Implement a part of multidict python helpers in C (#1096)

Author: asvetlov

.github/workflows/ci-cd.yml

@@ -365,7 +365,7 @@ jobs:
     - build-wheels-for-tested-arches
     - pre-setup  # transitive, for accessing settings
     runs-on: ubuntu-latest
-    timeout-minutes: 5
+    timeout-minutes: 10
     steps:
     - name: Checkout project
       uses: actions/checkout@v4

CHANGES/1096.misc

@@ -0,0 +1,3 @@
+Implemented views comparison and disjoints in C instead of Python helpers.
+
+The performance boost is about 40%.

multidict/_multidict.c

@@ -326,20 +326,20 @@ _multidict_extend(MultiDictObject *self, PyObject *args, PyObject *kwds,
 }
 
 static inline PyObject *
-_multidict_copy(MultiDictObject *self, PyTypeObject *multidict_tp_object)
+multidict_copy(MultiDictObject *self)
 {
     MultiDictObject *new_multidict = NULL;
 
     PyObject *arg_items = NULL,
              *items     = NULL;
 
     new_multidict = (MultiDictObject*)PyType_GenericNew(
-        multidict_tp_object, NULL, NULL);
+        Py_TYPE(self), NULL, NULL);
     if (new_multidict == NULL) {
         return NULL;
     }
 
-    if (multidict_tp_object->tp_init(
+    if (Py_TYPE(self)->tp_init(
         (PyObject*)new_multidict, NULL, NULL) < 0)
     {
         return NULL;
@@ -782,12 +782,6 @@ multidict_add(
     Py_RETURN_NONE;
 }
 
-static inline PyObject *
-multidict_copy(MultiDictObject *self)
-{
-    return _multidict_copy(self, &multidict_type);
-}
-
 static inline PyObject *
 multidict_extend(MultiDictObject *self, PyObject *args, PyObject *kwds)
 {
@@ -1163,27 +1157,6 @@ cimultidict_tp_init(MultiDictObject *self, PyObject *args, PyObject *kwds)
     return 0;
 }
 
-static inline PyObject *
-cimultidict_copy(MultiDictObject *self)
-{
-    return _multidict_copy(self, &cimultidict_type);
-}
-
-PyDoc_STRVAR(cimultidict_copy_doc,
-"Return a copy of itself.");
-
-static PyMethodDef cimultidict_methods[] = {
-    {
-        "copy",
-        (PyCFunction)cimultidict_copy,
-        METH_NOARGS,
-        cimultidict_copy_doc
-    },
-    {
-        NULL,
-        NULL
-    }   /* sentinel */
-};
 
 PyDoc_STRVAR(CIMultDict_doc,
 "Dictionary with the support for duplicate case-insensitive keys.");
@@ -1199,7 +1172,6 @@ static PyTypeObject cimultidict_type = {
     .tp_traverse = (traverseproc)multidict_tp_traverse,
     .tp_clear = (inquiry)multidict_tp_clear,
     .tp_weaklistoffset = offsetof(MultiDictObject, weaklist),
-    .tp_methods = cimultidict_methods,
     .tp_base = &multidict_type,
     .tp_init = (initproc)cimultidict_tp_init,
     .tp_alloc = PyType_GenericAlloc,
@@ -1607,18 +1579,15 @@ static inline void
 module_free(void *m)
 {
     Py_CLEAR(multidict_str_lower);
+    Py_CLEAR(viewbaseset_and_func);
+    Py_CLEAR(viewbaseset_or_func);
+    Py_CLEAR(viewbaseset_sub_func);
+    Py_CLEAR(viewbaseset_xor_func);
 }
 
 static PyMethodDef multidict_module_methods[] = {
-    {
-        "getversion",
-        (PyCFunction)getversion,
-        METH_O
-    },
-    {
-        NULL,
-        NULL
-    }   /* sentinel */
+    {"getversion", (PyCFunction)getversion, METH_O},
+    {NULL, NULL}   /* sentinel */
 };
 
 static PyModuleDef multidict_module = {
@@ -1659,21 +1628,6 @@ PyInit__multidict(void)
         goto fail;
     }
 
-#define WITH_MOD(NAME)                      \
-    Py_CLEAR(module);                       \
-    module = PyImport_ImportModule(NAME);   \
-    if (module == NULL) {                   \
-        goto fail;                          \
-    }
-
-#define GET_MOD_ATTR(VAR, NAME)                 \
-    VAR = PyObject_GetAttrString(module, NAME); \
-    if (VAR == NULL) {                          \
-        goto fail;                              \
-    }
-
-    Py_CLEAR(module);                       \
-
     /* Instantiate this module */
     module = PyModule_Create(&multidict_module);
     if (module == NULL) {
@@ -1746,7 +1700,4 @@ PyInit__multidict(void)
     Py_XDECREF(multidict_str_lower);
 
     return NULL;
-
-#undef WITH_MOD
-#undef GET_MOD_ATTR
 }

multidict/_multidict_base.py

@@ -1,59 +1,16 @@
 import sys
-from collections.abc import (
-    Container,
-    ItemsView,
-    Iterable,
-    KeysView,
-    Sequence,
-    Set,
-)
-from typing import Literal, Union
+from collections.abc import ItemsView, Iterable, KeysView, Sequence
+from typing import Union
 
 if sys.version_info >= (3, 10):
     from types import NotImplementedType
 else:
     from typing import Any as NotImplementedType
 
-if sys.version_info >= (3, 11):
-    from typing import assert_never
-else:
-    from typing_extensions import assert_never
-
 
 _ViewArg = Union[KeysView[str], ItemsView[str, object]]
 
 
-def _viewbaseset_richcmp(
-    view: _ViewArg, other: object, op: Literal[0, 1, 2, 3, 4, 5]
-) -> Union[bool, NotImplementedType]:
-    if not isinstance(other, Set):
-        return NotImplemented  # type: ignore[no-any-return]
-    if op == 0:  # <
-        return len(view) < len(other) and view <= other
-    elif op == 1:  # <=
-        if len(view) > len(other):
-            return False
-        for elem in view:
-            if elem not in other:
-                return False
-        return True
-    elif op == 2:  # ==
-        return len(view) == len(other) and view <= other
-    elif op == 3:  # !=
-        return not view == other
-    elif op == 4:  # >
-        return len(view) > len(other) and view >= other
-    elif op == 5:  # >=
-        if len(view) < len(other):
-            return False
-        for elem in other:
-            if elem not in view:
-                return False
-        return True
-    else:  # pragma: no cover
-        assert_never(op)
-
-
 def _viewbaseset_and(
     view: _ViewArg, other: object
 ) -> Union[set[Sequence[object]], NotImplementedType]:
@@ -92,19 +49,3 @@ def _viewbaseset_xor(
     lft = set(iter(view))
     rgt = set(iter(other))
     return lft ^ rgt
-
-
-def _itemsview_isdisjoint(view: Container[object], other: Iterable[object]) -> bool:
-    "Return True if two sets have a null intersection."
-    for v in other:
-        if v in view:
-            return False
-    return True
-
-
-def _keysview_isdisjoint(view: Container[object], other: Iterable[object]) -> bool:
-    "Return True if two sets have a null intersection."
-    for k in other:
-        if k in view:
-            return False
-    return True