fix(cuda.core): break _stream/_device <-> graph._graph_builder import cycle

cimport-ing GraphBuilder at the top of _stream.pyx and _device.pyx made
Cython emit a Python-level import of cuda.core.graph._graph_builder
during _stream module init. That triggered the chain
graph -> _graph_node -> _kernel_arg_handler -> _memory._buffer
-> _device, which then re-entered the still-initializing _stream module
via "from cuda.core._stream import IsStreamT", failing with
ImportError: cannot import name IsStreamT.

Restore the original lazy "import GraphBuilder" inside
create_graph_builder (Stream and Device) and Stream_accept. The return
annotations stay as bare names; "from __future__ import annotations" in
both files defers their evaluation, so they need not resolve at
function-definition time.

Made-with: Cursor

Author: Andy-Jost

cuda_core/cuda/core/_device.pyx

@@ -14,7 +14,6 @@ import threading
 from cuda.core._context cimport Context
 from cuda.core._context import ContextOptions
 from cuda.core._event cimport Event as cyEvent
-from cuda.core.graph._graph_builder cimport GraphBuilder
 from cuda.core._event import Event, EventOptions
 from cuda.core._memory._buffer cimport Buffer, MemoryResource
 from cuda.core._resource_handles cimport (
@@ -1371,6 +1370,8 @@ class Device:
             Newly created graph builder object.
 
         """
+        from cuda.core.graph._graph_builder import GraphBuilder
+
         self._check_context_initialized()
         return GraphBuilder._init(self.create_stream())
 

cuda_core/cuda/core/_stream.pyx

@@ -10,7 +10,6 @@ from libc.stdlib cimport strtol, getenv
 from cuda.bindings cimport cydriver
 
 from cuda.core._event cimport Event as cyEvent
-from cuda.core.graph._graph_builder cimport GraphBuilder
 from cuda.core._utils.cuda_utils cimport (
     check_or_create_options,
     HANDLE_RETURN,
@@ -372,6 +371,8 @@ cdef class Stream:
             Newly created graph builder object.
 
         """
+        from cuda.core.graph._graph_builder import GraphBuilder
+
         return GraphBuilder._init(self)
 
 
@@ -473,6 +474,8 @@ cdef cydriver.CUstream _handle_from_stream_protocol(obj) except*:
 # Helper for API functions that accept either Stream or GraphBuilder. Performs
 # needed checks and returns the relevant stream.
 cdef Stream Stream_accept(arg, bint allow_stream_protocol=False):
+    from cuda.core.graph._graph_builder import GraphBuilder
+
     if isinstance(arg, Stream):
         return <Stream>(arg)
     elif isinstance(arg, GraphBuilder):