betterproto2-compiler 0.0.1__py3-none-any.whl

betterproto2_compiler/_types.py

@@ -0,0 +1,13 @@
+from typing import (
+    TYPE_CHECKING,
+    TypeVar,
+)
+
+if TYPE_CHECKING:
+    from grpclib._typing import IProtoMessage
+
+    from . import Message
+
+# Bound type variable to allow methods to return `self` of subclasses
+T = TypeVar("T", bound="Message")
+ST = TypeVar("ST", bound="IProtoMessage")

betterproto2_compiler/casing.py

@@ -0,0 +1,140 @@
+import keyword
+import re
+
+# Word delimiters and symbols that will not be preserved when re-casing.
+# language=PythonRegExp
+SYMBOLS = "[^a-zA-Z0-9]*"
+
+# Optionally capitalized word.
+# language=PythonRegExp
+WORD = "[A-Z]*[a-z]*[0-9]*"
+
+# Uppercase word, not followed by lowercase letters.
+# language=PythonRegExp
+WORD_UPPER = "[A-Z]+(?![a-z])[0-9]*"
+
+
+def safe_snake_case(value: str) -> str:
+    """Snake case a value taking into account Python keywords."""
+    value = snake_case(value)
+    value = sanitize_name(value)
+    return value
+
+
+def snake_case(value: str, strict: bool = True) -> str:
+    """
+    Join words with an underscore into lowercase and remove symbols.
+
+    Parameters
+    -----------
+    value: :class:`str`
+        The value to convert.
+    strict: :class:`bool`
+        Whether or not to force single underscores.
+
+    Returns
+    --------
+    :class:`str`
+        The value in snake_case.
+    """
+
+    def substitute_word(symbols: str, word: str, is_start: bool) -> str:
+        if not word:
+            return ""
+        if strict:
+            delimiter_count = 0 if is_start else 1  # Single underscore if strict.
+        elif is_start:
+            delimiter_count = len(symbols)
+        elif word.isupper() or word.islower():
+            delimiter_count = max(1, len(symbols))  # Preserve all delimiters if not strict.
+        else:
+            delimiter_count = len(symbols) + 1  # Extra underscore for leading capital.
+
+        return ("_" * delimiter_count) + word.lower()
+
+    snake = re.sub(
+        f"(^)?({SYMBOLS})({WORD_UPPER}|{WORD})",
+        lambda groups: substitute_word(groups[2], groups[3], groups[1] is not None),
+        value,
+    )
+    return snake
+
+
+def pascal_case(value: str, strict: bool = True) -> str:
+    """
+    Capitalize each word and remove symbols.
+
+    Parameters
+    -----------
+    value: :class:`str`
+        The value to convert.
+    strict: :class:`bool`
+        Whether or not to output only alphanumeric characters.
+
+    Returns
+    --------
+    :class:`str`
+        The value in PascalCase.
+    """
+
+    def substitute_word(symbols, word):
+        if strict:
+            return word.capitalize()  # Remove all delimiters
+
+        if word.islower():
+            delimiter_length = len(symbols[:-1])  # Lose one delimiter
+        else:
+            delimiter_length = len(symbols)  # Preserve all delimiters
+
+        return ("_" * delimiter_length) + word.capitalize()
+
+    return re.sub(
+        f"({SYMBOLS})({WORD_UPPER}|{WORD})",
+        lambda groups: substitute_word(groups[1], groups[2]),
+        value,
+    )
+
+
+def camel_case(value: str, strict: bool = True) -> str:
+    """
+    Capitalize all words except first and remove symbols.
+
+    Parameters
+    -----------
+    value: :class:`str`
+        The value to convert.
+    strict: :class:`bool`
+        Whether or not to output only alphanumeric characters.
+
+    Returns
+    --------
+    :class:`str`
+        The value in camelCase.
+    """
+    return lowercase_first(pascal_case(value, strict=strict))
+
+
+def lowercase_first(value: str) -> str:
+    """
+    Lower cases the first character of the value.
+
+    Parameters
+    ----------
+    value: :class:`str`
+        The value to lower case.
+
+    Returns
+    -------
+    :class:`str`
+        The lower cased string.
+    """
+    return value[0:1].lower() + value[1:]
+
+
+def sanitize_name(value: str) -> str:
+    # https://www.python.org/dev/peps/pep-0008/#descriptive-naming-styles
+    if keyword.iskeyword(value):
+        return f"{value}_"
+    if not value.isidentifier():
+        return f"_{value}"
+    return value

betterproto2_compiler/compile/importing.py

@@ -0,0 +1,193 @@
+from __future__ import annotations
+
+import os
+from typing import (
+    TYPE_CHECKING,
+    Dict,
+    List,
+    Set,
+    Tuple,
+    Type,
+)
+
+from ..casing import safe_snake_case
+from ..lib.google import protobuf as google_protobuf
+from .naming import pythonize_class_name
+
+if TYPE_CHECKING:
+    from ..plugin.models import PluginRequestCompiler
+    from ..plugin.typing_compiler import TypingCompiler
+
+WRAPPER_TYPES: Dict[str, Type] = {
+    ".google.protobuf.DoubleValue": google_protobuf.DoubleValue,
+    ".google.protobuf.FloatValue": google_protobuf.FloatValue,
+    ".google.protobuf.Int32Value": google_protobuf.Int32Value,
+    ".google.protobuf.Int64Value": google_protobuf.Int64Value,
+    ".google.protobuf.UInt32Value": google_protobuf.UInt32Value,
+    ".google.protobuf.UInt64Value": google_protobuf.UInt64Value,
+    ".google.protobuf.BoolValue": google_protobuf.BoolValue,
+    ".google.protobuf.StringValue": google_protobuf.StringValue,
+    ".google.protobuf.BytesValue": google_protobuf.BytesValue,
+}
+
+
+def parse_source_type_name(field_type_name: str, request: "PluginRequestCompiler") -> Tuple[str, str]:
+    """
+    Split full source type name into package and type name.
+    E.g. 'root.package.Message' -> ('root.package', 'Message')
+         'root.Message.SomeEnum' -> ('root', 'Message.SomeEnum')
+
+    The function goes through the symbols that have been defined (names, enums,
+    packages) to find the actual package and name of the object that is referenced.
+    """
+    if field_type_name[0] != ".":
+        raise RuntimeError("relative names are not supported")
+    field_type_name = field_type_name[1:]
+    parts = field_type_name.split(".")
+
+    answer = None
+
+    # a.b.c:
+    # i=0: "", "a.b.c"
+    # i=1: "a", "b.c"
+    # i=2: "a.b", "c"
+    for i in range(len(parts)):
+        package_name, object_name = ".".join(parts[:i]), ".".join(parts[i:])
+
+        package = request.output_packages.get(package_name)
+
+        if not package:
+            continue
+
+        if object_name in package.messages or object_name in package.enums:
+            if answer:
+                # This should have already been handeled by protoc
+                raise ValueError(f"ambiguous definition: {field_type_name}")
+            answer = package_name, object_name
+
+    if answer:
+        return answer
+
+    raise ValueError(f"can't find type name: {field_type_name}")
+
+
+def get_type_reference(
+    *,
+    package: str,
+    imports: set,
+    source_type: str,
+    typing_compiler: TypingCompiler,
+    request: "PluginRequestCompiler",
+    unwrap: bool = True,
+    pydantic: bool = False,
+) -> str:
+    """
+    Return a Python type name for a proto type reference. Adds the import if
+    necessary. Unwraps well known type if required.
+    """
+    if unwrap:
+        if source_type in WRAPPER_TYPES:
+            wrapped_type = type(WRAPPER_TYPES[source_type]().value)
+            return typing_compiler.optional(wrapped_type.__name__)
+
+        if source_type == ".google.protobuf.Duration":
+            return "datetime.timedelta"
+
+        elif source_type == ".google.protobuf.Timestamp":
+            return "datetime.datetime"
+
+    source_package, source_type = parse_source_type_name(source_type, request)
+
+    current_package: List[str] = package.split(".") if package else []
+    py_package: List[str] = source_package.split(".") if source_package else []
+    py_type: str = pythonize_class_name(source_type)
+
+    compiling_google_protobuf = current_package == ["google", "protobuf"]
+    importing_google_protobuf = py_package == ["google", "protobuf"]
+    if importing_google_protobuf and not compiling_google_protobuf:
+        py_package = ["betterproto", "lib"] + (["pydantic"] if pydantic else []) + py_package
+
+    if py_package[:1] == ["betterproto"]:
+        return reference_absolute(imports, py_package, py_type)
+
+    if py_package == current_package:
+        return reference_sibling(py_type)
+
+    if py_package[: len(current_package)] == current_package:
+        return reference_descendent(current_package, imports, py_package, py_type)
+
+    if current_package[: len(py_package)] == py_package:
+        return reference_ancestor(current_package, imports, py_package, py_type)
+
+    return reference_cousin(current_package, imports, py_package, py_type)
+
+
+def reference_absolute(imports: Set[str], py_package: List[str], py_type: str) -> str:
+    """
+    Returns a reference to a python type located in the root, i.e. sys.path.
+    """
+    string_import = ".".join(py_package)
+    string_alias = safe_snake_case(string_import)
+    imports.add(f"import {string_import} as {string_alias}")
+    return f"{string_alias}.{py_type}"
+
+
+def reference_sibling(py_type: str) -> str:
+    """
+    Returns a reference to a python type within the same package as the current package.
+    """
+    return f"{py_type}"
+
+
+def reference_descendent(current_package: List[str], imports: Set[str], py_package: List[str], py_type: str) -> str:
+    """
+    Returns a reference to a python type in a package that is a descendent of the
+    current package, and adds the required import that is aliased to avoid name
+    conflicts.
+    """
+    importing_descendent = py_package[len(current_package) :]
+    string_from = ".".join(importing_descendent[:-1])
+    string_import = importing_descendent[-1]
+    if string_from:
+        string_alias = "_".join(importing_descendent)
+        imports.add(f"from .{string_from} import {string_import} as {string_alias}")
+        return f"{string_alias}.{py_type}"
+    else:
+        imports.add(f"from . import {string_import}")
+        return f"{string_import}.{py_type}"
+
+
+def reference_ancestor(current_package: List[str], imports: Set[str], py_package: List[str], py_type: str) -> str:
+    """
+    Returns a reference to a python type in a package which is an ancestor to the
+    current package, and adds the required import that is aliased (if possible) to avoid
+    name conflicts.
+
+    Adds trailing __ to avoid name mangling (python.org/dev/peps/pep-0008/#id34).
+    """
+    distance_up = len(current_package) - len(py_package)
+    if py_package:
+        string_import = py_package[-1]
+        string_alias = f"_{'_' * distance_up}{string_import}__"
+        string_from = f"..{'.' * distance_up}"
+        imports.add(f"from {string_from} import {string_import} as {string_alias}")
+        return f"{string_alias}.{py_type}"
+    else:
+        string_alias = f"{'_' * distance_up}{py_type}__"
+        imports.add(f"from .{'.' * distance_up} import {py_type} as {string_alias}")
+        return string_alias
+
+
+def reference_cousin(current_package: List[str], imports: Set[str], py_package: List[str], py_type: str) -> str:
+    """
+    Returns a reference to a python type in a package that is not descendent, ancestor
+    or sibling, and adds the required import that is aliased to avoid name conflicts.
+    """
+    shared_ancestry = os.path.commonprefix([current_package, py_package])  # type: ignore
+    distance_up = len(current_package) - len(shared_ancestry)
+    string_from = f".{'.' * distance_up}" + ".".join(py_package[len(shared_ancestry) : -1])
+    string_import = py_package[-1]
+    # Add trailing __ to avoid name mangling (python.org/dev/peps/pep-0008/#id34)
+    string_alias = f"{'_' * distance_up}" + safe_snake_case(".".join(py_package[len(shared_ancestry) :])) + "__"
+    imports.add(f"from {string_from} import {string_import} as {string_alias}")
+    return f"{string_alias}.{py_type}"

betterproto2_compiler/compile/naming.py

@@ -0,0 +1,21 @@
+from betterproto2_compiler import casing
+
+
+def pythonize_class_name(name: str) -> str:
+    return casing.pascal_case(name)
+
+
+def pythonize_field_name(name: str) -> str:
+    return casing.safe_snake_case(name)
+
+
+def pythonize_method_name(name: str) -> str:
+    return casing.safe_snake_case(name)
+
+
+def pythonize_enum_member_name(name: str, enum_name: str) -> str:
+    enum_name = casing.snake_case(enum_name).upper()
+    find = name.find(enum_name)
+    if find != -1:
+        name = name[find + len(enum_name) :].strip("_")
+    return casing.sanitize_name(name)

betterproto2_compiler/enum.py

@@ -0,0 +1,180 @@
+from __future__ import annotations
+
+from enum import (
+    EnumMeta,
+    IntEnum,
+)
+from types import MappingProxyType
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    Optional,
+    Tuple,
+)
+
+if TYPE_CHECKING:
+    from collections.abc import (
+        Generator,
+        Mapping,
+    )
+
+    from typing_extensions import (
+        Never,
+        Self,
+    )
+
+
+def _is_descriptor(obj: object) -> bool:
+    return hasattr(obj, "__get__") or hasattr(obj, "__set__") or hasattr(obj, "__delete__")
+
+
+class EnumType(EnumMeta if TYPE_CHECKING else type):
+    _value_map_: Mapping[int, Enum]
+    _member_map_: Mapping[str, Enum]
+
+    def __new__(mcs, name: str, bases: Tuple[type, ...], namespace: Dict[str, Any]) -> Self:
+        value_map = {}
+        member_map = {}
+
+        new_mcs = type(
+            f"{name}Type",
+            tuple(
+                dict.fromkeys([base.__class__ for base in bases if base.__class__ is not type] + [EnumType, type])
+            ),  # reorder the bases so EnumType and type are last to avoid conflicts
+            {"_value_map_": value_map, "_member_map_": member_map},
+        )
+
+        members = {
+            name: value for name, value in namespace.items() if not _is_descriptor(value) and not name.startswith("__")
+        }
+
+        cls = type.__new__(
+            new_mcs,
+            name,
+            bases,
+            {key: value for key, value in namespace.items() if key not in members},
+        )
+        # this allows us to disallow member access from other members as
+        # members become proper class variables
+
+        for name, value in members.items():
+            member = value_map.get(value)
+            if member is None:
+                member = cls.__new__(cls, name=name, value=value)  # type: ignore
+                value_map[value] = member
+            member_map[name] = member
+            type.__setattr__(new_mcs, name, member)
+
+        return cls
+
+    if not TYPE_CHECKING:
+
+        def __call__(cls, value: int) -> Enum:
+            try:
+                return cls._value_map_[value]
+            except (KeyError, TypeError):
+                raise ValueError(f"{value!r} is not a valid {cls.__name__}") from None
+
+        def __iter__(cls) -> Generator[Enum, None, None]:
+            yield from cls._member_map_.values()
+
+        def __reversed__(cls) -> Generator[Enum, None, None]:
+            yield from reversed(cls._member_map_.values())
+
+        def __getitem__(cls, key: str) -> Enum:
+            return cls._member_map_[key]
+
+        @property
+        def __members__(cls) -> MappingProxyType[str, Enum]:
+            return MappingProxyType(cls._member_map_)
+
+    def __repr__(cls) -> str:
+        return f"<enum {cls.__name__!r}>"
+
+    def __len__(cls) -> int:
+        return len(cls._member_map_)
+
+    def __setattr__(cls, name: str, value: Any) -> Never:
+        raise AttributeError(f"{cls.__name__}: cannot reassign Enum members.")
+
+    def __delattr__(cls, name: str) -> Never:
+        raise AttributeError(f"{cls.__name__}: cannot delete Enum members.")
+
+    def __contains__(cls, member: object) -> bool:
+        return isinstance(member, cls) and member.name in cls._member_map_
+
+
+class Enum(IntEnum if TYPE_CHECKING else int, metaclass=EnumType):
+    """
+    The base class for protobuf enumerations, all generated enumerations will
+    inherit from this. Emulates `enum.IntEnum`.
+    """
+
+    name: Optional[str]
+    value: int
+
+    if not TYPE_CHECKING:
+
+        def __new__(cls, *, name: Optional[str], value: int) -> Self:
+            self = super().__new__(cls, value)
+            super().__setattr__(self, "name", name)
+            super().__setattr__(self, "value", value)
+            return self
+
+    def __str__(self) -> str:
+        return self.name or "None"
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}.{self.name}"
+
+    def __setattr__(self, key: str, value: Any) -> Never:
+        raise AttributeError(f"{self.__class__.__name__} Cannot reassign a member's attributes.")
+
+    def __delattr__(self, item: Any) -> Never:
+        raise AttributeError(f"{self.__class__.__name__} Cannot delete a member's attributes.")
+
+    def __copy__(self) -> Self:
+        return self
+
+    def __deepcopy__(self, memo: Any) -> Self:
+        return self
+
+    @classmethod
+    def try_value(cls, value: int = 0) -> Self:
+        """Return the value which corresponds to the value.
+
+        Parameters
+        -----------
+        value: :class:`int`
+            The value of the enum member to get.
+
+        Returns
+        -------
+        :class:`Enum`
+            The corresponding member or a new instance of the enum if
+            ``value`` isn't actually a member.
+        """
+        try:
+            return cls._value_map_[value]
+        except (KeyError, TypeError):
+            return cls.__new__(cls, name=None, value=value)
+
+    @classmethod
+    def from_string(cls, name: str) -> Self:
+        """Return the value which corresponds to the string name.
+
+        Parameters
+        -----------
+        name: :class:`str`
+            The name of the enum member to get.
+
+        Raises
+        -------
+        :exc:`ValueError`
+            The member was not found in the Enum.
+        """
+        try:
+            return cls._member_map_[name]
+        except KeyError as e:
+            raise ValueError(f"Unknown value {name} for enum {cls.__name__}") from e