batteries/vec2.lua

619 lines
13 KiB
Lua

--[[
2d vector type
]]
local path = (...):gsub("vec2", "")
local class = require(path .. "class")
local math = require(path .. "mathx") --shadow global math module
local make_pooled = require(path .. "make_pooled")
local vec2 = class({
name = "vec2",
})
--stringification
function vec2:__tostring()
return ("(%.2f, %.2f)"):format(self.x, self.y)
end
--ctor
function vec2:new(x, y)
if type(x) == "number" then
self:scalar_set(x, y)
elseif type(x) == "table" then
if type(x.type) == "function" and x:type() == "vec2" then
self:vector_set(x)
elseif x[1] then
self:scalar_set(x[1], x[2])
else
self:scalar_set(x.x, x.y)
end
else
self:scalar_set(0)
end
end
--explicit ctors; mostly vestigial at this point
function vec2:copy()
return vec2(self.x, self.y)
end
function vec2:xy(x, y)
return vec2(x, y)
end
function vec2:polar(length, angle)
return vec2(length, 0):rotate_inplace(angle)
end
function vec2:filled(v)
return vec2(v, v)
end
function vec2:zero()
return vec2(0)
end
--unpack for multi-args
function vec2:unpack()
return self.x, self.y
end
--pack when a sequence is needed
function vec2:pack()
return {self:unpack()}
end
--shared pooled storage
make_pooled(vec2, 128)
--get a pooled copy of an existing vector
function vec2:pooled_copy()
return vec2:pooled(self)
end
--modify
function vec2:vector_set(v)
self.x = v.x
self.y = v.y
return self
end
function vec2:scalar_set(x, y)
if not y then y = x end
self.x = x
self.y = y
return self
end
function vec2:swap(v)
local sx, sy = self.x, self.y
self.x, self.y = v.x, v.y
v.x, v.y = sx, sy
return self
end
-----------------------------------------------------------
--equality comparison
-----------------------------------------------------------
--threshold for equality in each dimension
local EQUALS_EPSILON = 1e-9
--true if a and b are functionally equivalent
function vec2.equals(a, b)
return (
math.abs(a.x - b.x) <= EQUALS_EPSILON and
math.abs(a.y - b.y) <= EQUALS_EPSILON
)
end
--true if a and b are not functionally equivalent
--(very slightly faster than `not vec2.equals(a, b)`)
function vec2.nequals(a, b)
return (
math.abs(a.x - b.x) > EQUALS_EPSILON or
math.abs(a.y - b.y) > EQUALS_EPSILON
)
end
--alias
vec2.not_equals = vec2.nequals
-----------------------------------------------------------
--arithmetic
-----------------------------------------------------------
--vector
function vec2:vector_add_inplace(v)
self.x = self.x + v.x
self.y = self.y + v.y
return self
end
function vec2:vector_sub_inplace(v)
self.x = self.x - v.x
self.y = self.y - v.y
return self
end
function vec2:vector_mul_inplace(v)
self.x = self.x * v.x
self.y = self.y * v.y
return self
end
function vec2:vector_div_inplace(v)
self.x = self.x / v.x
self.y = self.y / v.y
return self
end
--(a + (b * t))
--useful for integrating physics and adding directional offsets
function vec2:fused_multiply_add_inplace(v, t)
self.x = self.x + (v.x * t)
self.y = self.y + (v.y * t)
return self
end
--scalar
function vec2:scalar_add_inplace(x, y)
if not y then y = x end
self.x = self.x + x
self.y = self.y + y
return self
end
function vec2:scalar_sub_inplace(x, y)
if not y then y = x end
self.x = self.x - x
self.y = self.y - y
return self
end
function vec2:scalar_mul_inplace(x, y)
if not y then y = x end
self.x = self.x * x
self.y = self.y * y
return self
end
function vec2:scalar_div_inplace(x, y)
if not y then y = x end
self.x = self.x / x
self.y = self.y / y
return self
end
-----------------------------------------------------------
-- geometric methods
-----------------------------------------------------------
function vec2:length_squared()
return self.x * self.x + self.y * self.y
end
function vec2:length()
return math.sqrt(self:length_squared())
end
function vec2:distance_squared(other)
local dx = self.x - other.x
local dy = self.y - other.y
return dx * dx + dy * dy
end
function vec2:distance(other)
return math.sqrt(self:distance_squared(other))
end
function vec2:normalise_both_inplace()
local len = self:length()
if len == 0 then
return self, 0
end
return self:scalar_div_inplace(len), len
end
function vec2:normalise_inplace()
local v, len = self:normalise_both_inplace()
return v
end
function vec2:normalise_len_inplace()
local v, len = self:normalise_both_inplace()
return len
end
function vec2:inverse_inplace()
return self:scalar_mul_inplace(-1)
end
-- angle/direction specific
function vec2:rotate_inplace(angle)
local s = math.sin(angle)
local c = math.cos(angle)
local ox = self.x
local oy = self.y
self.x = c * ox - s * oy
self.y = s * ox + c * oy
return self
end
function vec2:rotate_around_inplace(angle, pivot)
self:vector_sub_inplace(pivot)
self:rotate_inplace(angle)
self:vector_add_inplace(pivot)
return self
end
--fast quarter/half rotations
function vec2:rot90r_inplace()
local ox = self.x
local oy = self.y
self.x = -oy
self.y = ox
return self
end
function vec2:rot90l_inplace()
local ox = self.x
local oy = self.y
self.x = oy
self.y = -ox
return self
end
vec2.rot180_inplace = vec2.inverse_inplace --alias
--get the angle of this vector relative to (1, 0)
function vec2:angle()
return math.atan2(self.y, self.x)
end
--get the normalised difference in angle between two vectors
function vec2:angle_difference(v)
return math.angle_difference(self:angle(), v:angle())
end
--lerp towards the direction of a provided vector
--(length unchanged)
function vec2:lerp_direction_inplace(v, t)
return self:rotate_inplace(self:angle_difference(v) * t)
end
-----------------------------------------------------------
-- per-component clamping ops
-----------------------------------------------------------
function vec2:min_inplace(v)
self.x = math.min(self.x, v.x)
self.y = math.min(self.y, v.y)
return self
end
function vec2:max_inplace(v)
self.x = math.max(self.x, v.x)
self.y = math.max(self.y, v.y)
return self
end
function vec2:clamp_inplace(min, max)
self.x = math.clamp(self.x, min.x, max.x)
self.y = math.clamp(self.y, min.y, max.y)
return self
end
-----------------------------------------------------------
-- absolute value
-----------------------------------------------------------
function vec2:abs_inplace()
self.x = math.abs(self.x)
self.y = math.abs(self.y)
return self
end
-----------------------------------------------------------
-- sign
-----------------------------------------------------------
function vec2:sign_inplace()
self.x = math.sign(self.x)
self.y = math.sign(self.y)
return self
end
-----------------------------------------------------------
-- truncation/rounding
-----------------------------------------------------------
function vec2:floor_inplace()
self.x = math.floor(self.x)
self.y = math.floor(self.y)
return self
end
function vec2:ceil_inplace()
self.x = math.ceil(self.x)
self.y = math.ceil(self.y)
return self
end
function vec2:round_inplace()
self.x = math.round(self.x)
self.y = math.round(self.y)
return self
end
-----------------------------------------------------------
-- interpolation
-----------------------------------------------------------
function vec2:lerp_inplace(other, amount)
self.x = math.lerp(self.x, other.x, amount)
self.y = math.lerp(self.y, other.y, amount)
return self
end
function vec2:lerp_eps_inplace(other, amount, eps)
self.x = math.lerp_eps(self.x, other.x, amount, eps)
self.y = math.lerp_eps(self.y, other.y, amount, eps)
return self
end
-----------------------------------------------------------
-- vector products and projections
-----------------------------------------------------------
function vec2:dot(other)
return self.x * other.x + self.y * other.y
end
--"fake", but useful - also called the wedge product apparently
function vec2:cross(other)
return self.x * other.y - self.y * other.x
end
function vec2:scalar_projection(other)
local len = other:length()
if len == 0 then
return 0
end
return self:dot(other) / len
end
function vec2:vector_projection_inplace(other)
local div = other:dot(other)
if div == 0 then
return self:scalar_set(0)
end
local fac = self:dot(other) / div
return self:vector_set(other):scalar_mul_inplace(fac)
end
function vec2:vector_rejection_inplace(other)
local tx, ty = self.x, self.y
self:vector_projection_inplace(other)
self:scalar_set(tx - self.x, ty - self.y)
return self
end
--get the winding side of p, relative to the line a-b
-- (this is based on the signed area of the triangle a-b-p)
-- return value:
-- >0 when p left of line
-- =0 when p on line
-- <0 when p right of line
function vec2.winding_side(a, b, p)
return (b.x - a.x) * (p.y - a.y)
- (p.x - a.x) * (b.y - a.y)
end
--return whether a is nearer to v than b
function vec2.nearer(v, a, b)
return v:distance_squared(a) < v:distance_squared(b)
end
-----------------------------------------------------------
-- vector extension methods for special purposes
-- (any common vector ops worth naming)
-----------------------------------------------------------
--"physical" friction
function vec2:apply_friction_inplace(mu, dt)
local friction = self:pooled_copy():scalar_mul_inplace(mu * dt)
if friction:length_squared() > self:length_squared() then
self:scalar_set(0, 0)
else
self:vector_sub_inplace(friction)
end
friction:release()
return self
end
--"gamey" friction in one dimension
local function _friction_1d(v, mu, dt)
local friction = mu * v * dt
if math.abs(friction) > math.abs(v) then
return 0
else
return v - friction
end
end
--"gamey" friction in both dimensions
function vec2:apply_friction_xy_inplace(mu_x, mu_y, dt)
self.x = _friction_1d(self.x, mu_x, dt)
self.y = _friction_1d(self.y, mu_y, dt)
return self
end
--minimum/maximum components
function vec2:mincomp()
return math.min(self.x, self.y)
end
function vec2:maxcomp()
return math.max(self.x, self.y)
end
-- meta functions for mathmatical operations
function vec2.__add(a, b)
return a:vector_add(b)
end
function vec2.__sub(a, b)
return a:vector_sub(b)
end
function vec2.__mul(a, b)
if type(a) == "number" then
return b:scalar_mul(a)
elseif type(b) == "number" then
return a:scalar_mul(b)
else
return a:vector_mul(b)
end
end
function vec2.__div(a, b)
if type(b) == "number" then
return a:scalar_div(b)
else
return a:vector_div(b)
end
end
-- mask out min component, with preference to keep x
function vec2:major_inplace()
if self.x > self.y then
self.y = 0
else
self.x = 0
end
return self
end
-- mask out max component, with preference to keep x
function vec2:minor_inplace()
if self.x < self.y then
self.y = 0
else
self.x = 0
end
return self
end
--vector_ free alias; we're a vector library, so semantics should default to vector
vec2.add_inplace = vec2.vector_add_inplace
vec2.sub_inplace = vec2.vector_sub_inplace
vec2.mul_inplace = vec2.vector_mul_inplace
vec2.div_inplace = vec2.vector_div_inplace
vec2.set = vec2.vector_set
--american spelling alias
vec2.normalize_both_inplace = vec2.normalise_both_inplace
vec2.normalize_inplace = vec2.normalise_inplace
vec2.normalize_len_inplace = vec2.normalise_len_inplace
--garbage generating functions that return a new vector rather than modifying self
for _, inplace_name in ipairs({
"vector_add_inplace",
"vector_sub_inplace",
"vector_mul_inplace",
"vector_div_inplace",
"fused_multiply_add_inplace",
"add_inplace",
"sub_inplace",
"mul_inplace",
"div_inplace",
"scalar_add_inplace",
"scalar_sub_inplace",
"scalar_mul_inplace",
"scalar_div_inplace",
"normalise_both_inplace",
"normalise_inplace",
"normalise_len_inplace",
"normalize_both_inplace",
"normalize_inplace",
"normalize_len_inplace",
"inverse_inplace",
"rotate_inplace",
"rotate_around_inplace",
"rot90r_inplace",
"rot90l_inplace",
"lerp_direction_inplace",
"min_inplace",
"max_inplace",
"clamp_inplace",
"abs_inplace",
"sign_inplace",
"floor_inplace",
"ceil_inplace",
"round_inplace",
"lerp_inplace",
"lerp_eps_inplace",
"vector_projection_inplace",
"vector_rejection_inplace",
"apply_friction_inplace",
"apply_friction_xy_inplace",
"major_inplace",
"minor_inplace",
}) do
local garbage_name = inplace_name:gsub("_inplace", "")
vec2[garbage_name] = function(self, ...)
self = self:copy()
return self[inplace_name](self, ...)
end
end
--"hungarian" shorthand aliases for compatibility and short names
--
--i do encourage using the longer versions above as it makes code easier
--to understand when you come back, but i also appreciate wanting short code
for _, v in ipairs({
{"sset", "scalar_set"},
{"sadd", "scalar_add"},
{"ssub", "scalar_sub"},
{"smul", "scalar_mul"},
{"sdiv", "scalar_div"},
{"vset", "vector_set"},
{"vadd", "vector_add"},
{"vsub", "vector_sub"},
{"vmul", "vector_mul"},
{"vdiv", "vector_div"},
--(no plain addi etc, imo it's worth differentiating vaddi vs saddi)
{"fma", "fused_multiply_add"},
{"vproj", "vector_projection"},
{"vrej", "vector_rejection"},
--just for the _inplace -> i shorthand, mostly for backwards compatibility
{"min", "min"},
{"max", "max"},
{"clamp", "clamp"},
{"abs", "abs"},
{"sign", "sign"},
{"floor", "floor"},
{"ceil", "ceil"},
{"round", "round"},
{"lerp", "lerp"},
{"rotate", "rotate"},
{"normalise", "normalise"},
{"normalize", "normalize"},
}) do
local shorthand, original = v[1], v[2]
if vec2[shorthand] == nil then
vec2[shorthand] = vec2[original]
end
--and inplace version
shorthand = shorthand .. "i"
original = original .. "_inplace"
if vec2[shorthand] == nil then
vec2[shorthand] = vec2[original]
end
end
return vec2