Rect

Rect represents a rectangle defined by four floating point numbers x0, y0, x1, y1. They are viewed as being coordinates of two diagonally opposite points. The first two numbers are regarded as the “top left” corner P_x0,y0 and P_x1,y1 as the “bottom right” one. However, these two properties need not coincide with their intuitive meanings - read on.

The following remarks are also valid for IRect objects:

• Rectangle borders are always parallel to the respective X- and Y-axes.
• The constructing points can be anywhere in the plane - they need not even be different, and e.g. “top left” need not be the geometrical “north-western” point.
• For any given quadruple of numbers, the geometrically “same” rectangle can be defined in (up to) four different ways: Rect(P_x0,y0, P_x1,y1), Rect(P_x1,y1, P_x0,y0), Rect(P_x0,y1, P_x1,y0), and Rect(P_x1,y0, P_x0,y1).

Hence some useful classification:

• A rectangle is called finite if x0 <= x1 and y0 <= y1 (i.e. the bottom right point is “south-eastern” to the top left one), otherwise infinite. Of the four alternatives above, only one is finite (disregarding degenerate cases).
• A rectangle is called empty if x0 = x1 or y0 = y1, i.e. if its area is zero.

Class API

class Rect

__init__(self)

__init__(self, x0, y0, x1, y1)

__init__(self, top_left, bottom_right)

__init__(self, top_left, x1, y1)

__init__(self, x0, y0, bottom_right)

__init__(self, rect)

__init__(self, list)

Overloaded constructors: top_left, bottom_right stand for Point objects, list is a Python sequence type with length 4, rect means another Rect, while the other parameters mean float coordinates. If list is specified, it is the user’s responsibility to only provide numeric entries - no error checking is done, and invalid entries will receive a value of -1.0.

If rect is specified, the constructor creates a new copy of rect.

Without parameters, the rectangle Rect(0.0, 0.0, 0.0, 0.0) is created.

round()

Creates the smallest containing IRect (this is not the same as simply rounding the rectangle’s edges!).

1. If the rectangle is infinite, the “normalized” (finite) version of it will be taken. The result of this method is always a finite IRect.
2. If the rectangle is empty, the result is also empty.
3. Possible paradox: The result may be empty, even if the rectangle is not empty! In such cases, the result obviously does not contain the rectangle. This is because MuPDF’s algorithm allows for a small tolerance (1e-3). Example:

>>> r = fitz.Rect(100, 100, 200, 100.001)
>>> r.isEmpty
False
>>> r.round()
fitz.IRect(100, 100, 200, 100)
>>> r.round().isEmpty
True

To reproduce the effect on your platform, you may need to adjust the numbers a little.

Return type:	IRect

transform(m)

Transforms the rectangle with a matrix and replaces the original. If the rectangle is empty or infinite, this is a no-operation.

Parameters:	m (Matrix) – The matrix for the transformation.
Return type:	Rect
Returns:	the smallest rectangle that contains the transformed original.

intersect(r)

The intersection (common rectangular area) of the current rectangle and r is calculated and replaces the current rectangle. If either rectangle is empty, the result is also empty. If r is infinite, this is a no-operation.

Parameters:	r (Rect) – Second rectangle

includeRect(r)

The smallest rectangle containing the current one and r is calculated and replaces the current one. If either rectangle is infinite, the result is also infinite. If one is empty, the other one will be taken as the result.

Parameters:	r (Rect) – Second rectangle

includePoint(p)

The smallest rectangle containing the current one and point p is calculated and replaces the current one. Infinite rectangles remain unchanged. To create a rectangle containing a series of points, start with (the empty) fitz.Rect(p1, p1) and successively perform includePoint operations for the other points.

Parameters:	p (Point) – Point to include.

getRectArea([unit])

getArea([unit])

Calculate the area of the rectangle and, with no parameter, equals abs(rect). Like an empty rectangle, the area of an infinite rectangle is also zero. So, at least one of fitz.Rect(p1, p2) and fitz.Rect(p2, p1) has a zero area.

Parameters:	unit (str) – Specify required unit: respective squares of px (pixels, default), in (inches), cm (centimeters), or mm (millimeters).
Return type:	float

contains(x)

Checks whether x is contained in the rectangle. It may be an IRect, Rect, Point or number. If x is an empty rectangle, this is always true. If the rectangle is empty this is always False for all non-empty rectangles and for all points. If x is a number, it will be checked against the four components. x in rect and rect.contains(x) are equivalent.

Parameters:	x (IRect or Rect or Point or number) – the object to check.
Return type:	bool

intersects(r)

Checks whether the rectangle and r (a Rect or IRect) have a non-empty rectangle in common. This will always be False if either is infinite or empty.

Parameters:	r (IRect or Rect) – the rectangle to check.
Return type:	bool

normalize()

Replace the rectangle with its finite version. This is done by shuffling the rectangle corners. After completion of this method, the bottom right corner will indeed be south-eastern to the top left one.

irect

Equals result of method round().

top_left

tl

Equals Point(x0, y0).

Type:	Point

top_right

tr

Equals Point(x1, y0).

Type:	Point

bottom_left

bl

Equals Point(x0, y1).

Type:	Point

bottom_right

br

Equals Point(x1, y1).

Type:	Point

width

Contains the width of the rectangle. Equals x1 - x0.

Return type:	float

height

Contains the height of the rectangle. Equals y1 - y0.

Return type:	float

x0

X-coordinate of the left corners.

Type:	float

y0

Y-coordinate of the top corners.

Type:	float

x1

X-coordinate of the right corners.

Type:	float

y1

Y-coordinate of the bottom corners.

Type:	float

isInfinite

True if rectangle is infinite, False otherwise.

Type:	bool

isEmpty

True if rectangle is empty, False otherwise.

Type:	bool

>>> p1 = fitz.Point(10, 10)
>>> p2 = fitz.Point(300, 450)
>>>
>>> fitz.Rect(p1, p2)
fitz.Rect(10.0, 10.0, 300.0, 450.0)
>>>
>>> fitz.Rect(10, 10, 300, 450)
fitz.Rect(10.0, 10.0, 300.0, 450.0)
>>>
>>> fitz.Rect(10, 10, p2)
fitz.Rect(10.0, 10.0, 300.0, 450.0)
>>>
>>> fitz.Rect(p1, 300, 450)
fitz.Rect(10.0, 10.0, 300.0, 450.0)

>>> r = fitz.Rect(0.5, -0.01, 123.88, 455.123456)
>>>
>>> r
fitz.Rect(0.5, -0.009999999776482582, 123.87999725341797, 455.1234436035156)
>>>
>>> r.round()     # = r.irect
fitz.IRect(0, -1, 124, 456)

>>> m = fitz.Matrix(45)
>>> r = fitz.Rect(10, 10, 410, 610)
>>> r * m
fitz.Rect(-424.2640686035156, 14.142135620117188, 282.84271240234375, 721.2489013671875)
>>>
>>> r | fitz.Point(5, 5)
fitz.Rect(5.0, 5.0, 410.0, 610.0)
>>>
>>> r + 5
fitz.Rect(15.0, 15.0, 415.0, 615.0)
>>>
>>> r & fitz.Rect(0, 0, 15, 15)
fitz.Rect(10.0, 10.0, 15.0, 15.0)

>>> r = fitz.Rect(...)     # any rectangle
>>> ir = r.irect           # its IRect version
>>> # even though you get ...
>>> ir in r
True
>>> # ... and ...
>>> r in ir
True
>>> # ... r and ir are still different types!
>>> r == ir
False
>>> # corners are always part of non-epmpty rectangles
>>> r.bottom_left in r
True
>>>
>>> # numbers are checked against coordinates
>>> r.x0 in r
True

>>> r = fitz.Rect(...)     # any rectangle
>>>
>>> # alternative 1
>>> s = fitz.Rect(r.top_left, r.top_left)   # just a point
>>> s | r.bottom_right     # s is a finite rectangle!
>>>
>>> # alternative 2
>>> s = (+r).normalize()
>>> # r.normalize() changes r itself!

>>> r  = fitz.Rect(...)
>>> r1 = r + (-5, -5, 5, 5)

>>> r /= (1, 2, 3, 4, 5, 6)