automated rolling-average signal smoothing.
• Unified Reporting: Generates highly detailed CSV tables containing unweighted/weighted residuals, parameter uncertainty matrices, and standard errors.

🛠️ Interactive CLI & Workflow

• Sensor-Agnostic Parsing: Automatically maps generic laboratory column headers (e.g., R, Voltage, Tstd, Resistance) for plug-and-play data loading.
• Domain Normalization Modes: Built-in physical scaling transformations (e.g., linear compression, natural logs, W = R/R_TPW ratios).
• Automated Plotting: Generates multi-panel diagnostic dashboards (Studentized residuals, global stitched fits, topology landscapes) using matplotlib.

📂 Included Validation Data

The v1.0.0a release includes a validated examples/ directory featuring a complete calibration pipeline for a 4.4 K – 25 K cryogenic sensor. It serves as a benchmark for reproducing the global fit, piecewise segmentation, and cross-validation workflows.

📥 Installation

git clone [https://github.com/GrzegorzSzklarz/PrecisionThermometryFramework.git](https://github.com/GrzegorzSzklarz/PrecisionThermometryFramework.git)
cd PrecisionThermometryFramework
conda env create -f environment.yml
conda activate calibration_framework

🎓 Citation & DOI

If this framework accelerates your research, please cite the software and the accompanying methodology paper.

Software Citation:

Szklarz, G. (2026). PrecisionThermometryFramework (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.18924874

Methodology Paper (Under Review):

Szklarz, G., Nikonkov, R., and Kowal, A., "Analytical framework for high-precision cryogenic thermometry: Characterization of RhFe and PtCo sensors below 25 K", [Journal Name], (2026).

Developed and maintained by @GrzegorzSzklarz.