perfect-shape 0.3.5 → 0.5.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ab1f48d0551ed07a6a2b5f2d51abba6c6596eef82e156e68df57489319104e4c
-  data.tar.gz: bb67e137bbf23df2ed291a574ef45ae97113dd140d0a17269c9b60c27b7261e9
+  metadata.gz: cf595115a6a9344cedfce9e93801cb5aba1960f4e970d064043e5f813716a0be
+  data.tar.gz: bcb60359b840662739767882546e360ceb052e734f5d3d7836406e25ff7b72ae
 SHA512:
-  metadata.gz: 6d365e47635c7daa094f02849df24696784b667c55bffa479bdbaacc41abf1b9ad1114084a4ed7bd0145e6dd18ad233f9b09e6bad8792828417aca49db2076c4
-  data.tar.gz: d8440bb0d35e1539dcf5c8cad79e2930d74e10e32a408391085027d773c0ea7e4272a6011a7244677c035c8f6a7954606065611282ae8cd14aab71cbbe6b2bba
+  metadata.gz: c57fb7619dd930974af23d8004ad134bc3fefb51c6dd03f5ae2e7a46cccae3011d7dd74ca9a08e8055c41ad93c9a4ef92132f76d7349fd5cc20ffefc5c7d371c
+  data.tar.gz: ca7c2586198a35ae8e9dc1485a87298c11b4151e11aab16e54302775c4539ac94c7867a3b48598914dfb8367b87fa6ae166cdea5dbe9da32ee0161788692b6d2

data/CHANGELOG.md

@@ -1,5 +1,36 @@
 # Change Log
 
+## 0.5.2
+
+- `QuadraticBezierCurve#intersect?(rectangle)`
+
+## 0.5.1
+
+- `Point#intersect?(rectangle)` (equivalent to `Rectangle#contain?(point)`)
+
+## 0.5.0
+
+- `Line#intersect?(rectangle)`
+- `Rectangle#out_state(x_or_point, y = nil)`
+
+## 0.4.0
+
+- `PerfectShape::AffineTransform#new`
+- `PerfectShape::AffineTransform#==`
+- `PerfectShape::AffineTransform#transform_point`
+- `PerfectShape::AffineTransform#transform_points`
+- `PerfectShape::AffineTransform#identity!` (alias: `reset!`)
+- `PerfectShape::AffineTransform#invert!`
+- `PerfectShape::AffineTransform#invertible?`
+- `PerfectShape::AffineTransform#multiply!`
+- `PerfectShape::AffineTransform#translate!`
+- `PerfectShape::AffineTransform#scale!`
+- `PerfectShape::AffineTransform#rotate!`
+- `PerfectShape::AffineTransform#shear!` (alias: `skew!`)
+- `PerfectShape::AffineTransform#clone`
+- `PerfectShape::AffineTransform#inverse_transform_point`
+- `PerfectShape::AffineTransform#inverse_transform_points`
+
 ## 0.3.5
 
 - Check point containment in composite shape outline with distance tolerance (new method signature: `PerfectShape::CompositeShape#contain?(x_or_point, y = nil, outline: false, distance_tolerance: 0)`)

data/README.md

@@ -1,9 +1,9 @@
-# Perfect Shape 0.3.5
+# Perfect Shape 0.5.2
 ## Geometric Algorithms
 [![Gem Version](https://badge.fury.io/rb/perfect-shape.svg)](http://badge.fury.io/rb/perfect-shape)
 [![Test](https://github.com/AndyObtiva/perfect-shape/actions/workflows/ruby.yml/badge.svg)](https://github.com/AndyObtiva/perfect-shape/actions/workflows/ruby.yml)
 
-[`PerfectShape`](https://rubygems.org/gems/perfect-shape) is a collection of pure Ruby geometric algorithms that are mostly useful for GUI (Graphical User Interface) manipulation like checking containment of a mouse click [point](#perfectshapepoint) in popular geometry shapes such as [rectangle](#perfectshaperectangle), [square](#perfectshapesquare), [arc](#perfectshapearc) (open, chord, and pie), [ellipse](#perfectshapeellipse), [circle](#perfectshapecircle), [polygon](#perfectshapepolygon), and [paths](#perfectshapepath) containing [lines](#perfectshapeline), [quadratic bézier curves](#perfectshapequadraticbeziercurve), and [cubic bezier curves](#perfectshapecubicbeziercurve) (including both [Ray Casting Algorithm](https://en.wikipedia.org/wiki/Point_in_polygon#Ray_casting_algorithm), aka [Even-odd Rule](https://en.wikipedia.org/wiki/Even%E2%80%93odd_rule), and [Winding Number Algorithm](https://en.wikipedia.org/wiki/Point_in_polygon#Winding_number_algorithm), aka [Nonzero Rule](https://en.wikipedia.org/wiki/Nonzero-rule)).
+[`PerfectShape`](https://rubygems.org/gems/perfect-shape) is a collection of pure Ruby geometric algorithms that are mostly useful for GUI (Graphical User Interface) manipulation like checking viewport rectangle intersection or containment of a mouse click [point](#perfectshapepoint) in popular geometry shapes such as [rectangle](#perfectshaperectangle), [square](#perfectshapesquare), [arc](#perfectshapearc) (open, chord, and pie), [ellipse](#perfectshapeellipse), [circle](#perfectshapecircle), [polygon](#perfectshapepolygon), and [paths](#perfectshapepath) containing [lines](#perfectshapeline), [quadratic bézier curves](#perfectshapequadraticbeziercurve), and [cubic bezier curves](#perfectshapecubicbeziercurve), potentially with [affine transforms](#perfectshapeaffinetransform) applied like translation, scale, rotation, shear/skew, and inversion (including both [Ray Casting Algorithm](https://en.wikipedia.org/wiki/Point_in_polygon#Ray_casting_algorithm), aka [Even-odd Rule](https://en.wikipedia.org/wiki/Even%E2%80%93odd_rule), and [Winding Number Algorithm](https://en.wikipedia.org/wiki/Point_in_polygon#Winding_number_algorithm), aka [Nonzero Rule](https://en.wikipedia.org/wiki/Nonzero-rule)).
 
 Additionally, [`PerfectShape::Math`](#perfectshapemath) contains some purely mathematical algorithms, like [IEEE 754-1985 Remainder](https://en.wikipedia.org/wiki/IEEE_754-1985).
 
@@ -14,13 +14,13 @@ To ensure high accuracy, this library does all its mathematical operations with
 Run:
 
 ```
-gem install perfect-shape -v 0.3.5
+gem install perfect-shape -v 0.5.2
 ```
 
 Or include in Bundler `Gemfile`:
 
 ```ruby
-gem 'perfect-shape', '~> 0.3.5'
+gem 'perfect-shape', '~> 0.5.2'
 ```
 
 And, run:
@@ -57,7 +57,6 @@ This is a base class for all shapes. It is not meant to be used directly. Subcla
 - `#center_y`: center y
 - `#bounding_box`: bounding box is a rectangle with x = min x, y = min y, and width/height just as those of shape
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
-- `#normalize_point(x_or_point, y = nil)`: normalizes point into an `Array` of `[x,y]` coordinates
 - `#contain?(x_or_point, y=nil, outline: false, distance_tolerance: 0)`: checks if point is inside if `outline` is `false` or if point is on the outline if `outline` is `true`. `distance_tolerance` can be used as a fuzz factor when `outline` is `true`, for example, to help GUI users mouse-click-select a shape from its outline more successfully
 
 ### `PerfectShape::PointLocation`
@@ -86,6 +85,71 @@ Includes `PerfectShape::PointLocation`
 - `#max_x`: max x
 - `#max_y`: max y
 
+### `PerfectShape::AffineTransform`
+
+Class
+
+Affine transforms have the following matrix:
+
+[ xxp xyp xt ]<br>
+[ yxp yyp yt ]
+
+The matrix is used to transform (x,y) point coordinates as follows:
+
+[ xxp xyp xt ] * [x] = [ xxp * x + xyp * y + xt ]<br>
+[ yxp yyp yt ] * [y] = [ yxp * x + yyp * y + yt ]
+
+`xxp` is the x coordinate x product (`m11`)<br>
+`xyp` is the x coordinate y product (`m12`)<br>
+`yxp` is the y coordinate x product (`m21`)<br>
+`yyp` is the y coordinate y product (`m22`)<br>
+`xt` is the x coordinate translation (`m13`)<br>
+`yt` is the y coordinate translation (`m23`)
+
+Affine transform mutation operations ending with `!` can be chained as they all return `self`.
+
+- `::new(xxp_element = nil, xyp_element = nil, yxp_element = nil, yyp_element = nil, xt_element = nil, yt_element = nil,
+         xxp: nil, xyp: nil, yxp: nil, yyp: nil, xt: nil, yt: nil,
+         m11: nil, m12: nil, m21: nil, m22: nil, m13: nil, m23: nil)`:
+  The constructor accepts either the (x,y)-operation related argument/kwarg names or traditional matrix element kwarg names. If no arguments are supplied, it constructs an identity matrix (i.e. like calling `::new(xxp: 1, xyp: 0, yxp: 0, yyp: 1, xt: 0, yt: 0)`).
+- `#matrix_3d`: Returns Ruby `Matrix` object representing affine transform in 3D (used internally for performing multiplication)
+- `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
+- `#identity!` (alias: `reset!`): Resets to identity matrix (i.e. like calling `::new(xxp: 1, xyp: 0, yxp: 0, yyp: 1, xt: 0, yt: 0)`)
+- `#invertible?` Returns `true` if matrix is invertible and `false` otherwise
+- `#invert!`: Inverts affine transform matrix if invertible or raises an error otherwise
+- `#multiply!(other)`: Multiplies affine transform with another affine transform, storing resulting changes in matrix elements
+- `#translate!(x_or_point, y=nil)`: Translates affine transform with (x, y) translation values
+- `#scale!(x_or_point, y=nil)`: Scales affine transform with (x, y) scale values
+- `#rotate!(degrees)`: Rotates by angle degrees counter-clockwise if angle value is positive or clockwise if angle value is negative. Note that it returns very close approximate results for rotations that are 90/180/270 degrees (good enough for inverse-transform GUI point containment checks needed when checking if mouse-click-point is inside a transformed shape).
+- `#shear!(x_or_point, y=nil)`: Shears by x and y factors
+- `#clone`: Returns a new AffineTransform with the same matrix elements
+- `#transform_point(x_or_point, y=nil)`: returns `[xxp * x + xyp * y + xt, yxp * x + yyp * y + yt]`. Note that result is a close approximation, but should be good enough for GUI mouse-click-point containment checks.
+- `#transform_points(*xy_coordinates_or_points)`: returns `Array` of (x,y) pair `Array`s transformed with `#transform_point` method
+- `#inverse_transform_point(x_or_point, y=nil)`: returns inverse transform of a point (x,y) coordinates (clones self and inverts clone, and then transforms point). Note that result is a close approximation, but should be good enough for GUI mouse-click-point containment checks.
+- `#inverse_transform_points(*xy_coordinates_or_points)`: returns inverse transforms of a point `Array` of (x,y) coordinates
+
+Example:
+
+```ruby
+xxp = 2
+xyp = 3
+yxp = 4
+yyp = 5
+xt = 6
+yt = 7
+affine_transform1 = PerfectShape::AffineTransform.new(xxp: xxp, xyp: xyp, yxp: yxp, yyp: yyp, xt: xt, yt: yt) # (x,y)-operation kwarg names
+affine_transform2 = PerfectShape::AffineTransform.new(m11: xxp, m12: xyp, m21: yxp, m22: yyp, m13: xt, m23: yt) # traditional matrix element kwarg names
+affine_transform3 = PerfectShape::AffineTransform.new(xxp, xyp, yxp, yyp, xt, yt) # standard arguments
+
+affine_transform2.matrix_3d == affine_transform1.matrix_3d # => true
+affine_transform3.matrix_3d == affine_transform1.matrix_3d # => true
+
+affine_transform = PerfectShape::AffineTransform.new.translate!(30, 20).scale!(2, 3)
+
+affine_transform.transform_point(10, 10) # => approximately [50, 50]
+affine_transform.inverse_transform_point(50, 50) # => approximately [10, 10]
+```
+
 ### `PerfectShape::Point`
 
 Class
@@ -99,6 +163,7 @@ Includes `PerfectShape::PointLocation`
 Points are simply represented by an `Array` of `[x,y]` coordinates when used within other shapes, but when needing point-specific operations like `point_distance`, the `PerfectShape::Point` class can come in handy.
 
 - `::point_distance(x, y, px, py)`: Returns the distance from a point to another point
+- `::normalize_point(x_or_point, y = nil)`: Normalizes point args whether two-number `point` `Array` or `x`, `y` args, returning normalized point `Array` of two `BigDecimal`'s
 - `::new(x_or_point=nil, y_arg=nil, x: nil, y: nil)`: constructs a point with (x,y) pair (default: 0,0) whether specified as `Array` of (x,y) pair, flat `x,y` args, or `x:, y:` kwargs.
 - `#min_x`: min x (always x)
 - `#min_y`: min y (always y)
@@ -112,6 +177,7 @@ Points are simply represented by an `Array` of `[x,y]` coordinates when used wit
 - `#bounding_box`: bounding box is a rectangle with x = min x, y = min y, and width/height of shape
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
 - `#contain?(x_or_point, y=nil, outline: true, distance_tolerance: 0)`: checks if point matches self, with a distance tolerance (0 by default). Distance tolerance provides a fuzz factor that for example enables GUI users to mouse-click-select a point shape more successfully. `outline` option makes no difference on point
+- `#intersect?(rectangle)`: Returns `true` if intersecting with interior of rectangle or `false` otherwise. This is useful for GUI optimization checks of whether a shape appears in a GUI viewport rectangle and needs redrawing
 - `#point_distance(x_or_point, y=nil)`: Returns the distance from a point to another point
 
 Example:
@@ -139,7 +205,7 @@ Includes `PerfectShape::MultiPoint`
 
 ![line](https://raw.githubusercontent.com/AndyObtiva/perfect-shape/master/images/line.png)
 
-- `::relative_counterclockwise(x1, y1, x2, y2, px, py)`: Returns an indicator of where the specified point (px,py) lies with respect to the line segment from (x1,y1) to (x2,y2). The return value can be either 1, -1, or 0 and indicates in which direction the specified line must pivot around its first end point, (x1,y1), in order to point at the specified point (px,py). A return value of 1 indicates that the line segment must turn in the direction that takes the positive X axis towards the negative Y axis. In the default coordinate system used by Java 2D, this direction is counterclockwise. A return value of -1 indicates that the line segment must turn in the direction that takes the positive X axis towards the positive Y axis. In the default coordinate system, this direction is clockwise. A return value of 0 indicates that the point lies exactly on the line segment. Note that an indicator value of 0 is rare and not useful for determining collinearity because of floating point rounding issues. If the point is colinear with the line segment, but not between the end points, then the value will be -1 if the point lies “beyond (x1,y1)” or 1 if the point lies “beyond (x2,y2)”.
+- `::relative_counterclockwise(x1, y1, x2, y2, px, py)`: Returns an indicator of where the specified point (px,py) lies with respect to the line segment from (x1,y1) to (x2,y2). The return value can be either 1, -1, or 0 and indicates in which direction the specified line must pivot around its first end point, (x1,y1), in order to point at the specified point (px,py). A return value of 1 indicates that the line segment must turn in the direction that takes the positive X axis towards the negative Y axis. In the default coordinate system, this direction is counterclockwise. A return value of -1 indicates that the line segment must turn in the direction that takes the positive X axis towards the positive Y axis. In the default coordinate system, this direction is clockwise. A return value of 0 indicates that the point lies exactly on the line segment. Note that an indicator value of 0 is rare and not useful for determining collinearity because of floating point rounding issues. If the point is colinear with the line segment, but not between the end points, then the value will be -1 if the point lies “beyond (x1,y1)” or 1 if the point lies “beyond (x2,y2)”.
 - `::point_distance_square(x1, y1, x2, y2, px, py)`: Returns the square of distance from a point to a line segment.
 - `::point_distance(x1, y1, x2, y2, px, py)`: Returns the distance from a point to a line segment.
 - `::new(points: [])`: constructs a line with two `points` as `Array` of `Array`s of `[x,y]` pairs or flattened `Array` of alternating x and y coordinates
@@ -155,7 +221,8 @@ Includes `PerfectShape::MultiPoint`
 - `#bounding_box`: bounding box is a rectangle with x = min x, y = min y, and width/height of shape
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
 - `#contain?(x_or_point, y=nil, outline: true, distance_tolerance: 0)`: checks if point lies on line, with a distance tolerance (0 by default). Distance tolerance provides a fuzz factor that for example enables GUI users to mouse-click-select a line shape more successfully. `outline` option makes no difference on line
-- `#relative_counterclockwise(x_or_point, y=nil)`: Returns an indicator of where the specified point (px,py) lies with respect to the line segment from (x1,y1) to (x2,y2). The return value can be either 1, -1, or 0 and indicates in which direction the specified line must pivot around its first end point, (x1,y1), in order to point at the specified point (px,py). A return value of 1 indicates that the line segment must turn in the direction that takes the positive X axis towards the negative Y axis. In the default coordinate system used by Java 2D, this direction is counterclockwise. A return value of -1 indicates that the line segment must turn in the direction that takes the positive X axis towards the positive Y axis. In the default coordinate system, this direction is clockwise. A return value of 0 indicates that the point lies exactly on the line segment. Note that an indicator value of 0 is rare and not useful for determining collinearity because of floating point rounding issues. If the point is colinear with the line segment, but not between the end points, then the value will be -1 if the point lies “beyond (x1,y1)” or 1 if the point lies “beyond (x2,y2)”.
+- `#intersect?(rectangle)`: Returns `true` if intersecting with interior of rectangle or `false` otherwise. This is useful for GUI optimization checks of whether a shape appears in a GUI viewport rectangle and needs redrawing
+- `#relative_counterclockwise(x_or_point, y=nil)`: Returns an indicator of where the specified point (px,py) lies with respect to the line segment from (x1,y1) to (x2,y2). The return value can be either 1, -1, or 0 and indicates in which direction the specified line must pivot around its first end point, (x1,y1), in order to point at the specified point (px,py). A return value of 1 indicates that the line segment must turn in the direction that takes the positive X axis towards the negative Y axis. In the default coordinate system, this direction is counterclockwise. A return value of -1 indicates that the line segment must turn in the direction that takes the positive X axis towards the positive Y axis. In the default coordinate system, this direction is clockwise. A return value of 0 indicates that the point lies exactly on the line segment. Note that an indicator value of 0 is rare and not useful for determining collinearity because of floating point rounding issues. If the point is colinear with the line segment, but not between the end points, then the value will be -1 if the point lies “beyond (x1,y1)” or 1 if the point lies “beyond (x2,y2)”.
 - `#point_distance(x_or_point, y=nil)`: Returns the distance from a point to a line segment.
 
 Example:
@@ -183,6 +250,10 @@ Includes `PerfectShape::MultiPoint`
 
 ![quadratic_bezier_curve](https://raw.githubusercontent.com/AndyObtiva/perfect-shape/master/images/quadratic_bezier_curve.png)
 
+- `::tag(coord, low, high)`: Determine where coord lies with respect to the range from low to high.  It is assumed that low < high.  The return value is one of the 5 values BELOW, LOWEDGE, INSIDE, HIGHEDGE, or ABOVE.
+- `::eqn(val, c1, cp, c2)`: Fill an array with the coefficients of the parametric equation in t, ready for solving against val with solve_quadratic. We currently have: val = Py(t) = C1*(1-t)^2 + 2*CP*t*(1-t) + C2*t^2 = C1 - 2*C1*t + C1*t^2 + 2*CP*t - 2*CP*t^2 + C2*t^2 = C1 + (2*CP - 2*C1)*t + (C1 - 2*CP + C2)*t^2; 0 = (C1 - val) + (2*CP - 2*C1)*t + (C1 - 2*CP + C2)*t^2; 0 = C + Bt + At^2; C = C1 - val; B = 2*CP - 2*C1; A = C1 - 2*CP + C2
+- `::solve_quadratic(eqn)`: Solves the quadratic whose coefficients are in the eqn array and places the non-complex roots into the res array, returning the number of roots. The quadratic solved is represented by the equation: <pre>eqn = {C, B, A}; ax^2 + bx + c = 0</pre> A return value of {@code -1} is used to distinguish a constant equation, which might be always 0 or never 0, from an equation that has no zeroes.
+- `::eval_quadratic(vals, num, include0, include1, inflect, c1, ctrl, c2)`: Evaluate the t values in the first num slots of the vals[] array and place the evaluated values back into the same array.  Only evaluate t values that are within the range <, >, including the 0 and 1 ends of the range iff the include0 or include1 booleans are true.  If an "inflection" equation is handed in, then any points which represent a point of inflection for that quadratic equation are also ignored.
 - `::new(points: [])`: constructs a quadratic bézier curve with three `points` (start point, control point, and end point) as `Array` of `Array`s of `[x,y]` pairs or flattened `Array` of alternating x and y coordinates
 - `#points`: points (start point, control point, and end point)
 - `#min_x`: min x
@@ -197,6 +268,7 @@ Includes `PerfectShape::MultiPoint`
 - `#bounding_box`: bounding box is a rectangle with x = min x, y = min y, and width/height of shape (bounding box only guarantees that the shape is within it, but it might be bigger than the shape)
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
 - `#contain?(x_or_point, y=nil, outline: false, distance_tolerance: 0)`: checks if point is inside when `outline` is `false` or if point is on the outline when `outline` is `true`. `distance_tolerance` can be used as a fuzz factor when `outline` is `true`, for example, to help GUI users mouse-click-select a quadratic bezier curve shape from its outline more successfully
+- `#intersect?(rectangle)`: Returns `true` if intersecting with interior of rectangle or `false` otherwise. This is useful for GUI optimization checks of whether a shape appears in a GUI viewport rectangle and needs redrawing
 - `#curve_center_point`: point at the center of the curve outline (not the center of the bounding box area like `center_x` and `center_y`)
 - `#curve_center_x`: point x coordinate at the center of the curve outline (not the center of the bounding box area like `center_x` and `center_y`)
 - `#curve_center_y`: point y coordinate at the center of the curve outline (not the center of the bounding box area like `center_x` and `center_y`)
@@ -297,6 +369,7 @@ Includes `PerfectShape::RectangularShape`
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
 - `#contain?(x_or_point, y=nil, outline: false, distance_tolerance: 0)`: checks if point is inside when `outline` is `false` or if point is on the outline when `outline` is `true`. `distance_tolerance` can be used as a fuzz factor when `outline` is `true`, for example, to help GUI users mouse-click-select a rectangle shape from its outline more successfully
 - `#edges`: edges of rectangle as `PerfectShape::Line` objects
+- `#out_state(x_or_point, y = nil)`: Returns "out state" of specified point (x,y) (whether it lies to the left, right, top, bottom of rectangle). If point is outside rectangle, it returns a bit mask combination of `Rectangle::OUT_LEFT`, `Rectangle::OUT_RIGHT`, `Rectangle::OUT_TOP`, or `Rectangle::OUT_BOTTOM`. Otherwise, it returns `0` if point is inside the rectangle.
 
 Example:
 
@@ -735,7 +808,7 @@ A composite shape is simply an aggregate of multiple shapes (e.g. square and pol
 - `#center_y`: center y
 - `#bounding_box`: bounding box is a rectangle with x = min x, y = min y, and width/height of shape (bounding box only guarantees that the shape is within it, but it might be bigger than the shape)
 - `#==(other)`: Returns `true` if equal to `other` or `false` otherwise
-- `#contain?(x_or_point, y=nil)`: When `outline` is `false`, it checks if point is inside any of the shapes owned by the composite shape. Otherwise, when `outline` is `true`, it checks if point is on the outline of any of the shapes owned by the composite shape. `distance_tolerance` can be used as a fuzz factor when `outline` is `true`, for example, to help GUI users mouse-click-select a composite shape from its outline more successfully
+- `#contain?(x_or_point, y=nil, outline: false, distance_tolerance: 0)`: When `outline` is `false`, it checks if point is inside any of the shapes owned by the composite shape. Otherwise, when `outline` is `true`, it checks if point is on the outline of any of the shapes owned by the composite shape. `distance_tolerance` can be used as a fuzz factor when `outline` is `true`, for example, to help GUI users mouse-click-select a composite shape from its outline more successfully
 
 Example:
 
@@ -778,7 +851,10 @@ shape.contain?([145, 189], outline: true, distance_tolerance: 1) # => true
 ## Resources
 
 - Rubydoc: https://www.rubydoc.info/gems/perfect-shape
-- AWT Geom Javadoc (inspiration): https://docs.oracle.com/javase/8/docs/api/java/awt/geom/package-summary.html
+- Point in Polygon: https://en.wikipedia.org/wiki/Point_in_polygon
+- Even-Odd Rule: https://en.wikipedia.org/wiki/Even%E2%80%93odd_rule
+- Nonzero Rule: https://en.wikipedia.org/wiki/Nonzero-rule
+- IEEE 754-1985 Remainder: https://en.wikipedia.org/wiki/IEEE_754-1985
 
 ## TODO
 

data/VERSION

@@ -1 +1 @@
-0.3.5
+0.5.2

data/lib/perfect-shape.rb

@@ -23,8 +23,8 @@ $LOAD_PATH.unshift File.expand_path('.', __dir__)
 
 require 'bigdecimal'
 require 'equalizer'
+require 'matrix'
 
-# Perfect Shape algorithms are mostly ported from java.awt.geom: https://docs.oracle.com/javase/8/docs/api/java/awt/geom/package-summary.html
 module PerfectShape
 end
 

data/lib/perfect_shape/affine_transform.rb

@@ -0,0 +1,235 @@
+# Copyright (c) 2021-2022 Andy Maleh
+#
+# Permission is hereby granted, free of charge, to any person obtaining
+# a copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+#
+# The above copyright notice and this permission notice shall be
+# included in all copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+# LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+# OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+# WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+require 'perfect_shape/math'
+require 'perfect_shape/shape'
+require 'perfect_shape/point'
+require 'perfect_shape/multi_point'
+
+module PerfectShape
+  # Represents an affine transform
+  class AffineTransform
+    include Equalizer.new(:xxp, :xyp, :yxp, :yyp, :xt, :yt)
+  
+    attr_reader :xxp, :xyp, :yxp, :yyp, :xt, :yt
+    alias m11 xxp
+    alias m12 xyp
+    alias m21 yxp
+    alias m22 yyp
+    alias m13 xt
+    alias m23 yt
+  
+    # Creates a new AffineTransform with the following Matrix:
+    #
+    # [ xxp xyp xt ]
+    # [ yxp yyp yt ]
+    #
+    # The Matrix is used to transform (x,y) point coordinates as follows:
+    #
+    # [ xxp xyp xt ] * [x] = [ xxp * x + xyp * y + xt ]
+    # [ yxp yyp yt ] * [y] = [ yxp * x + yyp * y + yt ]
+    #
+    # xxp is the x coordinate x product (m11)
+    # xyp is the x coordinate y product (m12)
+    # yxp is the y coordinate x product (m21)
+    # yyp is the y coordinate y product (m22)
+    # xt is the x coordinate translation (m13)
+    # yt is the y coordinate translation (m23)
+    #
+    # The constructor accepts either the (x,y)-operation related argument/kwarg names
+    # or traditional Matrix element kwarg names
+    #
+    # Example with (x,y)-operation kwarg names:
+    #
+    # AffineTransform.new(xxp: 2, xyp: 3, yxp: 4, yyp: 5, xt: 6, yt: 7)
+    #
+    # Example with traditional Matrix element kwarg names:
+    #
+    # AffineTransform.new(m11: 2, m12: 3, m21: 4, m22: 5, m13: 6, m23: 7)
+    #
+    # Example with standard arguments:
+    #
+    # AffineTransform.new(2, 3, 4, 5, 6, 7)
+    #
+    # If no arguments are supplied, it constructs an identity matrix
+    # (i.e. like calling `::new(xxp: 1, xyp: 0, yxp: 0, yyp: 1, xt: 0, yt: 0)`)
+    def initialize(xxp_element = nil, xyp_element = nil, yxp_element = nil, yyp_element = nil, xt_element = nil, yt_element = nil,
+                   xxp: nil, xyp: nil, yxp: nil, yyp: nil, xt: nil, yt: nil,
+                   m11: nil, m12: nil, m21: nil, m22: nil, m13: nil, m23: nil)
+      self.xxp = xxp_element || xxp || m11 || 1
+      self.xyp = xyp_element || xyp || m12 || 0
+      self.yxp = yxp_element || yxp || m21 || 0
+      self.yyp = yyp_element || yyp || m22 || 1
+      self.xt  = xt_element  || xt  || m13 || 0
+      self.yt  = yt_element  || yt  || m23 || 0
+    end
+    
+    def xxp=(value)
+      @xxp = BigDecimal(value.to_s)
+    end
+    alias m11= xxp=
+    
+    def xyp=(value)
+      @xyp = BigDecimal(value.to_s)
+    end
+    alias m12= xyp=
+    
+    def yxp=(value)
+      @yxp = BigDecimal(value.to_s)
+    end
+    alias m21= yxp=
+    
+    def yyp=(value)
+      @yyp = BigDecimal(value.to_s)
+    end
+    alias m22= yyp=
+    
+    def xt=(value)
+      @xt = BigDecimal(value.to_s)
+    end
+    alias m13= xt=
+    
+    def yt=(value)
+      @yt = BigDecimal(value.to_s)
+    end
+    alias m23= yt=
+    
+    # Resets to identity matrix
+    # Returns self to support fluent interface chaining
+    def identity!
+      self.xxp = 1
+      self.xyp = 0
+      self.yxp = 0
+      self.yyp = 1
+      self.xt  = 0
+      self.yt  = 0
+      
+      self
+    end
+    alias reset! identity!
+
+    # Inverts AffineTransform matrix if invertible
+    # Raises an error if affine transform matrix is not invertible
+    # Returns self to support fluent interface chaining
+    def invert!
+      raise 'Cannot invert (matrix is not invertible)!' if !invertible?
+      
+      self.matrix_3d = matrix_3d.inverse
+      
+      self
+    end
+    
+    def invertible?
+      (m11 * m22 - m12 * m21) != 0
+    end
+    
+    # Multiplies by other AffineTransform
+    def multiply!(other_affine_transform)
+      self.matrix_3d = matrix_3d*other_affine_transform.matrix_3d
+      
+      self
+    end
+    
+    # Translates AffineTransform
+    def translate!(x_or_point, y = nil)
+      x, y = Point.normalize_point(x_or_point, y)
+      return unless x && y
+      
+      translation_affine_transform = AffineTransform.new(xxp: 1, xyp: 0, yxp: 0, yyp: 1, xt: x, yt: y)
+      multiply!(translation_affine_transform)
+      
+      self
+    end
+    
+    # Scales AffineTransform
+    def scale!(x_or_point, y = nil)
+      x, y = Point.normalize_point(x_or_point, y)
+      return unless x && y
+      
+      scale_affine_transform = AffineTransform.new(xxp: x, xyp: 0, yxp: 0, yyp: y, xt: 0, yt: 0)
+      multiply!(scale_affine_transform)
+      
+      self
+    end
+    
+    # Rotates AffineTransform counter-clockwise for positive angle value in degrees
+    # or clockwise for negative angle value in degrees
+    def rotate!(degrees)
+      degrees = Math.normalize_degrees(degrees)
+      radians = Math.degrees_to_radians(degrees)
+      
+      rotation_affine_transform = AffineTransform.new(xxp: Math.cos(radians), xyp: -Math.sin(radians), yxp: Math.sin(radians), yyp: Math.cos(radians), xt: 0, yt: 0)
+      multiply!(rotation_affine_transform)
+      
+      self
+    end
+    
+    # Shears AffineTransform by (x,y) amount
+    def shear!(x_or_point, y = nil)
+      x, y = Point.normalize_point(x_or_point, y)
+      return unless x && y
+      
+      shear_affine_transform = AffineTransform.new(xxp: 1 + x*y, xyp: x, yxp: y, yyp: 1, xt: 0, yt: 0)
+      multiply!(shear_affine_transform)
+      
+      self
+    end
+    alias skew! shear!
+    
+    # Sets elements from a Ruby Matrix representing Affine Transform matrix elements in 3D
+    def matrix_3d=(the_matrix_3d)
+      self.xxp = the_matrix_3d[0, 0]
+      self.xyp = the_matrix_3d[0, 1]
+      self.xt = the_matrix_3d[0, 2]
+      self.yxp = the_matrix_3d[1, 0]
+      self.yyp = the_matrix_3d[1, 1]
+      self.yt = the_matrix_3d[1, 2]
+    end
+    
+    # Returns Ruby Matrix representing Affine Transform matrix elements in 3D
+    def matrix_3d
+      Matrix[[xxp, xyp, xt], [yxp, yyp, yt], [0, 0, 1]]
+    end
+    
+    def transform_point(x_or_point, y = nil)
+      x, y = Point.normalize_point(x_or_point, y)
+      return unless x && y
+      
+      [xxp*x + xyp*y + xt, yxp*x + yyp*y + yt]
+    end
+    
+    def inverse_transform_point(x_or_point, y = nil)
+      clone.invert!.transform_point(x_or_point, y)
+    end
+    
+    def transform_points(*xy_coordinates_or_points)
+      points = xy_coordinates_or_points.first.is_a?(Array) ? xy_coordinates_or_points.first : xy_coordinates_or_points
+      points = MultiPoint.normalize_point_array(points)
+      points.map { |point| transform_point(point) }
+    end
+    
+    def inverse_transform_points(*xy_coordinates_or_points)
+      points = xy_coordinates_or_points.first.is_a?(Array) ? xy_coordinates_or_points.first : xy_coordinates_or_points
+      points = MultiPoint.normalize_point_array(points)
+      points.map { |point| inverse_transform_point(point) }
+    end
+  end
+end

data/lib/perfect_shape/arc.rb

@@ -21,10 +21,10 @@
 
 require 'perfect_shape/shape'
 require 'perfect_shape/rectangular_shape'
+require 'perfect_shape/point'
 require 'perfect_shape/line'
 
 module PerfectShape
-  # Mostly ported from java.awt.geom: https://docs.oracle.com/javase/8/docs/api/java/awt/geom/Arc2D.html
   class Arc < Shape
     include RectangularShape
     include Equalizer.new(:type, :x, :y, :width, :height, :start, :extent)
@@ -145,7 +145,7 @@ module PerfectShape
     # the arc, {@code false} if the point lies outside of the
     # arc's bounds.
     def contain?(x_or_point, y = nil, outline: false, distance_tolerance: 0)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       if outline
         if type == :pie && x == center_x && y == center_y

data/lib/perfect_shape/composite_shape.rb

@@ -64,7 +64,7 @@ module PerfectShape
     # the composite shape or false if the point lies outside of the
     # path's bounds.
     def contain?(x_or_point, y = nil, outline: false, distance_tolerance: 0)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       
       shapes.any? { |shape| shape.contain?(x, y, outline: outline, distance_tolerance: distance_tolerance) }

data/lib/perfect_shape/cubic_bezier_curve.rb

@@ -20,10 +20,10 @@
 # WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 require 'perfect_shape/shape'
+require 'perfect_shape/point'
 require 'perfect_shape/multi_point'
 
 module PerfectShape
-  # Mostly ported from java.awt.geom: https://docs.oracle.com/javase/8/docs/api/java/awt/geom/QuadCurve2D.html
   class CubicBezierCurve < Shape
     class << self
       # Calculates the number of times the cubic bézier curve from (x1,y1) to (x2,y2)
@@ -86,7 +86,7 @@ module PerfectShape
     # the cubic bézier curve, {@code false} if the point lies outside of the
     # cubic bézier curve's bounds.
     def contain?(x_or_point, y = nil, outline: false, distance_tolerance: 0)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       
       if outline
@@ -120,7 +120,7 @@ module PerfectShape
     # +1 is added for each crossing where the Y coordinate is increasing
     # -1 is added for each crossing where the Y coordinate is decreasing
     def point_crossings(x_or_point, y = nil, level = 0)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       CubicBezierCurve.point_crossings(points[0][0], points[0][1], points[1][0], points[1][1], points[2][0], points[2][1], points[3][0], points[3][1], x, y, level)
     end
@@ -183,7 +183,7 @@ module PerfectShape
     end
     
     def point_distance(x_or_point, y = nil, minimum_distance_threshold: OUTLINE_MINIMUM_DISTANCE_THRESHOLD)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       
       point = Point.new(x, y)

data/lib/perfect_shape/ellipse.rb

@@ -22,7 +22,6 @@
 require 'perfect_shape/arc'
 
 module PerfectShape
-  # Mostly ported from java.awt.geom: https://docs.oracle.com/javase/8/docs/api/java/awt/geom/Ellipse2D.html
   class Ellipse < Arc
     MESSAGE_CANNOT_UPDATE_ATTRIUBTE = "Ellipse %s cannot be updated. If you want to update type, use Arc instead!"
     
@@ -65,7 +64,7 @@ module PerfectShape
     # ellipse's bounds.
     def contain?(x_or_point, y = nil, outline: false, distance_tolerance: 0)
       # This is implemented again even though super would have just worked to have an optimized algorithm for Ellipse.
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       if outline
         super(x, y, outline: true, distance_tolerance: distance_tolerance)

data/lib/perfect_shape/line.rb

@@ -20,10 +20,10 @@
 # WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 require 'perfect_shape/shape'
+require 'perfect_shape/point'
 require 'perfect_shape/multi_point'
 
 module PerfectShape
-  # Mostly ported from java.awt.geom: https://docs.oracle.com/javase/8/docs/api/java/awt/geom/Line2D.html
   class Line < Shape
     class << self
       # Returns an indicator of where the specified point (px,py) lies with respect to the line segment from
@@ -213,20 +213,20 @@ module PerfectShape
     # the line, {@code false} if the point lies outside of the
     # line's bounds.
     def contain?(x_or_point, y = nil, outline: true, distance_tolerance: 0)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       distance_tolerance = BigDecimal(distance_tolerance.to_s)
       point_distance(x, y) <= distance_tolerance
     end
     
     def point_distance(x_or_point, y = nil)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       Line.point_distance(points[0][0], points[0][1], points[1][0], points[1][1], x, y)
     end
     
     def relative_counterclockwise(x_or_point, y = nil)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       Line.relative_counterclockwise(points[0][0], points[0][1], points[1][0], points[1][1], x, y)
     end
@@ -237,9 +237,34 @@ module PerfectShape
     # +1 is returned for a crossing where the Y coordinate is increasing
     # -1 is returned for a crossing where the Y coordinate is decreasing
     def point_crossings(x_or_point, y = nil)
-      x, y = normalize_point(x_or_point, y)
+      x, y = Point.normalize_point(x_or_point, y)
       return unless x && y
       Line.point_crossings(points[0][0], points[0][1], points[1][0], points[1][1], x, y)
     end
+    
+    def intersect?(rectangle)
+      require 'perfect_shape/rectangle'
+      x1 = points[0][0]
+      y1 = points[0][1]
+      x2 = points[1][0]
+      y2 = points[1][1]
+      out1 = out2 = nil
+      return true if (out2 = rectangle.out_state(x2, y2)) == 0
+      while (out1 = rectangle.out_state(x1, y1)) != 0
+        return false if (out1 & out2) != 0
+        if (out1 & (Rectangle::OUT_LEFT | Rectangle::OUT_RIGHT)) != 0
+          x = rectangle.x
+          x += rectangle.width if (out1 & Rectangle::OUT_RIGHT) != 0
+          y1 = y1 + (x - x1) * (y2 - y1) / (x2 - x1)
+          x1 = x
+        else
+          y = rectangle.y
+          y += rectangle.height if (out1 & Rectangle::OUT_BOTTOM) != 0
+          x1 = x1 + (y - y1) * (x2 - x1) / (y2 - y1)
+          y1 = y
+        end
+      end
+      true
+    end
   end
 end