forked from M-Labs/nac3
1
0
Fork 0

core: categorize np_{transpose,reshape} as 'view functions'

This commit is contained in:
lyken 2024-08-20 16:17:52 +08:00
parent ee66cac8c2
commit d37d0c8da4
No known key found for this signature in database
GPG Key ID: 3BD5FC6AC8325DD8
3 changed files with 67 additions and 61 deletions

View File

@ -517,6 +517,10 @@ impl<'a> BuiltinBuilder<'a> {
self.build_ndarray_property_getter_function(prim)
}
PrimDef::FunNpTranspose | PrimDef::FunNpReshape => {
self.build_ndarray_view_function(prim)
}
PrimDef::FunStr => self.build_str_function(),
PrimDef::FunFloor | PrimDef::FunFloor64 | PrimDef::FunCeil | PrimDef::FunCeil64 => {
@ -582,10 +586,6 @@ impl<'a> BuiltinBuilder<'a> {
| PrimDef::FunNpHypot
| PrimDef::FunNpNextAfter => self.build_np_2ary_function(prim),
PrimDef::FunNpTranspose | PrimDef::FunNpReshape => {
self.build_np_sp_ndarray_function(prim)
}
PrimDef::FunNpDot
| PrimDef::FunNpLinalgCholesky
| PrimDef::FunNpLinalgQr
@ -1463,6 +1463,57 @@ impl<'a> BuiltinBuilder<'a> {
}
}
/// Build np/sp functions that take as input `NDArray` only
fn build_ndarray_view_function(&mut self, prim: PrimDef) -> TopLevelDef {
debug_assert_prim_is_allowed(prim, &[PrimDef::FunNpTranspose, PrimDef::FunNpReshape]);
match prim {
PrimDef::FunNpTranspose => {
let ndarray_ty = self.unifier.get_fresh_var_with_range(
&[self.ndarray_num_ty],
Some("T".into()),
None,
);
create_fn_by_codegen(
self.unifier,
&into_var_map([ndarray_ty]),
prim.name(),
self.ndarray_num_ty,
&[(self.ndarray_num_ty, "x")],
Box::new(move |ctx, _, fun, args, generator| {
let arg_ty = fun.0.args[0].ty;
let arg_val =
args[0].1.clone().to_basic_value_enum(ctx, generator, arg_ty)?;
Ok(Some(ndarray_transpose(generator, ctx, (arg_ty, arg_val))?))
}),
)
}
// NOTE: on `ndarray_factory_fn_shape_arg_tvar` and
// the `param_ty` for `create_fn_by_codegen`.
//
// Similar to `build_ndarray_from_shape_factory_function` we delegate the responsibility of typechecking
// to [`typecheck::type_inferencer::Inferencer::fold_numpy_function_call_shape_argument`],
// and use a dummy [`TypeVar`] `ndarray_factory_fn_shape_arg_tvar` as a placeholder for `param_ty`.
PrimDef::FunNpReshape => create_fn_by_codegen(
self.unifier,
&VarMap::new(),
prim.name(),
self.ndarray_num_ty,
&[(self.ndarray_num_ty, "x"), (self.ndarray_factory_fn_shape_arg_tvar.ty, "shape")],
Box::new(move |ctx, _, fun, args, generator| {
let x1_ty = fun.0.args[0].ty;
let x1_val = args[0].1.clone().to_basic_value_enum(ctx, generator, x1_ty)?;
let x2_ty = fun.0.args[1].ty;
let x2_val = args[1].1.clone().to_basic_value_enum(ctx, generator, x2_ty)?;
Ok(Some(ndarray_reshape(generator, ctx, (x1_ty, x1_val), (x2_ty, x2_val))?))
}),
),
_ => unreachable!(),
}
}
/// Build the `str()` function.
fn build_str_function(&mut self) -> TopLevelDef {
let prim = PrimDef::FunStr;
@ -1950,57 +2001,6 @@ impl<'a> BuiltinBuilder<'a> {
}
}
/// Build np/sp functions that take as input `NDArray` only
fn build_np_sp_ndarray_function(&mut self, prim: PrimDef) -> TopLevelDef {
debug_assert_prim_is_allowed(prim, &[PrimDef::FunNpTranspose, PrimDef::FunNpReshape]);
match prim {
PrimDef::FunNpTranspose => {
let ndarray_ty = self.unifier.get_fresh_var_with_range(
&[self.ndarray_num_ty],
Some("T".into()),
None,
);
create_fn_by_codegen(
self.unifier,
&into_var_map([ndarray_ty]),
prim.name(),
self.ndarray_num_ty,
&[(self.ndarray_num_ty, "x")],
Box::new(move |ctx, _, fun, args, generator| {
let arg_ty = fun.0.args[0].ty;
let arg_val =
args[0].1.clone().to_basic_value_enum(ctx, generator, arg_ty)?;
Ok(Some(ndarray_transpose(generator, ctx, (arg_ty, arg_val))?))
}),
)
}
// NOTE: on `ndarray_factory_fn_shape_arg_tvar` and
// the `param_ty` for `create_fn_by_codegen`.
//
// Similar to `build_ndarray_from_shape_factory_function` we delegate the responsibility of typechecking
// to [`typecheck::type_inferencer::Inferencer::fold_numpy_function_call_shape_argument`],
// and use a dummy [`TypeVar`] `ndarray_factory_fn_shape_arg_tvar` as a placeholder for `param_ty`.
PrimDef::FunNpReshape => create_fn_by_codegen(
self.unifier,
&VarMap::new(),
prim.name(),
self.ndarray_num_ty,
&[(self.ndarray_num_ty, "x"), (self.ndarray_factory_fn_shape_arg_tvar.ty, "shape")],
Box::new(move |ctx, _, fun, args, generator| {
let x1_ty = fun.0.args[0].ty;
let x1_val = args[0].1.clone().to_basic_value_enum(ctx, generator, x1_ty)?;
let x2_ty = fun.0.args[1].ty;
let x2_val = args[1].1.clone().to_basic_value_enum(ctx, generator, x2_ty)?;
Ok(Some(ndarray_reshape(generator, ctx, (x1_ty, x1_val), (x2_ty, x2_val))?))
}),
),
_ => unreachable!(),
}
}
/// Build `np_linalg` and `sp_linalg` functions
///
/// The input to these functions must be floating point `NDArray`

View File

@ -57,6 +57,10 @@ pub enum PrimDef {
FunNpShape,
FunNpStrides,
// NumPy ndarray view functions
FunNpTranspose,
FunNpReshape,
// Miscellaneous NumPy & SciPy functions
FunNpRound,
FunNpFloor,
@ -104,8 +108,6 @@ pub enum PrimDef {
FunNpLdExp,
FunNpHypot,
FunNpNextAfter,
FunNpTranspose,
FunNpReshape,
// Linalg functions
FunNpDot,
@ -248,6 +250,10 @@ impl PrimDef {
PrimDef::FunNpShape => fun("np_shape", None),
PrimDef::FunNpStrides => fun("np_strides", None),
// NumPy NDArray view functions
PrimDef::FunNpTranspose => fun("np_transpose", None),
PrimDef::FunNpReshape => fun("np_reshape", None),
// Miscellaneous NumPy & SciPy functions
PrimDef::FunNpRound => fun("np_round", None),
PrimDef::FunNpFloor => fun("np_floor", None),
@ -295,8 +301,6 @@ impl PrimDef {
PrimDef::FunNpLdExp => fun("np_ldexp", None),
PrimDef::FunNpHypot => fun("np_hypot", None),
PrimDef::FunNpNextAfter => fun("np_nextafter", None),
PrimDef::FunNpTranspose => fun("np_transpose", None),
PrimDef::FunNpReshape => fun("np_reshape", None),
// Linalg functions
PrimDef::FunNpDot => fun("np_dot", None),

View File

@ -179,6 +179,10 @@ def patch(module):
module.np_identity = np.identity
module.np_array = np.array
# NumPy NDArray view functions
module.np_transpose = np.transpose
module.np_reshape = np.reshape
# NumPy NDArray property getters
module.np_size = np.size
module.np_shape = np.shape
@ -223,8 +227,6 @@ def patch(module):
module.np_ldexp = np.ldexp
module.np_hypot = np.hypot
module.np_nextafter = np.nextafter
module.np_transpose = np.transpose
module.np_reshape = np.reshape
# SciPy Math functions
module.sp_spec_erf = special.erf