ramtej/nac3
1
0
Fork 0
forked from M-Labs/nac3
Code Pull Requests Activity

core/ndstrides: implement broadcasting & np_broadcast_to()

Browse Source
This commit is contained in:
lyken 2024-08-22 09:59:58 +08:00 committed by David Mak
parent 926e7e93a8
commit 9359ed9685
14 changed files with 363 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
nac3core/irrt/irrt.cpp
View File

@ -10,4 +10,5 @@
#include "irrt/ndarray/iter.hpp"
#include "irrt/ndarray/indexing.hpp"
#include "irrt/ndarray/array.hpp"
#include "irrt/ndarray/reshape.hpp"
#include "irrt/ndarray/reshape.hpp"
#include "irrt/ndarray/broadcast.hpp"

165
nac3core/irrt/irrt/ndarray/broadcast.hpp Normal file
View File

@ -0,0 +1,165 @@
#pragma once
#include "irrt/int_types.hpp"
#include "irrt/ndarray/def.hpp"
#include "irrt/slice.hpp"
namespace {
template<typename SizeT>
struct ShapeEntry {
SizeT ndims;
SizeT* shape;
};
} // namespace
namespace {
namespace ndarray {
namespace broadcast {
/**
* @brief Return true if `src_shape` can broadcast to `dst_shape`.
*
* See https://numpy.org/doc/stable/user/basics.broadcasting.html
*/
template<typename SizeT>
bool can_broadcast_shape_to(SizeT target_ndims, const SizeT* target_shape, SizeT src_ndims, const SizeT* src_shape) {
if (src_ndims > target_ndims) {
return false;
}
for (SizeT i = 0; i < src_ndims; i++) {
SizeT target_dim = target_shape[target_ndims - i - 1];
SizeT src_dim = src_shape[src_ndims - i - 1];
if (!(src_dim == 1 || target_dim == src_dim)) {
return false;
}
}
return true;
}
/**
* @brief Performs `np.broadcast_shapes(<shapes>)`
*
* @param num_shapes Number of entries in `shapes`
* @param shapes The list of shape to do `np.broadcast_shapes` on.
* @param dst_ndims The length of `dst_shape`.
* `dst_ndims` must be `max([shape.ndims for shape in shapes])`, but the caller has to calculate it/provide it.
* for this function since they should already know in order to allocate `dst_shape` in the first place.
* @param dst_shape The resulting shape. Must be pre-allocated by the caller. This function calculate the result
* of `np.broadcast_shapes` and write it here.
*/
template<typename SizeT>
void broadcast_shapes(SizeT num_shapes, const ShapeEntry<SizeT>* shapes, SizeT dst_ndims, SizeT* dst_shape) {
for (SizeT dst_axis = 0; dst_axis < dst_ndims; dst_axis++) {
dst_shape[dst_axis] = 1;
}
#ifdef IRRT_DEBUG_ASSERT
SizeT max_ndims_found = 0;
#endif
for (SizeT i = 0; i < num_shapes; i++) {
ShapeEntry<SizeT> entry = shapes[i];
// Check pre-condition: `dst_ndims` must be `max([shape.ndims for shape in shapes])`
debug_assert(SizeT, entry.ndims <= dst_ndims);
#ifdef IRRT_DEBUG_ASSERT
max_ndims_found = max(max_ndims_found, entry.ndims);
#endif
for (SizeT j = 0; j < entry.ndims; j++) {
SizeT entry_axis = entry.ndims - j - 1;
SizeT dst_axis = dst_ndims - j - 1;
SizeT entry_dim = entry.shape[entry_axis];
SizeT dst_dim = dst_shape[dst_axis];
if (dst_dim == 1) {
dst_shape[dst_axis] = entry_dim;
} else if (entry_dim == 1 || entry_dim == dst_dim) {
// Do nothing
} else {
raise_exception(SizeT, EXN_VALUE_ERROR,
"shape mismatch: objects cannot be broadcast "
"to a single shape.",
NO_PARAM, NO_PARAM, NO_PARAM);
}
}
}
// Check pre-condition: `dst_ndims` must be `max([shape.ndims for shape in shapes])`
debug_assert_eq(SizeT, max_ndims_found, dst_ndims);
}
/**
* @brief Perform `np.broadcast_to(<ndarray>, <target_shape>)` and appropriate assertions.
*
* This function attempts to broadcast `src_ndarray` to a new shape defined by `dst_ndarray.shape`,
* and return the result by modifying `dst_ndarray`.
*
* # Notes on `dst_ndarray`
* The caller is responsible for allocating space for the resulting ndarray.
* Here is what this function expects from `dst_ndarray` when called:
* - `dst_ndarray->data` does not have to be initialized.
* - `dst_ndarray->itemsize` does not have to be initialized.
* - `dst_ndarray->ndims` must be initialized, determining the length of `dst_ndarray->shape`
* - `dst_ndarray->shape` must be allocated, and must contain the desired target broadcast shape.
* - `dst_ndarray->strides` must be allocated, through it can contain uninitialized values.
* When this function call ends:
* - `dst_ndarray->data` is set to `src_ndarray->data` (`dst_ndarray` is just a view to `src_ndarray`)
* - `dst_ndarray->itemsize` is set to `src_ndarray->itemsize`
* - `dst_ndarray->ndims` is unchanged.
* - `dst_ndarray->shape` is unchanged.
* - `dst_ndarray->strides` is updated accordingly by how ndarray broadcast_to works.
*/
template<typename SizeT>
void broadcast_to(const NDArray<SizeT>* src_ndarray, NDArray<SizeT>* dst_ndarray) {
if (!ndarray::broadcast::can_broadcast_shape_to(dst_ndarray->ndims, dst_ndarray->shape, src_ndarray->ndims,
src_ndarray->shape)) {
raise_exception(SizeT, EXN_VALUE_ERROR, "operands could not be broadcast together", NO_PARAM, NO_PARAM,
NO_PARAM);
}
dst_ndarray->data = src_ndarray->data;
dst_ndarray->itemsize = src_ndarray->itemsize;
for (SizeT i = 0; i < dst_ndarray->ndims; i++) {
SizeT src_axis = src_ndarray->ndims - i - 1;
SizeT dst_axis = dst_ndarray->ndims - i - 1;
if (src_axis < 0 || (src_ndarray->shape[src_axis] == 1 && dst_ndarray->shape[dst_axis] != 1)) {
// Freeze the steps in-place
dst_ndarray->strides[dst_axis] = 0;
} else {
dst_ndarray->strides[dst_axis] = src_ndarray->strides[src_axis];
}
}
}
} // namespace broadcast
} // namespace ndarray
} // namespace
extern "C" {
using namespace ndarray::broadcast;
void __nac3_ndarray_broadcast_to(NDArray<int32_t>* src_ndarray, NDArray<int32_t>* dst_ndarray) {
broadcast_to(src_ndarray, dst_ndarray);
}
void __nac3_ndarray_broadcast_to64(NDArray<int64_t>* src_ndarray, NDArray<int64_t>* dst_ndarray) {
broadcast_to(src_ndarray, dst_ndarray);
}
void __nac3_ndarray_broadcast_shapes(int32_t num_shapes,
const ShapeEntry<int32_t>* shapes,
int32_t dst_ndims,
int32_t* dst_shape) {
broadcast_shapes(num_shapes, shapes, dst_ndims, dst_shape);
}
void __nac3_ndarray_broadcast_shapes64(int64_t num_shapes,
const ShapeEntry<int64_t>* shapes,
int64_t dst_ndims,
int64_t* dst_shape) {
broadcast_shapes(num_shapes, shapes, dst_ndims, dst_shape);
}
}

29
nac3core/src/codegen/irrt/mod.rs
View File

@ -21,7 +21,7 @@ use super::{
model::{function::FnCall, *},
object::{
list::List,
ndarray::{indexing::NDIndex, nditer::NDIter, NDArray},
ndarray::{broadcast::ShapeEntry, indexing::NDIndex, nditer::NDIter, NDArray},
},
stmt::gen_for_callback_incrementing,
CodeGenContext, CodeGenerator,
@ -1176,3 +1176,30 @@ pub fn call_nac3_ndarray_reshape_resolve_and_check_new_shape<'ctx, G: CodeGenera
);
FnCall::builder(generator, ctx, &name).arg(size).arg(new_ndims).arg(new_shape).returning_void();
}
pub fn call_nac3_ndarray_broadcast_to<'ctx, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
src_ndarray: Instance<'ctx, Ptr<Struct<NDArray>>>,
dst_ndarray: Instance<'ctx, Ptr<Struct<NDArray>>>,
) {
let name = get_sizet_dependent_function_name(generator, ctx, "__nac3_ndarray_broadcast_to");
FnCall::builder(generator, ctx, &name).arg(src_ndarray).arg(dst_ndarray).returning_void();
}
pub fn call_nac3_ndarray_broadcast_shapes<'ctx, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
num_shape_entries: Instance<'ctx, Int<SizeT>>,
shape_entries: Instance<'ctx, Ptr<Struct<ShapeEntry>>>,
dst_ndims: Instance<'ctx, Int<SizeT>>,
dst_shape: Instance<'ctx, Ptr<Int<SizeT>>>,
) {
let name = get_sizet_dependent_function_name(generator, ctx, "__nac3_ndarray_broadcast_shapes");
FnCall::builder(generator, ctx, &name)
.arg(num_shape_entries)
.arg(shape_entries)
.arg(dst_ndims)
.arg(dst_shape)
.returning_void();
}

138
nac3core/src/codegen/object/ndarray/broadcast.rs Normal file
View File

@ -0,0 +1,138 @@
use itertools::Itertools;
use super::NDArrayObject;
use crate::codegen::{
irrt::{call_nac3_ndarray_broadcast_shapes, call_nac3_ndarray_broadcast_to},
model::*,
CodeGenContext, CodeGenerator,
};
/// Fields of [`ShapeEntry`]
pub struct ShapeEntryFields<'ctx, F: FieldTraversal<'ctx>> {
pub ndims: F::Output<Int<SizeT>>,
pub shape: F::Output<Ptr<Int<SizeT>>>,
}
/// An IRRT structure used in broadcasting.
#[derive(Debug, Clone, Copy, Default)]
pub struct ShapeEntry;
impl<'ctx> StructKind<'ctx> for ShapeEntry {
type Fields<F: FieldTraversal<'ctx>> = ShapeEntryFields<'ctx, F>;
fn iter_fields<F: FieldTraversal<'ctx>>(&self, traversal: &mut F) -> Self::Fields<F> {
Self::Fields { ndims: traversal.add_auto("ndims"), shape: traversal.add_auto("shape") }
}
}
impl<'ctx> NDArrayObject<'ctx> {
/// Create a broadcast view on this ndarray with a target shape.
///
/// The input shape will be checked to make sure that it contains no negative values.
///
/// * `target_ndims` - The ndims type after broadcasting to the given shape.
/// The caller has to figure this out for this function.
/// * `target_shape` - An array pointer pointing to the target shape.
#[must_use]
pub fn broadcast_to<G: CodeGenerator + ?Sized>(
&self,
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
target_ndims: u64,
target_shape: Instance<'ctx, Ptr<Int<SizeT>>>,
) -> Self {
let broadcast_ndarray = NDArrayObject::alloca(generator, ctx, self.dtype, target_ndims);
broadcast_ndarray.copy_shape_from_array(generator, ctx, target_shape);
call_nac3_ndarray_broadcast_to(generator, ctx, self.instance, broadcast_ndarray.instance);
broadcast_ndarray
}
}
/// A result produced by [`broadcast_all_ndarrays`]
#[derive(Debug, Clone)]
pub struct BroadcastAllResult<'ctx> {
/// The statically known `ndims` of the broadcast result.
pub ndims: u64,
/// The broadcasting shape.
pub shape: Instance<'ctx, Ptr<Int<SizeT>>>,
/// Broadcasted views on the inputs.
///
/// All of them will have `shape` [`BroadcastAllResult::shape`] and
/// `ndims` [`BroadcastAllResult::ndims`]. The length of the vector
/// is the same as the input.
pub ndarrays: Vec<NDArrayObject<'ctx>>,
}
/// Helper function to call `call_nac3_ndarray_broadcast_shapes`
fn broadcast_shapes<'ctx, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
in_shape_entries: &[(Instance<'ctx, Ptr<Int<SizeT>>>, u64)], // (shape, shape's length/ndims)
broadcast_ndims: u64,
broadcast_shape: Instance<'ctx, Ptr<Int<SizeT>>>,
) {
// Prepare input shape entries to be passed to `call_nac3_ndarray_broadcast_shapes`.
let num_shape_entries = Int(SizeT).const_int(
generator,
ctx.ctx,
u64::try_from(in_shape_entries.len()).unwrap(),
false,
);
let shape_entries = Struct(ShapeEntry).array_alloca(generator, ctx, num_shape_entries.value);
for (i, (in_shape, in_ndims)) in in_shape_entries.iter().enumerate() {
let pshape_entry = shape_entries.offset_const(ctx, i64::try_from(i).unwrap());
let in_ndims = Int(SizeT).const_int(generator, ctx.ctx, *in_ndims, false);
pshape_entry.set(ctx, |f| f.ndims, in_ndims);
pshape_entry.set(ctx, |f| f.shape, *in_shape);
}
let broadcast_ndims = Int(SizeT).const_int(generator, ctx.ctx, broadcast_ndims, false);
call_nac3_ndarray_broadcast_shapes(
generator,
ctx,
num_shape_entries,
shape_entries,
broadcast_ndims,
broadcast_shape,
);
}
impl<'ctx> NDArrayObject<'ctx> {
/// Broadcast all ndarrays according to `np.broadcast()` and return a [`BroadcastAllResult`]
/// containing all the information of the result of the broadcast operation.
pub fn broadcast<G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
ndarrays: &[Self],
) -> BroadcastAllResult<'ctx> {
assert!(!ndarrays.is_empty());
// Infer the broadcast output ndims.
let broadcast_ndims_int = ndarrays.iter().map(|ndarray| ndarray.ndims).max().unwrap();
let broadcast_ndims = Int(SizeT).const_int(generator, ctx.ctx, broadcast_ndims_int, false);
let broadcast_shape = Int(SizeT).array_alloca(generator, ctx, broadcast_ndims.value);
let shape_entries = ndarrays
.iter()
.map(|ndarray| (ndarray.instance.get(generator, ctx, |f| f.shape), ndarray.ndims))
.collect_vec();
broadcast_shapes(generator, ctx, &shape_entries, broadcast_ndims_int, broadcast_shape);
// Broadcast all the inputs to shape `dst_shape`.
let broadcast_ndarrays: Vec<_> = ndarrays
.iter()
.map(|ndarray| {
ndarray.broadcast_to(generator, ctx, broadcast_ndims_int, broadcast_shape)
})
.collect_vec();
BroadcastAllResult {
ndims: broadcast_ndims_int,
shape: broadcast_shape,
ndarrays: broadcast_ndarrays,
}
}
}

1
nac3core/src/codegen/object/ndarray/mod.rs
View File

@ -22,6 +22,7 @@ use crate::{
};
pub mod array;
pub mod broadcast;
pub mod factory;
pub mod indexing;
pub mod nditer;

22
nac3core/src/toplevel/builtins.rs
View File

@ -523,7 +523,7 @@ impl<'a> BuiltinBuilder<'a> {
self.build_ndarray_property_getter_function(prim)
}
PrimDef::FunNpTranspose | PrimDef::FunNpReshape => {
PrimDef::FunNpBroadcastTo | PrimDef::FunNpTranspose | PrimDef::FunNpReshape => {
self.build_ndarray_view_function(prim)
}
@ -1471,7 +1471,10 @@ 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]);
debug_assert_prim_is_allowed(
prim,
&[PrimDef::FunNpBroadcastTo, PrimDef::FunNpTranspose, PrimDef::FunNpReshape],
);
let in_ndarray_ty = self.unifier.get_fresh_var_with_range(
&[self.primitives.ndarray],
@ -1499,7 +1502,10 @@ impl<'a> BuiltinBuilder<'a> {
// 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 => {
PrimDef::FunNpBroadcastTo | PrimDef::FunNpReshape => {
// These two functions have the same function signature.
// Mixed together for convenience.
let ret_ty = self.unifier.get_dummy_var().ty; // Handled by special holding
create_fn_by_codegen(
@ -1530,7 +1536,15 @@ impl<'a> BuiltinBuilder<'a> {
let (_, ndims) = unpack_ndarray_var_tys(&mut ctx.unifier, fun.0.ret);
let ndims = extract_ndims(&ctx.unifier, ndims);
let new_ndarray = ndarray.reshape_or_copy(generator, ctx, ndims, shape);
let new_ndarray = match prim {
PrimDef::FunNpBroadcastTo => {
ndarray.broadcast_to(generator, ctx, ndims, shape)
}
PrimDef::FunNpReshape => {
ndarray.reshape_or_copy(generator, ctx, ndims, shape)
}
_ => unreachable!(),
};
Ok(Some(new_ndarray.instance.value.as_basic_value_enum()))
}),
)

2
nac3core/src/toplevel/helper.rs
View File

@ -60,6 +60,7 @@ pub enum PrimDef {
FunNpStrides,
// NumPy ndarray view functions
FunNpBroadcastTo,
FunNpTranspose,
FunNpReshape,
@ -253,6 +254,7 @@ impl PrimDef {
PrimDef::FunNpStrides => fun("np_strides", None),
// NumPy NDArray view functions
PrimDef::FunNpBroadcastTo => fun("np_broadcast_to", None),
PrimDef::FunNpTranspose => fun("np_transpose", None),
PrimDef::FunNpReshape => fun("np_reshape", None),

2
nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__generic_class.snap
View File

@ -5,7 +5,7 @@ expression: res_vec
[
"Class {\nname: \"Generic_A\",\nancestors: [\"Generic_A[V]\", \"B\"],\nfields: [\"aa\", \"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"foo\", \"fn[[b:T], none]\"), (\"fun\", \"fn[[a:int32], V]\")],\ntype_vars: [\"V\"]\n}\n",
"Function {\nname: \"Generic_A.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"Generic_A.fun\",\nsig: \"fn[[a:int32], V]\",\nvar_id: [TypeVarId(248)]\n}\n",
"Function {\nname: \"Generic_A.fun\",\nsig: \"fn[[a:int32], V]\",\nvar_id: [TypeVarId(250)]\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B\"],\nfields: [\"aa\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"foo\", \"fn[[b:T], none]\")],\ntype_vars: []\n}\n",
"Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"B.foo\",\nsig: \"fn[[b:T], none]\",\nvar_id: []\n}\n",

2
nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__inheritance_override.snap
View File

@ -7,7 +7,7 @@ expression: res_vec
"Function {\nname: \"A.__init__\",\nsig: \"fn[[t:T], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"A.fun\",\nsig: \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\",\nvar_id: []\n}\n",
"Function {\nname: \"A.foo\",\nsig: \"fn[[c:C], none]\",\nvar_id: []\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B[typevar237]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: [\"typevar237\"]\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B[typevar239]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: [\"typevar239\"]\n}\n",
"Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"B.fun\",\nsig: \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\",\nvar_id: []\n}\n",
"Class {\nname: \"C\",\nancestors: [\"C\", \"B[bool]\", \"A[float]\"],\nfields: [\"a\", \"b\", \"c\", \"d\", \"e\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:int32, b:T], list[virtual[B[bool]]]]\"), (\"foo\", \"fn[[c:C], none]\")],\ntype_vars: []\n}\n",

4
nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__list_tuple_generic.snap
View File

@ -5,8 +5,8 @@ expression: res_vec
[
"Function {\nname: \"foo\",\nsig: \"fn[[a:list[int32], b:tuple[T, float]], A[B, bool]]\",\nvar_id: []\n}\n",
"Class {\nname: \"A\",\nancestors: [\"A[T, V]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[v:V], none]\"), (\"fun\", \"fn[[a:T], V]\")],\ntype_vars: [\"T\", \"V\"]\n}\n",
"Function {\nname: \"A.__init__\",\nsig: \"fn[[v:V], none]\",\nvar_id: [TypeVarId(250)]\n}\n",
"Function {\nname: \"A.fun\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(255)]\n}\n",
"Function {\nname: \"A.__init__\",\nsig: \"fn[[v:V], none]\",\nvar_id: [TypeVarId(252)]\n}\n",
"Function {\nname: \"A.fun\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(257)]\n}\n",
"Function {\nname: \"gfun\",\nsig: \"fn[[a:A[list[float], int32]], none]\",\nvar_id: []\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B\"],\nfields: [],\nmethods: [(\"__init__\", \"fn[[], none]\")],\ntype_vars: []\n}\n",
"Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",

2
nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__self1.snap
View File

@ -3,7 +3,7 @@ source: nac3core/src/toplevel/test.rs
expression: res_vec
---
[
"Class {\nname: \"A\",\nancestors: [\"A[typevar236, typevar237]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[a:A[float, bool], b:B], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\")],\ntype_vars: [\"typevar236\", \"typevar237\"]\n}\n",
"Class {\nname: \"A\",\nancestors: [\"A[typevar238, typevar239]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[a:A[float, bool], b:B], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\")],\ntype_vars: [\"typevar238\", \"typevar239\"]\n}\n",
"Function {\nname: \"A.__init__\",\nsig: \"fn[[a:A[float, bool], b:B], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"A.fun\",\nsig: \"fn[[a:A[float, bool]], A[bool, int32]]\",\nvar_id: []\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B\", \"A[int64, bool]\"],\nfields: [\"a\", \"b\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[a:A[float, bool]], A[bool, int32]]\"), (\"foo\", \"fn[[b:B], B]\"), (\"bar\", \"fn[[a:A[list[B], int32]], tuple[A[virtual[A[B, int32]], bool], B]]\")],\ntype_vars: []\n}\n",

4
nac3core/src/toplevel/snapshots/nac3core__toplevel__test__test_analyze__simple_class_compose.snap
View File

@ -6,12 +6,12 @@ expression: res_vec
"Class {\nname: \"A\",\nancestors: [\"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
"Function {\nname: \"A.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"A.fun\",\nsig: \"fn[[b:B], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"A.foo\",\nsig: \"fn[[a:T, b:V], none]\",\nvar_id: [TypeVarId(256)]\n}\n",
"Function {\nname: \"A.foo\",\nsig: \"fn[[a:T, b:V], none]\",\nvar_id: [TypeVarId(258)]\n}\n",
"Class {\nname: \"B\",\nancestors: [\"B\", \"C\", \"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
"Function {\nname: \"B.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Class {\nname: \"C\",\nancestors: [\"C\", \"A\"],\nfields: [\"a\"],\nmethods: [(\"__init__\", \"fn[[], none]\"), (\"fun\", \"fn[[b:B], none]\"), (\"foo\", \"fn[[a:T, b:V], none]\")],\ntype_vars: []\n}\n",
"Function {\nname: \"C.__init__\",\nsig: \"fn[[], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"C.fun\",\nsig: \"fn[[b:B], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"foo\",\nsig: \"fn[[a:A], none]\",\nvar_id: []\n}\n",
"Function {\nname: \"ff\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(264)]\n}\n",
"Function {\nname: \"ff\",\nsig: \"fn[[a:T], V]\",\nvar_id: [TypeVarId(266)]\n}\n",
]

2
nac3core/src/typecheck/type_inferencer/mod.rs
View File

@ -1595,7 +1595,7 @@ impl<'a> Inferencer<'a> {
}));
}
// 2-argument ndarray n-dimensional factory functions
if id == &"np_reshape".into() && args.len() == 2 {
if ["np_reshape".into(), "np_broadcast_to".into()].contains(id) && args.len() == 2 {
let arg0 = self.fold_expr(args.remove(0))?;
let shape_expr = args.remove(0);

1
nac3standalone/demo/interpret_demo.py
View File

@ -180,6 +180,7 @@ def patch(module):
module.np_array = np.array
# NumPy NDArray view functions
module.np_broadcast_to = np.broadcast_to
module.np_transpose = np.transpose
module.np_reshape = np.reshape