Inline Cython exception handler (rapidsai#13411)

I originally placed the exception handler into a separate C++ header file that could be included by the Cython header because I figured that reading C++ inlined in Cython would be more confusing to devs. Unfortunately, the current approach complicates the build system due to the need to ensure that the directory containing the C++ header is always in the include path, which becomes problematic depending on where the files including the exception handler are (anywhere outside of `_lib` becomes problematic). Inlining is the simplest solution to this problem.

Authors:
  - Vyas Ramasubramani (https://github.com/vyasr)

Approvers:
  - Ashwin Srinath (https://github.com/shwina)
  - Bradley Dice (https://github.com/bdice)

URL: rapidsai#13411

python/cudf/cudf/_lib/CMakeLists.txt

@@ -63,11 +63,6 @@ rapids_cython_create_modules(
   LINKED_LIBRARIES "${linked_libraries}" ASSOCIATED_TARGETS cudf
 )
 
-# All modules need to include the header containing the exception handler.
-foreach(target IN LISTS RAPIDS_CYTHON_CREATED_TARGETS)
-  target_include_directories(${target} PRIVATE ${CMAKE_CURRENT_LIST_DIR})
-endforeach()
-
 target_link_libraries(strings_udf cudf_strings_udf)
 
 # TODO: Finding NumPy currently requires finding Development due to a bug in CMake. This bug was

python/cudf/cudf/_lib/exception_handler.pxd

@@ -1,5 +1,69 @@
 # Copyright (c) 2023, NVIDIA CORPORATION.
 
 
-cdef extern from "exception_handler.hpp" namespace "cudf_python::exceptions":
+# See
+# https://github.com/cython/cython/blob/master/Cython/Utility/CppSupport.cpp
+# for the original Cython exception handler.
+cdef extern from *:
+    """
+    #include <Python.h>
+    #include <cudf/utilities/error.hpp>
+    #include <ios>
+    #include <stdexcept>
+
+    namespace {
+
+    /**
+     * @brief Exception handler to map C++ exceptions to Python ones in Cython
+     *
+     * This exception handler extends the base exception handler provided by
+     * Cython. In addition to the exceptions that Cython itself supports, this
+     * file adds support for additional exceptions thrown by libcudf that need
+     * to be mapped to specific Python exceptions.
+     *
+     * Since this function interoperates with Python's exception state, it
+     * does not throw any C++ exceptions.
+     */
+    void cudf_exception_handler()
+    {
+      // Catch a handful of different errors here and turn them into the
+      // equivalent Python errors.
+      try {
+        if (PyErr_Occurred())
+          ;  // let latest Python exn pass through and ignore the current one
+        throw;
+      } catch (const std::bad_alloc& exn) {
+        PyErr_SetString(PyExc_MemoryError, exn.what());
+      } catch (const std::bad_cast& exn) {
+        PyErr_SetString(PyExc_TypeError, exn.what());
+      } catch (const std::domain_error& exn) {
+        PyErr_SetString(PyExc_ValueError, exn.what());
+      } catch (const cudf::data_type_error& exn) {
+        // Catch subclass (data_type_error) before parent (invalid_argument)
+        PyErr_SetString(PyExc_TypeError, exn.what());
+      } catch (const std::invalid_argument& exn) {
+        PyErr_SetString(PyExc_ValueError, exn.what());
+      } catch (const std::ios_base::failure& exn) {
+        // Unfortunately, in standard C++ we have no way of distinguishing EOF
+        // from other errors here; be careful with the exception mask
+        PyErr_SetString(PyExc_IOError, exn.what());
+      } catch (const std::out_of_range& exn) {
+        // Change out_of_range to IndexError
+        PyErr_SetString(PyExc_IndexError, exn.what());
+      } catch (const std::overflow_error& exn) {
+        PyErr_SetString(PyExc_OverflowError, exn.what());
+      } catch (const std::range_error& exn) {
+        PyErr_SetString(PyExc_ArithmeticError, exn.what());
+      } catch (const std::underflow_error& exn) {
+        PyErr_SetString(PyExc_ArithmeticError, exn.what());
+        // The below is the default catch-all case.
+      } catch (const std::exception& exn) {
+        PyErr_SetString(PyExc_RuntimeError, exn.what());
+      } catch (...) {
+        PyErr_SetString(PyExc_RuntimeError, "Unknown exception");
+      }
+    }
+
+    }  // anonymous namespace
+    """
     cdef void cudf_exception_handler()

python/cudf_kafka/pyproject.toml

@@ -21,7 +21,7 @@ authors = [
 license = { text = "Apache 2.0" }
 requires-python = ">=3.9"
 dependencies = [
-    "cudf==23.6.*",
+    "cudf==23.8.*",
 ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`.
 
 [project.optional-dependencies]

python/custreamz/pyproject.toml

@@ -19,8 +19,8 @@ license = { text = "Apache 2.0" }
 requires-python = ">=3.9"
 dependencies = [
     "confluent-kafka>=1.9.0,<1.10.0a0",
-    "cudf==23.6.*",
-    "cudf_kafka==23.6.*",
+    "cudf==23.8.*",
+    "cudf_kafka==23.8.*",
     "streamz",
 ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`.
 classifiers = [

python/dask_cudf/pyproject.toml

@@ -39,7 +39,7 @@ dynamic = ["entry-points"]
 
 [project.optional-dependencies]
 test = [
-    "dask-cuda==23.6.*",
+    "dask-cuda==23.8.*",
     "numba>=0.57",
     "pytest",
     "pytest-cov",