Scalar, viewport and interoperability helpers¶
Import these functions and GLfloat from gem.common.
Source, utility tests,
viewport contracts,
angles.
These historical raw sequence utilities preserve inputs and return independent
containers unless noted. Shape/type validation is limited; short/malformed
sequences can raise IndexError/TypeError from direct indexing or ctypes conversion.
Do not infer a uniform invalid-input API from ordinary examples.
GLfloat is exactly ctypes.c_float, a binary32 type alias, not an OpenGL dependency.
It is the only intended public common constant/type alias; imported math, sm and
ct are implementation dependencies, not mathematical exports.
| Exact source declaration | Parameters, result and behavior |
|---|---|
convertArr(l, n) |
l: flat sliceable sequence, n: positive integer chunk width; return new list of slices l[i:i+n], final chunk may be shorter. n=0 raises ValueError from range; negative step yields empty result for ordinary input. Slice ownership follows supplied sequence semantics. |
mulV4(v1, v2) |
v1,v2: raw four-component numeric sequences; fresh list of four componentwise products, not dot/matrix multiplication. |
conv_list(listIn, cType) |
listIn: numeric/value list; cType: ctypes element type; fresh initialized cType*len array. Native ctypes conversion errors propagate; inputs preserved. |
conv_list_2d(listIn, cType) |
listIn: nonempty rectangular row lists; cType: ctypes type; fresh row-major nested ctypes array with lengths inferred from row count/first row. Empty outer list raises IndexError; mismatched shapes retain index behavior. |
list_2d_to_1d(inlist) |
inlist: nonempty rectangular rows; fresh row-major flat list of numeric elements. Width from first row; no ragged validation or deep copying of arbitrary objects. |
convertM4to3(matrix) |
Raw nested matrix: at least leading 3×3 entries, normally 4×4; fresh copy of leading 3×3 rows. Not a Matrix wrapper or affine transform adapter. |
sinc(x) |
Numeric x: 1.0 when abs(x)<1e−4, otherwise sin(x)/x; historical approximation at small angles, no mutation. Not the exact center-limit implementation of angular-probe quadrature. |
scalarLerp(a, b, time) |
Numeric a,b,time; scalar a+time*(b−a), no clamping or mutation. |
getViewPort(coords, width, height) |
coords: finite nonzero Vector2/3/4, numeric width/height; fresh four-element list from whole-vector normalization with original XY offsets. Zero Vector raises ZeroDivisionError; inputs preserved. Formula below. |
radiansToDegrees(degrees) |
Numeric degrees (misleading retained parameter name) actually represents radians; return input*180/pi. |
degreesToRadians(radians) |
Numeric radians (misleading retained parameter name) actually represents degrees; return input*pi/180. |
sign(x) |
Numeric x; return +1.0 if x≥0, including signed zero, otherwise −1.0. NaN follows else branch; no new nonfinite policy. |
Historical viewport¶
For L=norm(coords), output is [(x/L+1)width/2+x, (y/L+1)height/2+y, width, height]. Z/W affect the normalization for Vector3/4; original XY are offsets. This is neither OpenGL glViewport nor conventional NDC-to-window projection. Use project/unproject for those coordinate conversions. Widths/heights and wider malformed/nonfinite domains are not standardized here.
ctypes ownership¶
ctypes helpers create independent arrays; converting to binary32 can round/overflow binary64 values. Keep the array alive while foreign code uses a pointer. Matrix snapshots are documented in matrix; no foreign graphics upload is verified by these memory checks.
import ctypes
import math
from gem.common import GLfloat, conv_list_2d, list_2d_to_1d, convertArr, getViewPort, radiansToDegrees, degreesToRadians
from gem.vector import Vector
rows = [[1, 2], [3, 4]]
array = conv_list_2d(rows, GLfloat)
assert GLfloat is ctypes.c_float and list(array[1]) == [3.0, 4.0]
rows[1][0] = 99
assert array[1][0] == 3.0
assert list_2d_to_1d([[1, 2], [3, 4]]) == [1, 2, 3, 4]
assert convertArr([1, 2, 3], 2) == [[1, 2], [3]]
coords = Vector(2, [3, 4])
assert getViewPort(coords, 100, 200) == [83.0, 184.0, 100, 200]
assert coords.vector == [3, 4]
assert radiansToDegrees(math.pi) == 180.0
assert degreesToRadians(180) == math.pi