DLManagedTensor lifetime management implemented per array-API specs

1. The pycapsule destructor only calls DLManagedTensor.deleter is the name is "dltensor"
2. Code consuming the DLPack capsule renamed the capsule (to avoid destructor calling
   the deleter) and instead creates an internal object to do that and uses
   that internal object as the base of _Memory object

`from_dlpack_capsule` function should handle NULL data field
For zero-elements arrays in DLPack, allocate 1 element

Proper support for strides added.

Expanded docstring of `dpctl.tensor.from_dlpack`

Author: oleksandr-pavlyk

dpctl/tensor/_dlpack.pyx

@@ -36,6 +36,7 @@ from ._usmarray cimport usm_ndarray
 import numpy as np
 
 import dpctl
+import dpctl.memory as dpmem
 
 
 cdef extern from './include/dlpack/dlpack.h' nogil:
@@ -95,10 +96,6 @@ cdef void pycapsule_deleter(object dlt_capsule):
         dlm_tensor = <DLManagedTensor*>cpython.PyCapsule_GetPointer(
             dlt_capsule, 'dltensor')
         dlm_tensor.deleter(dlm_tensor)
-    elif cpython.PyCapsule_IsValid(dlt_capsule, 'used_dltensor'):
-        dlm_tensor = <DLManagedTensor*>cpython.PyCapsule_GetPointer(
-            dlt_capsule, 'used_dltensor')
-        dlm_tensor.deleter(dlm_tensor)
 
 
 cdef void managed_tensor_deleter(DLManagedTensor *dlm_tensor) with gil:
@@ -133,7 +130,11 @@ cpdef to_dlpack_capsule(usm_ndarray usm_ary) except+:
     cdef int64_t *shape_strides_ptr = NULL
     cdef int i = 0
     cdef int device_id = -1
+    cdef char* base_ptr = NULL
+    cdef Py_ssize_t element_offset = 0
+    cdef Py_ssize_t byte_offset = 0
 
+    ary_base = usm_ary.get_base()
     ary_sycl_queue = usm_ary.get_sycl_queue()
     ary_sycl_device = ary_sycl_queue.get_sycl_device()
 
@@ -176,11 +177,13 @@ cpdef to_dlpack_capsule(usm_ndarray usm_ary) except+:
 
     ary_dt = usm_ary.dtype
     ary_dtk = ary_dt.kind
+    element_offset = usm_ary.get_offset()
+    byte_offset = element_offset * (<Py_ssize_t>ary_dt.itemsize)
 
     dl_tensor = &dlm_tensor.dl_tensor
-    dl_tensor.data = <void*>data_ptr
+    dl_tensor.data = <void*>(data_ptr - byte_offset)
     dl_tensor.ndim = nd
-    dl_tensor.byte_offset = <uint64_t>0
+    dl_tensor.byte_offset = <uint64_t>byte_offset
     dl_tensor.shape = &shape_strides_ptr[0]
     if strides_ptr is NULL:
         dl_tensor.strides = NULL
@@ -212,6 +215,24 @@ cpdef to_dlpack_capsule(usm_ndarray usm_ary) except+:
     return cpython.PyCapsule_New(dlm_tensor, 'dltensor', pycapsule_deleter)
 
 
+cdef class _DLManagedTensorOwner:
+    """Helper class managing lifetimes of the DLManagedTensor struct"""
+    cdef DLManagedTensor *dlm_tensor
+
+    def __cinit__(self):
+        self.dlm_tensor = NULL
+
+    def __dealloc__(self):
+        if self.dlm_tensor:
+            self.dlm_tensor.deleter(self.dlm_tensor)
+
+    @staticmethod
+    cdef _DLManagedTensorOwner _create(DLManagedTensor *dlm_tensor_src):
+        cdef _DLManagedTensorOwner res = _DLManagedTensorOwner.__new__(_DLManagedTensorOwner)
+        res.dlm_tensor = dlm_tensor_src
+        return res
+
+
 cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
     """Reconstructs instance of usm_ndarray from named Python
     capsule object referencing instance of `DLManagedTensor` without
@@ -221,6 +242,11 @@ cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
     cdef size_t sz = 1
     cdef int i
     cdef int element_bytesize = 0
+    cdef Py_ssize_t offset_min = 0
+    cdef Py_ssize_t offset_max = 0
+    cdef int64_t stride_i
+    cdef char* mem_ptr = NULL
+    cdef Py_ssize_t element_offset = 0
 
     if not cpython.PyCapsule_IsValid(py_caps, 'dltensor'):
         if cpython.PyCapsule_IsValid(py_caps, 'used_dltensor'):
@@ -237,9 +263,12 @@ cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
     # Verify that we can work with this device
     if dlm_tensor.dl_tensor.device.device_type == kDLOneAPI:
         q = dpctl.SyclQueue(str(<int>dlm_tensor.dl_tensor.device.device_id))
-        usm_type = c_dpmem._Memory.get_pointer_type(
-            <DPCTLSyclUSMRef> dlm_tensor.dl_tensor.data,
-            <c_dpctl.SyclContext>q.sycl_context)
+        if dlm_tensor.dl_tensor.data is NULL:
+            usm_type = b"device"
+        else:
+            usm_type = c_dpmem._Memory.get_pointer_type(
+                <DPCTLSyclUSMRef> dlm_tensor.dl_tensor.data,
+                <c_dpctl.SyclContext>q.sycl_context)
         if usm_type == b"unknown":
             raise ValueError(
                 f"Data pointer in DLPack is not bound to default sycl "
@@ -255,17 +284,45 @@ cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
             raise ValueError(
                 "Can not import DLPack tensor with lanes != 1"
             )
-        for i in range(dlm_tensor.dl_tensor.ndim):
-            sz = sz * dlm_tensor.dl_tensor.shape[i]
+        if dlm_tensor.dl_tensor.strides is NULL:
+            for i in range(dlm_tensor.dl_tensor.ndim):
+                sz = sz * dlm_tensor.dl_tensor.shape[i]
+        else:
+            offset_min = 0
+            offset_max = 0
+            for i in range(dlm_tensor.dl_tensor.ndim):
+                stride_i = dlm_tensor.dl_tensor.strides[i]
+                if stride_i > 0:
+                    offset_max = offset_max + stride_i * (
+                        dlm_tensor.dl_tensor.shape[i] - 1
+                    )
+                else:
+                    offset_min = offset_min + stride_i * (
+                        dlm_tensor.dl_tensor.shape[i] - 1
+                    )
+            sz = offset_max - offset_min + 1
+        if sz == 0:
+            sz = 1
 
         element_bytesize = (dlm_tensor.dl_tensor.dtype.bits // 8)
         sz = sz * element_bytesize
-        usm_mem = c_dpmem._Memory.create_from_usm_pointer_size_qref(
-            <DPCTLSyclUSMRef> dlm_tensor.dl_tensor.data,
-            sz,
-            (<c_dpctl.SyclQueue>q).get_queue_ref(),
-            memory_owner=py_caps
-        )
+        element_offset = dlm_tensor.dl_tensor.byte_offset // element_bytesize
+
+        # transfer dlm_tensor ownership
+        dlm_holder = _DLManagedTensorOwner._create(dlm_tensor)
+        cpython.PyCapsule_SetName(py_caps, 'used_dltensor')
+
+        if dlm_tensor.dl_tensor.data is NULL:
+            usm_mem = dpmem.MemoryUSMDevice(sz, q)
+        else:
+            mem_ptr = <char *>dlm_tensor.dl_tensor.data + dlm_tensor.dl_tensor.byte_offset
+            mem_ptr = mem_ptr - (element_offset * element_bytesize)
+            usm_mem = c_dpmem._Memory.create_from_usm_pointer_size_qref(
+                <DPCTLSyclUSMRef> mem_ptr,
+                sz,
+                (<c_dpctl.SyclQueue>q).get_queue_ref(),
+                memory_owner=dlm_holder
+            )
         py_shape = list()
         for i in range(dlm_tensor.dl_tensor.ndim):
             py_shape.append(dlm_tensor.dl_tensor.shape[i])
@@ -293,9 +350,9 @@ cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
             py_shape,
             dtype=ary_dt,
             buffer=usm_mem,
-            strides=py_strides
+            strides=py_strides,
+            offset=element_offset
         )
-        cpython.PyCapsule_SetName(py_caps, 'used_dltensor')
         return res_ary
     else:
         raise ValueError(
@@ -304,8 +361,16 @@ cpdef usm_ndarray from_dlpack_capsule(object py_caps) except +:
 
 
 cpdef from_dlpack(array):
-    """Constructs `usm_ndarray` from a Python object that implements
-    `__dlpack__` protocol.
+    """dpctl.tensor.from_dlpack(obj)
+
+    Constructs :class:`dpctl.tensor.usm_ndarray` instance from a Python
+    object `obj` that implements `__dlpack__` protocol. The output
+    array is always a zero-copy view of the input.
+
+    Raises:
+        TypeError: if `obj` does not implement `__dlpack__` method.
+        ValueError: if zero copy view can not be constructed because
+            the input array resides on an unsupported device.
     """
     if not hasattr(array, "__dlpack__"):
         raise TypeError(