Add Odb.read_header; Have Odb.read return ObjectType enum

pygit2/_pygit2.pyi

@@ -518,7 +518,8 @@ class Odb:
     def add_backend(self, backend: OdbBackend, priority: int) -> None: ...
     def add_disk_alternate(self, path: str | Path) -> None: ...
     def exists(self, oid: _OidArg) -> bool: ...
-    def read(self, oid: _OidArg) -> tuple[int, bytes]: ...
+    def read(self, oid: _OidArg) -> tuple[ObjectType, bytes]: ...
+    def read_header(self, oid: _OidArg) -> tuple[ObjectType, int]: ...
     def write(self, type: int, data: bytes | str) -> Oid: ...
     def __contains__(self, other: _OidArg) -> bool: ...
     def __iter__(self) -> Iterator[Oid]: ...  # Odb_as_iter

src/odb.c

@@ -37,6 +37,8 @@
 
 extern PyTypeObject OdbBackendType;
 
+extern PyObject *ObjectTypeEnum;
+
 static git_otype
 int_to_loose_object_type(int type_id)
 {
@@ -170,7 +172,7 @@ Odb_read_raw(git_odb *odb, const git_oid *oid, size_t len)
 }
 
 PyDoc_STRVAR(Odb_read__doc__,
-  "read(oid) -> type, data, size\n"
+  "read(oid: Oid) -> tuple[enums.ObjectType, bytes]\n"
   "\n"
   "Read raw object data from the object db.");
 
@@ -180,6 +182,8 @@ Odb_read(Odb *self, PyObject *py_hex)
     git_oid oid;
     git_odb_object *obj;
     size_t len;
+    git_object_t type;
+    PyObject* type_enum;
     PyObject* tuple;
 
     len = py_oid_to_git_oid(py_hex, &oid);
@@ -190,16 +194,58 @@ Odb_read(Odb *self, PyObject *py_hex)
     if (obj == NULL)
         return NULL;
 
+    // Convert type to ObjectType enum
+    type = git_odb_object_type(obj);
+    type_enum = pygit2_enum(ObjectTypeEnum, type);
+
     tuple = Py_BuildValue(
-        "(ny#)",
-        git_odb_object_type(obj),
+        "(Oy#)",
+        type_enum,
         git_odb_object_data(obj),
         git_odb_object_size(obj));
 
     git_odb_object_free(obj);
     return tuple;
 }
 
+PyDoc_STRVAR(Odb_read_header__doc__,
+    "read_header(oid: Oid) -> tuple[enums.ObjectType, size\n"
+    "\n"
+    "Read the header of an object from the database, without reading its full\n"
+    "contents.\n"
+    "\n"
+    "The header includes the type and the length of an object.\n"
+    "\n"
+    "Note that most backends do not support reading only the header of an object,\n"
+    "so the whole object may be read and then the header will be returned.");
+
+PyObject *
+Odb_read_header(Odb *self, PyObject *py_hex)
+{
+    git_oid oid;
+    int err;
+    size_t len;
+    git_object_t type;
+    PyObject* type_enum;
+    PyObject* tuple;
+
+    len = py_oid_to_git_oid(py_hex, &oid);
+    if (len == 0)
+        return NULL;
+
+    err = git_odb_read_header(&len, &type, self->odb, &oid);
+    if (err != 0) {
+        Error_set_oid(err, &oid, len);
+        return NULL;
+    }
+
+    // Convert type to ObjectType enum
+    type_enum = pygit2_enum(ObjectTypeEnum, type);
+
+    tuple = Py_BuildValue("(On)", type_enum, len);
+    return tuple;
+}
+
 PyDoc_STRVAR(Odb_write__doc__,
     "write(type: int, data: bytes) -> Oid\n"
     "\n"
@@ -301,6 +347,7 @@ Odb_add_backend(Odb *self, PyObject *args)
 PyMethodDef Odb_methods[] = {
     METHOD(Odb, add_disk_alternate, METH_O),
     METHOD(Odb, read, METH_O),
+    METHOD(Odb, read_header, METH_O),
     METHOD(Odb, write, METH_VARARGS),
     METHOD(Odb, exists, METH_O),
     METHOD(Odb, add_backend, METH_VARARGS),

src/pygit2.c

@@ -47,6 +47,7 @@ PyObject *FileModeEnum;
 PyObject *FileStatusEnum;
 PyObject *MergeAnalysisEnum;
 PyObject *MergePreferenceEnum;
+PyObject *ObjectTypeEnum;
 PyObject *ReferenceTypeEnum;
 
 extern PyTypeObject RepositoryType;
@@ -354,6 +355,7 @@ forget_enums(void)
     Py_CLEAR(FileStatusEnum);
     Py_CLEAR(MergeAnalysisEnum);
     Py_CLEAR(MergePreferenceEnum);
+    Py_CLEAR(ObjectTypeEnum);
     Py_CLEAR(ReferenceTypeEnum);
 }
 
@@ -389,6 +391,7 @@ _cache_enums(PyObject *self, PyObject *args)
     CACHE_PYGIT2_ENUM(FileStatus);
     CACHE_PYGIT2_ENUM(MergeAnalysis);
     CACHE_PYGIT2_ENUM(MergePreference);
+    CACHE_PYGIT2_ENUM(ObjectType);
     CACHE_PYGIT2_ENUM(ReferenceType);
 
 #undef CACHE_PYGIT2_ENUM

test/test_odb.py

@@ -41,6 +41,7 @@
 BLOB_HEX = 'af431f20fc541ed6d5afede3e2dc7160f6f01f16'
 BLOB_RAW = binascii.unhexlify(BLOB_HEX.encode('ascii'))
 BLOB_OID = Oid(raw=BLOB_RAW)
+BLOB_CONTENTS = b'a contents\n'
 
 
 def test_emptyodb(barerepo: Repository) -> None:
@@ -75,14 +76,29 @@ def test_read(odb: Odb) -> None:
     ab = odb.read(BLOB_OID)
     a = odb.read(BLOB_HEX)
     assert ab == a
-    assert (ObjectType.BLOB, b'a contents\n') == a
+    assert (ObjectType.BLOB, BLOB_CONTENTS) == a
+    assert isinstance(a[0], ObjectType)
 
     a2 = odb.read('7f129fd57e31e935c6d60a0c794efe4e6927664b')
     assert (ObjectType.BLOB, b'a contents 2\n') == a2
+    assert isinstance(a2[0], ObjectType)
 
     a_hex_prefix = BLOB_HEX[:4]
     a3 = odb.read(a_hex_prefix)
-    assert (ObjectType.BLOB, b'a contents\n') == a3
+    assert (ObjectType.BLOB, BLOB_CONTENTS) == a3
+    assert isinstance(a3[0], ObjectType)
+
+
+def test_read_header(odb: Odb) -> None:
+    with pytest.raises(TypeError):
+        odb.read_header(123)  # type: ignore
+    utils.assertRaisesWithArg(KeyError, '1' * 40, odb.read_header, '1' * 40)
+
+    ab = odb.read_header(BLOB_OID)
+    a = odb.read_header(BLOB_HEX)
+    assert ab == a
+    assert (ObjectType.BLOB, len(BLOB_CONTENTS)) == a
+    assert isinstance(a[0], ObjectType)
 
 
 def test_write(odb: Odb) -> None: