fix segfault for NaN values

.github/workflows/_build.yaml

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Template workflow to build and install BiasAdjustCXX
 #

.github/workflows/_build_doc.yaml

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Template workflow to build documentation.
 #

.github/workflows/_build_docker.yaml

@@ -1,7 +1,7 @@
 ##    Checks the code logic, style and more
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Workflow to build the docker image.
 

.github/workflows/_pre_commit.yaml

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Template workflow to run pre-commit.
 #

.github/workflows/_test.yaml

@@ -1,7 +1,7 @@
 ##    Checks the code logic, style and more
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Workflow to build the test suite and run the unit tests.
 

.github/workflows/cicd.yaml

@@ -1,7 +1,7 @@
 ##    Checks the code logic, style and more
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Workflow to apply pre-commit, build, test and upload the docker
 # image(s).

.github/workflows/release.yaml

@@ -1,7 +1,7 @@
 ##    Checks the code logic, style and more
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Workflow to apply pre-commit, build, test and upload the docker
 # image. This job is only triggered during a release.

.github/workflows/release.yml

@@ -1,7 +1,7 @@
 ##    Checks the code logic, style and more
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 # Workflow to apply pre-commit, build, test and upload the docker
 # image. This job is only triggered during a release.

Dockerfile

@@ -1,6 +1,6 @@
 # -*- coding:utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 #
 
 FROM alpine:3.17 AS builder

Makefile

@@ -1,7 +1,7 @@
 #!make
 # -*- coding: utf-8 -*-
 # Copyright (C) 2023 Benjamin Thomas Schwertfeger
-# Github: https://github.com/btschwertfeger
+# GitHub: https://github.com/btschwertfeger
 
 PROJECT := BiasAdjustCXX
 TEST_PROJECT := TestBiasAdjustCXX
@@ -46,12 +46,12 @@ redoc: clean doc
 
 ## ======= U N -/ I N S T A L L A T I O N =======
 ## install	Installation of the BiasAdjustCXX tool
-##		(after a successfull the build)
+##		(after a successful the build)
 ##
 install:
 	cd build && make install
 
-## uninstall	Uninstallation of the BiasAdjustCXX tool
+## uninstall	Remove of the BiasAdjustCXX tool
 ##
 uninstall:
 	cd build && make uninstall
@@ -73,6 +73,8 @@ uninstall-val:
 test:
 	cd build/tests && ctest
 
+retest: dev test
+
 ## ======= M I S C E L A N I O U S =======
 ## changelog	Create the changelog
 ##

src/CMethods.cxx

@@ -38,7 +38,9 @@
 #include "CMethods.hxx"
 
 #include <algorithm>
+#include <cmath>
 #include <iostream>
+#include <limits>
 #include <vector>
 
 #include "MathUtils.hxx"
@@ -424,6 +426,48 @@ void CMethods::Delta_Method(
     }
 }
 
+/**
+ * Helper function used in `get_xbins`
+ *
+ * It takes two float values and returns a bool indicating whether the first
+ * value is considered smaller than the second.
+ * example:
+ *     a_max = *std::max_element(a.begin(), a.end(), is_smaller),
+ *
+ * @param a some value
+ * @param b what a wonder... another value to compare to
+ *
+ * … will return the maximum value without failing if any value is NaN
+ */
+bool is_smaller(double a, double b) {
+    if (std::isnan(a))
+        return false;  // a is NaN, consider it as larger
+    if (std::isnan(b))
+        return true;  // b is NaN, consider a as smaller
+    return a < b;     // neither a nor b is NaN, compare normally
+}
+
+/**
+ * Helper function used in `get_xbins`
+ *
+ * It takes two float values and returns a bool indicating whether the first
+ * value is considered larger than the second.
+ * example:
+ *     a_max = *std::min_element(a.begin(), a.end(), is_larger),
+ *
+ * @param a some value
+ * @param b what a wonder... another value to compare to
+ *
+ * … will return the minimum value without failing if any value is NaN.
+ */
+bool is_larger(double a, double b) {
+    if (std::isnan(a))
+        return true;
+    if (std::isnan(b))
+        return false;
+    return a < b;
+}
+
 /**
  * Returns the probability boundaries based on two input time series
  * -> This is required to compute the bin boundaries for the Probability Density and Cumulative Distribution Function.
@@ -433,7 +477,7 @@ void CMethods::Delta_Method(
  * @param a time series
  * @param b another time series
  * @param n_quantiles number of quantiles to use/respect
- * @param kind regular or bounded: defines if mininum value of a (climate) variable can be lower than in the data
+ * @param kind regular or bounded: defines if minimum value of a (climate) variable can be lower than in the data
  *             or is set to 0 (ratio based variables => 0)
  * @return the probability boundaries mentioned above
  */
@@ -445,10 +489,10 @@ std::vector<double> CMethods::get_xbins(
 ) {
     if (kind == "regular") {
         const double
-            a_max = *std::max_element(std::begin(a), std::end(a)),
-            a_min = *std::min_element(std::begin(a), std::end(a)),
-            b_max = *std::max_element(std::begin(b), std::end(b)),
-            b_min = *std::min_element(std::begin(b), std::end(b));
+            a_max = *std::max_element(a.begin(), a.end(), is_smaller),
+            a_min = *std::min_element(a.begin(), a.end(), is_larger),
+            b_max = *std::max_element(b.begin(), b.end(), is_smaller),
+            b_min = *std::min_element(b.begin(), b.end(), is_larger);
 
         const double
             global_max = std::max(a_max, b_max),
@@ -463,8 +507,8 @@ std::vector<double> CMethods::get_xbins(
 
     } else if (kind == "bounded") {
         const double
-            a_max = *std::max_element(std::begin(a), std::end(a)),
-            b_max = *std::max_element(std::begin(b), std::end(b));
+            a_max = *std::max_element(std::begin(a), std::end(a), is_smaller),
+            b_max = *std::max_element(std::begin(b), std::end(b), is_smaller);
         const double global_max = std::max(a_max, b_max);
         const double wide = global_max / n_quantiles;