use runmat_accelerate_api::GpuTensorHandle;
use runmat_builtins::{ComplexTensor, Tensor, Value};
use runmat_macros::runtime_builtin;
use crate::builtins::common::random_args::complex_tensor_into_value;
use crate::builtins::common::{gpu_helpers, tensor};
use crate::builtins::math::trigonometry::degree_helpers::reduce_degrees;
use crate::builtins::math::type_resolvers::numeric_unary_type;
use crate::{build_runtime_error, BuiltinResult, RuntimeError};
const BUILTIN_NAME: &str = "sind";
const DEG_TO_RAD: f64 = std::f64::consts::PI / 180.0;
fn builtin_error(message: impl Into<String>) -> RuntimeError {
build_runtime_error(message)
.with_builtin(BUILTIN_NAME)
.build()
}
#[inline]
fn sind_scalar(x: f64) -> f64 {
let Some(phi) = reduce_degrees(x) else {
return f64::NAN;
};
if phi == 0.0 || phi == 180.0 {
0.0
} else if phi == 90.0 {
1.0
} else if phi == -90.0 {
-1.0
} else if phi == 30.0 || phi == 150.0 {
0.5
} else if phi == -30.0 || phi == -150.0 {
-0.5
} else {
(x * DEG_TO_RAD).sin()
}
}
#[inline]
fn sind_complex(re: f64, im: f64) -> (f64, f64) {
let scaled_re = re * DEG_TO_RAD;
let scaled_im = im * DEG_TO_RAD;
(
scaled_re.sin() * scaled_im.cosh(),
scaled_re.cos() * scaled_im.sinh(),
)
}
#[runtime_builtin(
name = "sind",
category = "math/trigonometry",
summary = "Sine of input expressed in degrees.",
keywords = "sind,sine,degrees,trigonometry",
accel = "unary",
type_resolver(numeric_unary_type),
builtin_path = "crate::builtins::math::trigonometry::sind"
)]
async fn sind_builtin(value: Value) -> BuiltinResult<Value> {
match value {
Value::GpuTensor(handle) => sind_gpu(handle).await,
Value::Complex(re, im) => {
let (out_re, out_im) = sind_complex(re, im);
Ok(Value::Complex(out_re, out_im))
}
Value::ComplexTensor(ct) => sind_complex_tensor(ct),
Value::String(_) | Value::StringArray(_) => {
Err(builtin_error("sind: expected numeric input"))
}
other => sind_real(other),
}
}
async fn sind_gpu(handle: GpuTensorHandle) -> BuiltinResult<Value> {
let tensor = gpu_helpers::gather_tensor_async(&handle).await?;
sind_tensor(tensor).map(tensor::tensor_into_value)
}
fn sind_real(value: Value) -> BuiltinResult<Value> {
let tensor = tensor::value_into_tensor_for(BUILTIN_NAME, value).map_err(builtin_error)?;
sind_tensor(tensor).map(tensor::tensor_into_value)
}
fn sind_tensor(tensor: Tensor) -> BuiltinResult<Tensor> {
let data = tensor
.data
.iter()
.map(|&value| sind_scalar(value))
.collect::<Vec<_>>();
Tensor::new(data, tensor.shape.clone()).map_err(|err| builtin_error(format!("sind: {err}")))
}
fn sind_complex_tensor(tensor: ComplexTensor) -> BuiltinResult<Value> {
let data = tensor
.data
.iter()
.map(|&(re, im)| sind_complex(re, im))
.collect::<Vec<_>>();
let converted = ComplexTensor::new(data, tensor.shape.clone())
.map_err(|err| builtin_error(format!("sind: {err}")))?;
Ok(complex_tensor_into_value(converted))
}
#[cfg(test)]
pub(crate) mod tests {
use super::*;
use futures::executor::block_on;
use runmat_builtins::{IntValue, LogicalArray, ResolveContext, Type};
fn sind_builtin(value: Value) -> BuiltinResult<Value> {
block_on(super::sind_builtin(value))
}
fn error_message(err: RuntimeError) -> String {
err.message().to_string()
}
fn expect_num(value: Value) -> f64 {
match value {
Value::Num(v) => v,
other => panic!("expected scalar result, got {other:?}"),
}
}
#[test]
fn sind_type_preserves_tensor_shape() {
let out = numeric_unary_type(
&[Type::Tensor {
shape: Some(vec![Some(2), Some(3)]),
}],
&ResolveContext::new(Vec::new()),
);
assert_eq!(
out,
Type::Tensor {
shape: Some(vec![Some(2), Some(3)])
}
);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_exact_values_first_period() {
assert_eq!(expect_num(sind_builtin(Value::Num(0.0)).unwrap()), 0.0);
assert_eq!(expect_num(sind_builtin(Value::Num(30.0)).unwrap()), 0.5);
assert_eq!(expect_num(sind_builtin(Value::Num(90.0)).unwrap()), 1.0);
assert_eq!(expect_num(sind_builtin(Value::Num(150.0)).unwrap()), 0.5);
assert_eq!(expect_num(sind_builtin(Value::Num(180.0)).unwrap()), 0.0);
assert_eq!(expect_num(sind_builtin(Value::Num(210.0)).unwrap()), -0.5);
assert_eq!(expect_num(sind_builtin(Value::Num(270.0)).unwrap()), -1.0);
assert_eq!(expect_num(sind_builtin(Value::Num(330.0)).unwrap()), -0.5);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_exact_values_negative_and_wrapped() {
assert_eq!(expect_num(sind_builtin(Value::Num(360.0)).unwrap()), 0.0);
assert_eq!(expect_num(sind_builtin(Value::Num(540.0)).unwrap()), 0.0);
assert_eq!(expect_num(sind_builtin(Value::Num(-30.0)).unwrap()), -0.5);
assert_eq!(expect_num(sind_builtin(Value::Num(-90.0)).unwrap()), -1.0);
assert_eq!(expect_num(sind_builtin(Value::Num(-180.0)).unwrap()), 0.0);
assert_eq!(expect_num(sind_builtin(Value::Num(450.0)).unwrap()), 1.0);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_int_input_returns_exact() {
assert_eq!(
expect_num(sind_builtin(Value::Int(IntValue::I32(180))).unwrap()),
0.0,
);
assert_eq!(
expect_num(sind_builtin(Value::Int(IntValue::I32(30))).unwrap()),
0.5,
);
assert_eq!(
expect_num(sind_builtin(Value::Int(IntValue::I64(-90))).unwrap()),
-1.0,
);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_non_exact_value_matches_radian_formula() {
let degrees = 45.0_f64;
let actual = expect_num(sind_builtin(Value::Num(degrees)).unwrap());
let expected = (degrees * DEG_TO_RAD).sin();
assert!((actual - expected).abs() < 1e-12);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_tensor_preserves_shape() {
let tensor = Tensor::new(vec![0.0, 30.0, 90.0, 180.0], vec![2, 2]).unwrap();
let result = sind_builtin(Value::Tensor(tensor)).expect("sind");
match result {
Value::Tensor(t) => {
assert_eq!(t.shape, vec![2, 2]);
assert_eq!(t.data, vec![0.0, 0.5, 1.0, 0.0]);
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_logical_array_promotes() {
let logical = LogicalArray::new(vec![0, 1], vec![1, 2]).unwrap();
let result = sind_builtin(Value::LogicalArray(logical)).expect("sind");
match result {
Value::Tensor(t) => {
assert_eq!(t.shape, vec![1, 2]);
assert_eq!(t.data[0], 0.0);
let expected = (1.0_f64 * DEG_TO_RAD).sin();
assert!((t.data[1] - expected).abs() < 1e-12);
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_nan_propagates() {
let result = expect_num(sind_builtin(Value::Num(f64::NAN)).unwrap());
assert!(result.is_nan());
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_inf_is_nan() {
let pos = expect_num(sind_builtin(Value::Num(f64::INFINITY)).unwrap());
let neg = expect_num(sind_builtin(Value::Num(f64::NEG_INFINITY)).unwrap());
assert!(pos.is_nan());
assert!(neg.is_nan());
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_complex_uses_radian_formula() {
let result = sind_builtin(Value::Complex(180.0, 0.0)).expect("sind");
match result {
Value::Complex(re, im) => {
let (expected_re, expected_im) = sind_complex(180.0, 0.0);
assert!((re - expected_re).abs() < 1e-15);
assert!((im - expected_im).abs() < 1e-15);
assert_eq!(im, 0.0);
assert!(re.abs() < 1e-12);
}
other => panic!("expected complex result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_complex_off_axis_matches_formula() {
let result = sind_builtin(Value::Complex(60.0, 30.0)).expect("sind");
match result {
Value::Complex(re, im) => {
let (expected_re, expected_im) = sind_complex(60.0, 30.0);
assert!((re - expected_re).abs() < 1e-12);
assert!((im - expected_im).abs() < 1e-12);
}
other => panic!("expected complex result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sind_string_errors() {
let err = sind_builtin(Value::String("90".into())).expect_err("expected error");
assert!(error_message(err).contains("sind: expected numeric input"));
}
}