fix: display polarhistograms with more than 89 bins properly

Change petal resolution to the larger one of (a) 360 points (359
segments) per circle or (b) n_histogram_bins+1
points (n_histogram_bins segments). Ensuring at least one segment per
bin petal.

Previously polarhistograms had a fixed resolution of 90 points (89
segments) per circle. Specifying more than 89 bins would result in
either an empty plot (due to a bug in linspace) or too narrow
petals (with this linspace bug fixed).

Author: j0hnnybash

source/matplot/axes_objects/histogram.cpp

@@ -102,9 +102,10 @@ namespace matplot {
 
         } else {
             // resolutions of the petals
-            constexpr double points_per_circle = 90;
+            constexpr double points_per_circle = 360;
             const double n_histogram_bins = bin_edges_.size() - 1.;
-            const double points_per_bin = points_per_circle / n_histogram_bins;
+            // ensure at least one line segment (2 points) per petal.
+            const size_t points_per_bin = static_cast<size_t>(max(2.0, points_per_circle / n_histogram_bins));
             if (!stairs_only_) {
                 for (size_t i = 0; i < values_.size(); ++i) {
                     // make a petal for each value
@@ -114,8 +115,7 @@ namespace matplot {
                     // theta = edge_end             rho = 0
                     ss << "    " << bin_edges_[i] << "  " << 0 << "\n";
                     auto arc_between_edges =
-                        linspace(bin_edges_[i], bin_edges_[i + 1],
-                                 static_cast<size_t>(ceil(points_per_bin)));
+                        linspace(bin_edges_[i], bin_edges_[i + 1], points_per_bin);
                     for (size_t j = 0; j < arc_between_edges.size(); ++j) {
                         ss << "    " << arc_between_edges[j] << "  "
                            << values_[i] << "\n";
@@ -129,8 +129,7 @@ namespace matplot {
                     // theta = edge_end             rho = bin_value
                     // theta = edge_end             rho = next bin value
                     auto arc_between_edges =
-                        linspace(bin_edges_[i], bin_edges_[i + 1],
-                                 static_cast<size_t>(ceil(points_per_bin)));
+                        linspace(bin_edges_[i], bin_edges_[i + 1], points_per_bin);
                     for (size_t j = 0; j < arc_between_edges.size(); ++j) {
                         ss << "    " << arc_between_edges[j] << "  "
                            << values_[i] << "\n";
@@ -840,4 +839,4 @@ namespace matplot {
         return this->num_bins();
     }
 
-} // namespace matplot
+} // namespace matplot