Files
llvm/mlir/lib/Bindings/Python/IRModule.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

268 lines
9.3 KiB
C++
Raw Normal View History

//===- IRModule.cpp - IR pybind module ------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
#include "IRModule.h"
#include <optional>
#include <vector>
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
#include "Globals.h"
#include "NanobindUtils.h"
#include "mlir-c/Bindings/Python/Interop.h"
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
#include "mlir-c/Support.h"
#include "mlir/Bindings/Python/Nanobind.h"
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
namespace nb = nanobind;
using namespace mlir;
using namespace mlir::python;
// -----------------------------------------------------------------------------
// PyGlobals
// -----------------------------------------------------------------------------
PyGlobals *PyGlobals::instance = nullptr;
PyGlobals::PyGlobals() {
assert(!instance && "PyGlobals already constructed");
instance = this;
// The default search path include {mlir.}dialects, where {mlir.} is the
// package prefix configured at compile time.
dialectSearchPrefixes.emplace_back(MAKE_MLIR_PYTHON_QUALNAME("dialects"));
}
PyGlobals::~PyGlobals() { instance = nullptr; }
bool PyGlobals::loadDialectModule(llvm::StringRef dialectNamespace) {
{
nb::ft_lock_guard lock(mutex);
if (loadedDialectModules.contains(dialectNamespace))
return true;
}
// Since re-entrancy is possible, make a copy of the search prefixes.
std::vector<std::string> localSearchPrefixes = dialectSearchPrefixes;
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object loaded = nb::none();
for (std::string moduleName : localSearchPrefixes) {
moduleName.push_back('.');
moduleName.append(dialectNamespace.data(), dialectNamespace.size());
try {
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
loaded = nb::module_::import_(moduleName.c_str());
} catch (nb::python_error &e) {
if (e.matches(PyExc_ModuleNotFoundError)) {
continue;
}
throw;
}
break;
}
if (loaded.is_none())
return false;
// Note: Iterator cannot be shared from prior to loading, since re-entrancy
// may have occurred, which may do anything.
nb::ft_lock_guard lock(mutex);
loadedDialectModules.insert(dialectNamespace);
return true;
}
void PyGlobals::registerAttributeBuilder(const std::string &attributeKind,
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::callable pyFunc, bool replace) {
nb::ft_lock_guard lock(mutex);
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object &found = attributeBuilderMap[attributeKind];
if (found && !replace) {
throw std::runtime_error((llvm::Twine("Attribute builder for '") +
[mlir][python bindings] generate all the enums This PR implements python enum bindings for *all* the enums - this includes `I*Attrs` (including positional/bit) and `Dialect/EnumAttr`. There are a few parts to this: 1. CMake: a small addition to `declare_mlir_dialect_python_bindings` and `declare_mlir_dialect_extension_python_bindings` to generate the enum, a boolean arg `GEN_ENUM_BINDINGS` to make it opt-in (even though it works for basically all of the dialects), and an optional `GEN_ENUM_BINDINGS_TD_FILE` for handling corner cases. 2. EnumPythonBindingGen.cpp: there are two weedy aspects here that took investigation: 1. If an enum attribute is not a `Dialect/EnumAttr` then the `EnumAttrInfo` record is canonical, as far as both the cases of the enum **and the `AttrDefName`**. On the otherhand, if an enum is a `Dialect/EnumAttr` then the `EnumAttr` record has the correct `AttrDefName` ("load bearing", i.e., populates `ods.ir.AttributeBuilder('<NAME>')`) but its `enum` field contains the cases, which is an instance of `EnumAttrInfo`. The solution is to generate an one enum class for both `Dialect/EnumAttr` and "independent" `EnumAttrInfo` but to make that class interopable with two builder registrations that both do the right thing (see next sub-bullet). 2. Because we don't have a good connection to cpp `EnumAttr`, i.e., only the `enum class` getters are exposed (like `DimensionAttr::get(Dimension value)`), we have to resort to parsing e.g., `Attribute.parse(f'#gpu<dim {x}>')`. This means that the set of supported `assemblyFormat`s (for the enum) is fixed at compile of MLIR (currently 2, the only 2 I saw). There might be some things that could be done here but they would require quite a bit more C API work to support generically (e.g., casting ints to enum cases and binding all the getters or going generically through the `symbolize*` methods, like `symbolizeDimension(uint32_t)` or `symbolizeDimension(StringRef)`). A few small changes: 1. In addition, since this patch registers default builders for attributes where people might've had their own builders already written, I added a `replace` param to `AttributeBuilder.insert` (`False` by default). 2. `makePythonEnumCaseName` can't handle all the different ways in which people write their enum cases, e.g., `llvm.CConv.Intel_OCL_BI`, which gets turned into `INTEL_O_C_L_B_I` (because `llvm::convertToSnakeFromCamelCase` doesn't look for runs of caps). So I dropped it. On the otherhand regularization does need to done because some enums have `None` as a case (and others might have other python keywords). 3. I turned on `llvm` dialect generation here in order to test `nvvm.WGMMAScaleIn`, which is an enum with [[ https://github.com/llvm/llvm-project/blob/d7e26b56207cbd8995296c5bb7c11ce676b649da/mlir/include/mlir/IR/EnumAttr.td#L22-L25 | no explicit discriminator ]] for the `neg` case. Note, dialects that didn't get a `GEN_ENUM_BINDINGS` don't have any enums to generate. Let me know if I should add more tests (the three trivial ones I added exercise both the supported `assemblyFormat`s and `replace=True`). Reviewed By: stellaraccident Differential Revision: https://reviews.llvm.org/D157934
2023-08-23 13:27:08 -05:00
attributeKind +
"' is already registered with func: " +
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::cast<std::string>(nb::str(found)))
.str());
}
found = std::move(pyFunc);
}
void PyGlobals::registerTypeCaster(MlirTypeID mlirTypeID,
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::callable typeCaster, bool replace) {
nb::ft_lock_guard lock(mutex);
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object &found = typeCasterMap[mlirTypeID];
if (found && !replace)
throw std::runtime_error("Type caster is already registered with caster: " +
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::cast<std::string>(nb::str(found)));
found = std::move(typeCaster);
}
2023-11-07 10:49:41 -06:00
void PyGlobals::registerValueCaster(MlirTypeID mlirTypeID,
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::callable valueCaster, bool replace) {
nb::ft_lock_guard lock(mutex);
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object &found = valueCasterMap[mlirTypeID];
2023-11-07 10:49:41 -06:00
if (found && !replace)
throw std::runtime_error("Value caster is already registered: " +
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::cast<std::string>(nb::repr(found)));
2023-11-07 10:49:41 -06:00
found = std::move(valueCaster);
}
void PyGlobals::registerDialectImpl(const std::string &dialectNamespace,
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object pyClass) {
nb::ft_lock_guard lock(mutex);
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object &found = dialectClassMap[dialectNamespace];
if (found) {
throw std::runtime_error((llvm::Twine("Dialect namespace '") +
dialectNamespace + "' is already registered.")
.str());
}
found = std::move(pyClass);
}
void PyGlobals::registerOperationImpl(const std::string &operationName,
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object pyClass, bool replace) {
nb::ft_lock_guard lock(mutex);
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
nb::object &found = operationClassMap[operationName];
[mlir][python] remove mixins (#68853) This PR replaces the mixin `OpView` extension mechanism with the standard inheritance mechanism. Why? Firstly, mixins are not very pythonic (inheritance is usually used for this), a little convoluted, and too "tight" (can only be used in the immediately adjacent `_ext.py`). Secondly, it (mixins) are now blocking are correct implementation of "value builders" (see [here](https://github.com/llvm/llvm-project/pull/68764)) where the problem becomes how to choose the correct base class that the value builder should call. This PR looks big/complicated but appearances are deceiving; 4 things were needed to make this work: 1. Drop `skipDefaultBuilders` in `OpPythonBindingGen::emitDefaultOpBuilders` 2. Former mixin extension classes are converted to inherit from the generated `OpView` instead of being "mixins" a. extension classes that simply were calling into an already generated `super().__init__` continue to do so b. (almost all) extension classes that were calling `self.build_generic` because of a lack of default builder being generated can now also just call `super().__init__` 3. To handle the [lone single use-case](https://sourcegraph.com/search?q=context%3Aglobal+select_opview_mixin&patternType=standard&sm=1&groupBy=repo) of `select_opview_mixin`, namely [linalg](https://github.com/llvm/llvm-project/blob/main/mlir/python/mlir/dialects/_linalg_ops_ext.py#L38), only a small change was necessary in `opdsl/lang/emitter.py` (thanks to the emission/generation of default builders/`__init__`s) 4. since the `extend_opview_class` decorator is removed, we need a way to register extension classes as the desired `OpView` that `op.opview` conjures into existence; so we do the standard thing and just enable replacing the existing registered `OpView` i.e., `register_operation(_Dialect, replace=True)`. Note, the upgrade path for the common case is to change an extension to inherit from the generated builder and decorate it with `register_operation(_Dialect, replace=True)`. In the slightly more complicated case where `super().__init(self.build_generic(...))` is called in the extension's `__init__`, this needs to be updated to call `__init__` in `OpView`, i.e., the grandparent (see updated docs). Note, also `<DIALECT>_ext.py` files/modules will no longer be automatically loaded. Note, the PR has 3 base commits that look funny but this was done for the purpose of tracking the line history of moving the `<DIALECT>_ops_ext.py` class into `<DIALECT>.py` and updating (commit labeled "fix").
2023-10-19 16:20:14 -05:00
if (found && !replace) {
throw std::runtime_error((llvm::Twine("Operation '") + operationName +
"' is already registered.")
.str());
}
found = std::move(pyClass);
}
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
std::optional<nb::callable>
PyGlobals::lookupAttributeBuilder(const std::string &attributeKind) {
nb::ft_lock_guard lock(mutex);
const auto foundIt = attributeBuilderMap.find(attributeKind);
if (foundIt != attributeBuilderMap.end()) {
assert(foundIt->second && "attribute builder is defined");
return foundIt->second;
}
return std::nullopt;
}
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
std::optional<nb::callable> PyGlobals::lookupTypeCaster(MlirTypeID mlirTypeID,
MlirDialect dialect) {
// Try to load dialect module.
(void)loadDialectModule(unwrap(mlirDialectGetNamespace(dialect)));
nb::ft_lock_guard lock(mutex);
const auto foundIt = typeCasterMap.find(mlirTypeID);
if (foundIt != typeCasterMap.end()) {
assert(foundIt->second && "type caster is defined");
return foundIt->second;
}
return std::nullopt;
}
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
std::optional<nb::callable> PyGlobals::lookupValueCaster(MlirTypeID mlirTypeID,
2023-11-07 10:49:41 -06:00
MlirDialect dialect) {
// Try to load dialect module.
(void)loadDialectModule(unwrap(mlirDialectGetNamespace(dialect)));
nb::ft_lock_guard lock(mutex);
2023-11-07 10:49:41 -06:00
const auto foundIt = valueCasterMap.find(mlirTypeID);
if (foundIt != valueCasterMap.end()) {
assert(foundIt->second && "value caster is defined");
return foundIt->second;
}
return std::nullopt;
}
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
std::optional<nb::object>
PyGlobals::lookupDialectClass(const std::string &dialectNamespace) {
// Make sure dialect module is loaded.
if (!loadDialectModule(dialectNamespace))
return std::nullopt;
nb::ft_lock_guard lock(mutex);
const auto foundIt = dialectClassMap.find(dialectNamespace);
if (foundIt != dialectClassMap.end()) {
assert(foundIt->second && "dialect class is defined");
return foundIt->second;
}
// Not found and loading did not yield a registration.
return std::nullopt;
}
[mlir python] Port Python core code to nanobind. (#120473) Relands #118583, with a fix for Python 3.8 compatibility. It was not possible to set the buffer protocol accessers via slots in Python 3.8. Why? https://nanobind.readthedocs.io/en/latest/why.html says it better than I can, but my primary motivation for this change is to improve MLIR IR construction time from JAX. For a complicated Google-internal LLM model in JAX, this change improves the MLIR lowering time by around 5s (out of around 30s), which is a significant speedup for simply switching binding frameworks. To a large extent, this is a mechanical change, for instance changing `pybind11::` to `nanobind::`. Notes: * this PR needs Nanobind 2.4.0, because it needs a bug fix (https://github.com/wjakob/nanobind/pull/806) that landed in that release. * this PR does not port the in-tree dialect extension modules. They can be ported in a future PR. * I removed the py::sibling() annotations from def_static and def_class in `PybindAdapters.h`. These ask pybind11 to try to form an overload with an existing method, but it's not possible to form mixed pybind11/nanobind overloads this ways and the parent class is now defined in nanobind. Better solutions may be possible here. * nanobind does not contain an exact equivalent of pybind11's buffer protocol support. It was not hard to add a nanobind implementation of a similar API. * nanobind is pickier about casting to std::vector<bool>, expecting that the input is a sequence of bool types, not truthy values. In a couple of places I added code to support truthy values during casting. * nanobind distinguishes bytes (`nb::bytes`) from strings (e.g., `std::string`). This required nb::bytes overloads in a few places.
2024-12-18 21:55:42 -05:00
std::optional<nb::object>
PyGlobals::lookupOperationClass(llvm::StringRef operationName) {
// Make sure dialect module is loaded.
auto split = operationName.split('.');
llvm::StringRef dialectNamespace = split.first;
if (!loadDialectModule(dialectNamespace))
return std::nullopt;
nb::ft_lock_guard lock(mutex);
auto foundIt = operationClassMap.find(operationName);
if (foundIt != operationClassMap.end()) {
assert(foundIt->second && "OpView is defined");
return foundIt->second;
}
// Not found and loading did not yield a registration.
return std::nullopt;
}
bool PyGlobals::TracebackLoc::locTracebacksEnabled() {
nanobind::ft_lock_guard lock(mutex);
return locTracebackEnabled_;
}
void PyGlobals::TracebackLoc::setLocTracebacksEnabled(bool value) {
nanobind::ft_lock_guard lock(mutex);
locTracebackEnabled_ = value;
}
size_t PyGlobals::TracebackLoc::locTracebackFramesLimit() {
nanobind::ft_lock_guard lock(mutex);
return locTracebackFramesLimit_;
}
void PyGlobals::TracebackLoc::setLocTracebackFramesLimit(size_t value) {
nanobind::ft_lock_guard lock(mutex);
locTracebackFramesLimit_ = std::min(value, kMaxFrames);
}
void PyGlobals::TracebackLoc::registerTracebackFileInclusion(
const std::string &file) {
nanobind::ft_lock_guard lock(mutex);
auto reg = "^" + llvm::Regex::escape(file);
if (userTracebackIncludeFiles.insert(reg).second)
rebuildUserTracebackIncludeRegex = true;
if (userTracebackExcludeFiles.count(reg)) {
if (userTracebackExcludeFiles.erase(reg))
rebuildUserTracebackExcludeRegex = true;
}
}
void PyGlobals::TracebackLoc::registerTracebackFileExclusion(
const std::string &file) {
nanobind::ft_lock_guard lock(mutex);
auto reg = "^" + llvm::Regex::escape(file);
if (userTracebackExcludeFiles.insert(reg).second)
rebuildUserTracebackExcludeRegex = true;
if (userTracebackIncludeFiles.count(reg)) {
if (userTracebackIncludeFiles.erase(reg))
rebuildUserTracebackIncludeRegex = true;
}
}
bool PyGlobals::TracebackLoc::isUserTracebackFilename(
const llvm::StringRef file) {
nanobind::ft_lock_guard lock(mutex);
if (rebuildUserTracebackIncludeRegex) {
userTracebackIncludeRegex.assign(
llvm::join(userTracebackIncludeFiles, "|"));
rebuildUserTracebackIncludeRegex = false;
isUserTracebackFilenameCache.clear();
}
if (rebuildUserTracebackExcludeRegex) {
userTracebackExcludeRegex.assign(
llvm::join(userTracebackExcludeFiles, "|"));
rebuildUserTracebackExcludeRegex = false;
isUserTracebackFilenameCache.clear();
}
if (!isUserTracebackFilenameCache.contains(file)) {
std::string fileStr = file.str();
bool include = std::regex_search(fileStr, userTracebackIncludeRegex);
bool exclude = std::regex_search(fileStr, userTracebackExcludeRegex);
isUserTracebackFilenameCache[file] = include || !exclude;
}
return isUserTracebackFilenameCache[file];
}