Merge Release 1.0.3

Release Notes:
- Added error message for infinite covariance matrix
- Added unittest for warning message
- Ensure most recent version is displayed ob pypi

README.md

@@ -1,5 +1,5 @@
 ![GitHub License](https://img.shields.io/github/license/davidkowalk/fitting_toolkit)
-![Version](https://img.shields.io/badge/version-1.0.2-green)
+![Version](https://img.shields.io/badge/version-1.0.3-green)
 ![GitHub commit activity](https://img.shields.io/github/commit-activity/m/davidkowalk/fitting_toolkit)
 ![GitHub Repo stars](https://img.shields.io/github/stars/davidkowalk/fitting_toolkit?style=flat&label=github%20stars)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/fitting-toolkit?label=pip%20installs)\

package_description.md

@@ -10,7 +10,7 @@
 
 
 # Fitting Toolkit
-This toolkit aims at providing flexible and powerful tools for data analysis and modelling, but remain easy to use.
+This toolkit aims at providing flexible and powerful tools for preliminary data analysis and modelling, but remain easy to use.
 This toolkit is aimed primarily at my peers, students of physics at the university of bonn, and to a degree at professionals within my field. I am optimizing this toolkit to be used on the scale typical of lab courses and homework assignments but if possible it should be powerful enough to run decently sized datasets on an average laptop.
 
 This package wraps numpy for fast data management and manipulation, scipy for `curve_fit()` and matplotlib for display options.

setup.py

@@ -9,7 +9,7 @@
 
 setup(
     name = "fitting_toolkit",
-    version = "1.0.2",
+    version = "1.0.3",
     package_dir={"": "src"},
     packages=[""],
     long_description=description,

src/fitting_toolkit.py

@@ -2,6 +2,7 @@
 from scipy.special import erf
 import numpy as np
 from matplotlib import pyplot as plt
+import warnings
 
 def generate_thresholds(data, lower_frac=0.15865, upper_frac=0.84135):
     """
@@ -105,6 +106,11 @@ def curve_fit(model, xdata: np.array, ydata: np.array, yerror = None, resamples
         raise ValueError(f"x-data and y-data have different lengths and thus cannot be broadcast together.\nx: {np.shape(xdata)}, y: {np.shape(ydata)}")
 
     params, cov = sc_curve_fit(f = model, xdata = xdata, ydata = ydata, sigma = yerror, **kwargs)
+
+    if np.inf in cov or -np.inf in cov:
+        warnings.warn("Covariance matrix includes infinite values. This usually occours due to a non convergent loss function, i.e. an unfittable model. Confidence interval could not be calculated.", RuntimeWarning)
+        return params, cov, None, None
+
     if not model_axis is None:
          resampled_points = model_axis
     elif model_resolution is None:

tests/test_functions.py

@@ -1,4 +1,5 @@
 import unittest
+from unittest.mock import patch
 import numpy as np
 from src.fitting_toolkit import get_sigma_probability, generate_thresholds, curve_fit
 from src.fitting_toolkit import plot_fit
@@ -58,4 +59,29 @@ def model(x, m, c):
         self.assertLessEqual(diff[1], sigmas[1]*2)
 
         fig, ax = plot_fit(x, y, model, params, lower, upper, yerror=dy)
-        fig.savefig("./tests/plot.png")
+        fig.savefig("./tests/plot.png")
+
+    def test_infinite_covariance_warning(self):
+
+        def mock_curve_fit(*args, **kwargs):
+            """Mock function to simulate infinite covariance matrix in curve_fit."""
+            params = np.array([1.0, 1.0])
+            cov = np.array([[np.inf, 0], [0, np.inf]])
+            return params, cov
+
+        def model(x, a, b):
+            return a * x + b
+
+        xdata = np.array([1, 2, 3, 4, 5])
+        ydata = model(xdata, 2, 1)
+        yerror = np.array([0.1] * len(xdata))
+
+        with patch('src.fitting_toolkit.sc_curve_fit', wraps=mock_curve_fit):
+            with self.assertWarns(RuntimeWarning) as cm:
+                params, cov, lower_conf, upper_conf = curve_fit(
+                    model, xdata, ydata, yerror=yerror
+                )
+
+            self.assertIn("Covariance matrix includes infinite values", str(cm.warning))
+            self.assertIsNone(lower_conf)
+            self.assertIsNone(upper_conf)