Reduce spam during tests and remove docstrings on some internal methods extending Base functions (#231)

Author: DanielVandH

.github/workflows/CI.yml

@@ -18,6 +18,7 @@ jobs:
       fail-fast: false
       matrix:
         version:
+          - '1'
           - '1.11'
           - 'pre'
           - 'lts'

NEWS.md

@@ -3,6 +3,8 @@
 ## 1.6.5
 
 - Clarified the counter-clockwise requirement for the `clip_polygon` argument in `voronoi`. See [#230](https://github.com/JuliaGeometry/DelaunayTriangulation.jl/pull/230).
+- Reduced spam during package tests. `@info` statements are now only shown on a failed test, and are now shown in the `Context` output of the failed test. See [#231](https://github.com/JuliaGeometry/DelaunayTriangulation.jl/pull/231).
+- Removed docstrings on some methods that extended `Base` functions for internal use only. You can still find documentation on them by going to the method location and viewing the commented-out docstring. See [#231](https://github.com/JuliaGeometry/DelaunayTriangulation.jl/pull/231). 
 
 ## 1.6.4 
 

docs/Project.toml

@@ -20,7 +20,7 @@ Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [compat]
 BenchmarkTools = "1.5"
-CairoMakie = "0.12"
+CairoMakie = "0.15"
 Dates = "1.11"
 Documenter = "1.5.0"
 DocumenterTools = "0.1"
@@ -34,3 +34,6 @@ SimpleGraphs = "0.8"
 StableRNGs = "1.0"
 StatsBase = "0.34"
 Test = "1.11"
+
+[sources]
+DelaunayTriangulation = { path = ".." }

docs/src/applications/cell_simulations.md

@@ -2,7 +2,19 @@
 EditURL = "https://github.com/JuliaGeometry/DelaunayTriangulation.jl/tree/main/docs/src/literate_applications/cell_simulations.jl"
 ```
 
-This example has been moved to Agent.jl's documentation. Please see the example [here]().
+# Cellular Biology
+This example has been moved to Agent.jl's documentation. Please
+see the example [here](https://juliadynamics.github.io/Agents.jl/stable/examples/delaunay/).
+
+## Just the code
+An uncommented version of this example is given below.
+You can view the source code for this file [here](https://github.com/JuliaGeometry/DelaunayTriangulation.jl/tree/main/docs/src/literate_applications/cell_simulations.jl).
+
+```julia
+
+```
+
+---
 
 *This page was generated using [Literate.jl](https://github.com/fredrikekre/Literate.jl).*
 

docs/src/math/clipped.md

@@ -25,7 +25,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example vornclip11
-fig # hide
+fig #hide
 ```
 
 At the end of this section, we also discuss the intersection of $\mathcal V(\mathcal P)$ with a rectangle, and then discuss clipping $\mathcal V(\mathcal P)$ more generically to a given convex polygon.
@@ -100,7 +100,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example vornclip11
-fig # hide
+fig #hide
 ```
 
 In the first figure above, we are considering the processing of an edge $e$. The Voronoi polygon $\mathcal V$ we consider incident to $e$ is shown, obtained by finding that the midpoint of $e$ is contained in $\mathcal V$. By just processing the intersections of $\mathcal V$, we find the intersections with $\mathcal C\mathcal H(\mathcal P)$ (shown in magenta) shown in orange. This alone is not enough, though, as we can see that we don't identify that the black dot between $e_\ell$ and $e$ should be included in this intersection. This dot is an example of a corner point we were discussing previously, showing the need for this extra processing step. Note also from the above figure that  unbounded polygons $\mathcal V$ alone are not sufficient for checking all intersections, as we can see that some of the bounded polygons also intersect with the convex hull. Eventually, after processing all edges in this way, we obtain the set of orange points shown in the second figure.
@@ -157,7 +157,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example suthodg
-fig # hide
+fig #hide
 ```
 
 In the figure above, the black polygon shows the subject polygon and the red polygon is the clip polygon. Our aim is to clip the subject polygon to the clip polygon to obtain the blue polygon shown. Letting $\mathcal P_S = \{p_1, \ldots, p_n\}$ and $\mathcal P_C = \{q_1, \ldots, q_m\}$ be the subject and clip polygons listed in counter-clockwise order with $p_1 \neq p_n$ and $q_1 \neq q_m$, respectively, the procedure for this clipping as follows:
@@ -208,7 +208,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example suthodg
-fig # hide
+fig #hide
 ```
 
 In the figure above, we show the individual steps of this algorithm. In the second panel, the blue line shows the extended edge of the clip polygon that we use to slice the subject polygon, clipping it onto the edge. For the next four panels, the blue line never touches the subject polygon, and so nothing happens. The last two panels show the last two clips needed to obtain the final polygon.
@@ -270,7 +270,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example suthodg
-fig # hide
+fig #hide
 ```
 
 In the figure above, the polygon we are interested is the one corresponding to the black dot, and we want to clip this polygon to the red rectangle. The following figures show how we grow the polygon's unbounded edges to begin.
@@ -324,7 +324,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example suthodg
-fig # hide
+fig #hide
 ```
 
 In these figures, the blue polygon shows the Voronoi polygon, and the green edges show the approximations to the unbounded rays; the top-most ray starts away from the polygon since the midpoint is further behind the associated circumcenter in this case. The magenta line shows the line segment joining the two approximations - the aim is to grow the green edges long enough such that the magenta line is completely outside of the red rectangle. Let's analyse each figure.
@@ -350,7 +350,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example suthodg
-fig # hide
+fig #hide
 ```
 
 ## Computing the intersection of the Voronoi tessellation with a convex polygon

docs/src/math/constrained.md

@@ -51,7 +51,7 @@ end
 ```
 
 ```@example segins1 
-fig # hide
+fig #hide
 ```
 
 To develop an algorithm, we need to notice one important thing from this figure: Since the blue segment will occlude visibility between any points on either side of the segment, the blue segment effectively divides, locally, the triangulation into two parts that no longer interact with each other. In the figure above, this means that any changes to the triangles bounded between the red curve and the blue segment will not interact those in the region bounded between the green curve and the bkue segment. This is a key observation that will allow us to develop an algorithm for inserting segments: we can handle the two sides separately.
@@ -151,7 +151,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example segins1
-fig # hide
+fig #hide
 ```
 
 ## Adding Points into a Constrained Delaunay Triangulation

docs/src/math/convex.md

@@ -41,7 +41,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example convpoly
-fig # hide
+fig #hide
 ```
 
 See that at each stage the vertex $v_u$ to be added, shown in red, it lies outside of the triangulation, and only a single edge $e_{vw}$ separates $v_u$ from the triangulation. Thus, we can identify that the point location step amounts to finding this edge $e_{vw}$ so that we inserting $v_u$ into the triangulation can be done by retriangulating the cavity formed by the union of the triangles $T_{uvw}$ and the triangles containing $u$ in their circumcircles.

docs/src/math/curve_bounded.md

@@ -65,7 +65,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvbonddom1
-fig # hide
+fig #hide
 ```
 
 In the figure above, we show an example of a set of subcurves $\{s_1, s_2, s_3, s_4, s_5, s_6\}$ together with a common apex vertex. We see that the curve $s_1$ forms no small angle with any other curve, and so it is not a part of any small angle complex. The curves $\{s_2, s_3, s_4\}$ together form a small angle complex since they form a contiguous set of small angles. The curve $s_5$ is not included in this complex since the angle between $s_4$ and $s_5$ is not small. Lastly, the curves $\{s_5, s_6\}$ define a small angle complex. See that it is possible for a single apex vertex to define several separate small angle complexes.
@@ -130,7 +130,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvbonddom1
-fig # hide
+fig #hide
 ```
 
 In the above figure, the subcurves of interest are the subcures defined by the curves from $a$ to each of the first blue points on each curve. These blue points and lines shows the piecewise linear approximation to each subcurve. The shortest edge length is then used to compute $r_{\max}$, giving the circle shown in red. The shell with radius $2^{\lfloor \log_2 r_{\max}}\rfloor$ is shown in green. We then split each of the subcurves at their intersection with this green curve, giving the points shown in green and the updated piecewise linear approximation in the second figure. All subsequent splits of these subcurves (those between $a$ and the new split points) will be at the next smaller power of two shell, i.e. $2^{\lfloor \log_2 r_{\max} \rfloor - 1}$.

docs/src/math/curves.md

@@ -100,7 +100,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvexs1 
-fig # hide
+fig #hide
 ```
 
 There are many orientation markers on this curve. Only two of these come from $\kappa(t) = 0$, with all the others coming from $x', y', x''$, and $y''$, as we can see from the graphs shown.
@@ -140,7 +140,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvexs1
-fig # hide
+fig #hide
 ```
 
 ### Position of a point relative to a curve 
@@ -173,7 +173,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvexs1
-fig # hide
+fig #hide
 ```
 
 ### Computing equidistant splits
@@ -217,7 +217,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example curvexs1
-fig # hide
+fig #hide
 ```
 
 ## LineSegment

docs/src/math/delaunay.md

@@ -58,7 +58,7 @@ triplot!(ax, tri, strokewidth = 3, show_points = true)
 ```
 
 ```@example delaunay_ex1
-fig # hide
+fig #hide
 ```
 
 We can see that the circumcircles of the triangles contain no points in their interior. This is a Delaunay triangulation.
@@ -148,7 +148,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example delaunay_ex1
-fig # hide
+fig #hide
 ```
 
 In the above figure, we are considering insertion the red point into the triangulation. We then then all triangles whose circumcircles contain this red point, shown in blue. Deleting these triangles leaves behind a cavity shown in the third figure. Finally, connecting the vertices of the cavity to the red point gives the final triangulation.
@@ -186,7 +186,7 @@ scatter!(ax, [p, q], color = :red, markersize = 17, strokecolor = :black, stroke
 ```
 
 ```@example delaunay_ex1 
-fig # hide
+fig #hide
 ```
 
 The algorithm draws a line connecting some initial point $q$ and the search point $p$, as shown in red, and then marches along triangles until $p$ is found, traversing the blue triangles shown above. 
@@ -254,7 +254,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example delaunay_ex1
-fig # hide
+fig #hide
 ```
 
 In this example, we are interested in the ghost triangle associated with the blue edge. The oriented outer halfplane for this edge can be defined simply as $H = \{(x, y) ∈ \mathbb R^2 : y > 1\} ∪ \{(x, y) ∈ \mathbb R^2 : 0 < x < 1, y = 1\}$. Using this definition, we see that the red point is not in $H$, the magenta point is in $H$ (in fact, it is exactly on $\partial H$, where $\partial H$ denotes the boundary of $H$), the black point is in $H$, and the green point is not in $H$.
@@ -301,7 +301,7 @@ resize_to_layout!(fig)
 ```
 
 ```@example delaunay_ex1
-fig # hide
+fig #hide
 ```
 
 The ghost edges are the blue lines, and the central point is the red dot. With this definition, we therefore see that we can uniquely each region in space with a triangle - be it a ghost triangle or a real (solid) triangle.