Releases: taichi-dev/taichi_glsl
v0.0.11
v0.0.10
PyPI packages:
- Remove unexpected
p
ands
directory in wheel file (sorry!) - Bump Taichi requirement to v0.6.32.
GUI system:
- Support choosing
ipython
ormatplotlib
as backend ints.Animation
, e.g.:
self.gui_backend = 'ipython'
This would allows you to display GUI result in Jupyter notebook, checkout the config file.
Field sampling:
dflSample(field, P, dfl)
: return a default valuedfl
whenP
out of range.
v0.0.9
- Upgrade to Taichi new syntax:
ti.var
->ti.field
ti.Vector
->ti.Vector.field
- Examples upgrade to Code Format.
- Fix
ts.randInt
not returning correct range for new Taichi version. - Fix
ts.ballBoundReflect
raisingNameError
. - Functions like
ts.mix
are now@ti.pyfunc
to be called from both Python/Taichi-scope.
v0.0.8
Color map:
- blueorange (the same one used in Taichi THREE)
SPH kernel functions:
- spiky
- ploy6
- dspiky
- dploy6
ODOP classes:
TaichiClass
for generic base of Taichi ODOP data classes (WIP)Complex
based onTaichiClass
(usec @ c
for multiply, WIP)
Taichi hacks:
- matrix.T, a shortcut for
matrix.transpose()
- vector.L, a shortcut for
vector.norm()
- vector.L2, a shortcut for
vector.norm_sqr()
- vector.N, a shortcut for
vector.normalized()
- ti.pi, a shortcut for
math.pi
- ti.tau, a shortcut for
math.tau
Vector math:
- vector.Yx, a shortcut for
tl.vec(vector.y, vector.x)
, and what's more: - If
u = tl.vec(2, 3, 4, 5)
, thenu.y_xW
is(3, 0, 2, -5)
Field sampling:
- D = tl.vec(1, 0, -1), e.g.: use
D.yxy
orD._x_
to gettl.vec(0, 1, 0)
v0.0.7
Field sampling:
- sample (nearest+clamp sample)
- bilerp (bilinear+clamp sample)
- superSample2x2 (anti-aliasing)
Vector math:
- normalizePow (use
normalizePow(x_ij, -2)
to calculate gravity)
Painting helpers:
- paintArrow (simply feed the image tensor)
- sdLine (SDF of a line segment)
Builtin images:
- imageTaichi (from
taichi/examples/taichi_logo.py
) - imageChess (NxN black-white checkers)
- imageGrid (NxN black stroke grid)
Particle simulation:
- ballBoundReflect (often used in cloth simulation)
Color mapping:
- grayscale
- normalmap
GUI base class:
- Set
self.auto_clean = True
to clear image beforeon_render
.
Taichi hacks:
- Turn off the annoying deprecation warning by default.
- No error on
vector = scalar
, usevector.fill(scalar)
by default. vector.product()
to get the product of all elements.ti.static
now accept any argument types.
v0.0.6
Added classes:
Animation
- a Shadertoy-alike GUI base class, including uniforms likeself.iMouse
! Click here for documentation.
Added functions:
boundReflect
- used to compute particles bouncing with rectangular boundaries.
Dependencies:
- taichi: bump from
>=0.6.9
to>=0.6.12
.
(Sorry for the delayed release! My computer gets wrong... and I have to reinstall everything from OS to Taichi..)
The Animation
class in this release is really a big job, please feel free to feed back if you have any idea!
v0.0.5
New features:
- Access vector shuffled components as vector using
v.xy
,v.yx
! For example:
v = vec3(1, 2, 3)
v.xy # vec2(1, 2)
v.yx # vec2(2, 1)
v.zyx # vec2(3, 2, 1)
v.z # 3
Added classes:
TaichiClass
for ODOP infrastructure, documentation WIP.
Added functions:
- inversesqrt
- invLength
- sqrLength
Taichi hijackers:
ti.Vector.var
now supported in lower Taichi version.ti.static
now support all type of inputs and no error will be raised!
Let me know if you feel like some of these changes or not by opening an issue on GitHub.
v0.0.4
Added features:
vec
now support GLSL-alike composing from vectors, e.g.:
u = vec(2, 3) # scalar + scalar = 2D vector
v = vec(u, 3) # 2D vector + scalar = 3D vector
vec2
, vec3
now also supports fill vector with a scalar, e.g.:
u = vec2(0) # u = vec(0, 0)
v = vec3(3) # v = vec(3, 3, 3)
New module ts.phys
with function momentumExchange
.
May be used to exchange momentum when collision.
Added functions:
- randUnit3D
- vec2
- vec3
- vec4
- momentumExchange
v0.0.3
News:
Documentation now available at https://taichi-glsl.rtfd.io!
Added functions:
- reflect
- refract
- rand
- randND
- randRange
- randNDRange
- randUnit2D
- randSolid2D
- rand
v0.0.2
Added functions:
- exp
- log
- tensor (unlike
array
shape default toNone
) - vec_tensor
- mat_tensor
- isinf
- isnan (taichi-dev/taichi#1223)