Merge pull request #74 from Amulet-Team/pybind

C++ Rewrite

.gitignore

@@ -119,9 +119,3 @@ temp
 *.vs/
 *.vcxproj*
 *.sln
-
-# C++ files
-*.c
-!**/_cpp/**/*.c
-*.cpp
-!**/_cpp/**/*.cpp

MANIFEST.in

@@ -1 +1,2 @@
-recursive-include src *.pyi py.typed *.py *.pyx *.pxd *.cpp *.hpp
+prune tests
+recursive-include src *.pyi py.typed *.py *.cpp *.hpp

docs_source/getting_started.rst

@@ -52,32 +52,32 @@
    bedrock_named_tag: NamedTag
 
    # Load binary NBT
-   named_tag = amulet_nbt.load(
+   named_tag = amulet_nbt.read_nbt(
       "the/path/to/your/binary/nbt/file",
       preset=java_encoding,
       compressed=True,  # These inputs must be specified as keyword inputs like this.
    )  # from a file
-   named_tag = amulet_nbt.load(
+   named_tag = amulet_nbt.read_nbt(
       "the/path/to/your/binary/nbt/file",
       compressed=True,  # These inputs must be specified as keyword inputs like this.
       little_endian=False,  # If you do not define them they will default to these values
       string_encoding=mutf8_encoding
    )  # from a file
-   named_tag = amulet_nbt.load(b'<nbt file bytes>')  # from a bytes object
+   named_tag = amulet_nbt.read_nbt(b'<nbt file bytes>')  # from a bytes object
 
    # Note that Java Edition usually uses compressed modified UTF-8.
-   java_named_tag = amulet_nbt.load(
+   java_named_tag = amulet_nbt.read_nbt(
       "the/path/to/your/binary/java/nbt/file",
       string_encoding=mutf8_encoding
    )
 
    # Bedrock edition data is stored in little endian format and uses non-compressed UTF-8 but can also have arbitrary bytes.
-   bedrock_named_tag = amulet_nbt.load(
+   bedrock_named_tag = amulet_nbt.read_nbt(
       "the/path/to/your/binary/bedrock/nbt/file",
       preset=bedrock_encoding,
       compressed=False,
    )
-   bedrock_named_tag = amulet_nbt.load(
+   bedrock_named_tag = amulet_nbt.read_nbt(
       "the/path/to/your/binary/bedrock/nbt/file",
       compressed=False,
       little_endian=True,
@@ -117,7 +117,7 @@
 
 
    # You can also parse the stringified NBT format used in Java commands.
-   tag = amulet_nbt.from_snbt('{key1: "value", key2: 0b, key3: 0.0f}')
+   tag = amulet_nbt.read_snbt('{key1: "value", key2: 0b, key3: 0.0f}')
    # tag should look like this
    # TAG_Compound(
    #   key1: TAG_String("value"),

docs_source/index.rst

@@ -186,24 +186,11 @@
 
 These are functions to load the binary and stringified NBT formats.
 
-.. autofunction:: amulet_nbt.load
-.. autofunction:: amulet_nbt.load_array
+.. autofunction:: amulet_nbt.read_nbt
+.. autofunction:: amulet_nbt.read_nbt_array
 .. autoclass:: amulet_nbt.ReadOffset
    :members:
-.. autofunction:: amulet_nbt.from_snbt
-
-
-############
- Exceptions
-############
-
-.. autoexception:: amulet_nbt.NBTError
-.. autoexception:: amulet_nbt.NBTLoadError
-   :show-inheritance:
-.. autoexception:: amulet_nbt.NBTFormatError
-   :show-inheritance:
-.. autoexception:: amulet_nbt.SNBTParseError
-   :show-inheritance:
+.. autofunction:: amulet_nbt.read_snbt
 
 
 #################
@@ -217,7 +204,7 @@ These are instances of a class storing C++ functions to encode and decode string
    :inherited-members:
    :undoc-members:
 
-They can be passed to the string_encoding argument in to_nbt, save_to, load and load_array to control the string encoding behaviour.
+They can be passed to the string_encoding argument in to_nbt, save_to, read_nbt and read_nbt_array to control the string encoding behaviour.
 
 The usual string encoding scheme is called UTF-8.
 
@@ -237,7 +224,7 @@ Java Edition uses a modified version of UTF-8 implemented by the Java programmin
  Encoding Presets
 ##################
 
-The string encoding and endianness can be defined separatly but for simplicity the following presets have been defined.
+The string encoding and endianness can be defined separately but for simplicity the following presets have been defined.
 
 .. autodata:: amulet_nbt.java_encoding
 .. autodata:: amulet_nbt.bedrock_encoding

pyproject.toml

@@ -2,8 +2,7 @@
 requires = [
     "setuptools >= 42",
     "wheel",
-    "cython ~= 3.0",
     "versioneer",
-    "numpy ~= 1.17"
+    "pybind11 ~= 2.12",
 ]
 build-backend = "setuptools.build_meta"

setup.cfg

@@ -15,11 +15,11 @@ platforms = any
 [options]
 package_dir=
     =src
-packages = find_namespace:
+packages = find:
 zip_safe = False
 python_requires = ~=3.11
 install_requires =
-    numpy ~= 1.17
+    numpy >= 1.17, < 3.0
 
 [options.packages.find]
 where=src
@@ -30,6 +30,9 @@ amulet_nbt =
     py.typed
     *.pyd
     *.so
+    *.dylib
+    **/*.hpp
+    cpp/**/*.cpp
 
 
 [options.extras_require]

setup.py

@@ -1,10 +1,10 @@
-from setuptools import setup
-from Cython.Build import cythonize
+from setuptools import setup, Extension
 import versioneer
-import numpy
 import sysconfig
 from distutils import ccompiler
 import sys
+import pybind11
+import glob
 
 if (sysconfig.get_config_var("CXX") or ccompiler.get_default_compiler()).split()[
     0
@@ -20,8 +20,24 @@
 setup(
     version=versioneer.get_version(),
     cmdclass=versioneer.get_cmdclass(),
-    include_dirs=[numpy.get_include()],
-    ext_modules=cythonize(
-        f"src/**/*.pyx", language_level=3, aliases={"CPPCARGS": CompileArgs}
-    ),
+    libraries=[
+        (
+            "amulet_nbt",
+            dict(
+                sources=glob.glob("src/amulet_nbt/cpp/**/*.cpp", recursive=True),
+                include_dirs=["src/amulet_nbt/include"],
+                cflags=CompileArgs,
+            ),
+        )
+    ],
+    ext_modules=[
+        Extension(
+            name="amulet_nbt._nbt",
+            sources=glob.glob("src/amulet_nbt/pybind/**/*.cpp", recursive=True),
+            include_dirs=["src/amulet_nbt/include", pybind11.get_include()],
+            libraries=["amulet_nbt"],
+            define_macros=[("PYBIND11_DETAILED_ERROR_MESSAGES", None)],
+            extra_compile_args=CompileArgs,
+        )
+    ],
 )

src/amulet_nbt/__init__.pyx → src/amulet_nbt/__init__.py

@@ -1,11 +1,94 @@
 import re
-from typing import Union
-
+from typing import TypeAlias
 from . import _version
 
 __version__ = _version.get_versions()["version"]
-__major__ = int(re.match(r"\d+", __version__)[0])
+__major__ = int(re.match(r"\d+", __version__)[0])  # type: ignore
+
+del re
+del _version
+
+
+def get_include() -> str:
+    import os
+
+    return os.path.join(__path__[0], "include")
+
+
+from ._nbt import (
+    # Abstract classes
+    AbstractBaseTag,
+    AbstractBaseImmutableTag,
+    AbstractBaseMutableTag,
+    AbstractBaseNumericTag,
+    AbstractBaseIntTag,
+    AbstractBaseFloatTag,
+    AbstractBaseArrayTag,
+    # Tag classes
+    ByteTag,
+    ShortTag,
+    IntTag,
+    LongTag,
+    FloatTag,
+    DoubleTag,
+    ByteArrayTag,
+    StringTag,
+    ListTag,
+    CompoundTag,
+    IntArrayTag,
+    LongArrayTag,
+    # Tag class aliases
+    ByteTag as TAG_Byte,
+    ShortTag as TAG_Short,
+    IntTag as TAG_Int,
+    LongTag as TAG_Long,
+    FloatTag as TAG_Float,
+    DoubleTag as TAG_Double,
+    ByteArrayTag as TAG_Byte_Array,
+    StringTag as TAG_String,
+    ListTag as TAG_List,
+    CompoundTag as TAG_Compound,
+    IntArrayTag as TAG_Int_Array,
+    LongArrayTag as TAG_Long_Array,
+    NamedTag,
+    read_nbt,
+    read_nbt_array,
+    ReadOffset,
+    read_snbt,
+    StringEncoding,
+    mutf8_encoding,
+    utf8_encoding,
+    utf8_escape_encoding,
+    EncodingPreset,
+    java_encoding,
+    bedrock_encoding,
+)
+
+SNBTType: TypeAlias = str
+IntType: TypeAlias = ByteTag | ShortTag | IntTag | LongTag
+FloatType: TypeAlias = FloatTag | DoubleTag
+NumberType: TypeAlias = ByteTag | ShortTag | IntTag | LongTag | FloatTag | DoubleTag
+ArrayType: TypeAlias = ByteArrayTag | IntArrayTag | LongArrayTag
+AnyNBT: TypeAlias = (
+    ByteTag
+    | ShortTag
+    | IntTag
+    | LongTag
+    | FloatTag
+    | DoubleTag
+    | ByteArrayTag
+    | StringTag
+    | ListTag
+    | CompoundTag
+    | IntArrayTag
+    | LongArrayTag
+)
+
+
 __all__ = [
+    "__version__",
+    "__major__",
+    "get_include",
     "AbstractBaseTag",
     "AbstractBaseImmutableTag",
     "AbstractBaseMutableTag",
@@ -38,14 +121,10 @@
     "CompoundTag",
     "TAG_Compound",
     "NamedTag",
-    "load",
-    "load_array",
+    "read_nbt",
+    "read_nbt_array",
     "ReadOffset",
-    "from_snbt",
-    "NBTError",
-    "NBTLoadError",
-    "NBTFormatError",
-    "SNBTParseError",
+    "read_snbt",
     "SNBTType",
     "IntType",
     "FloatType",
@@ -60,97 +139,3 @@
     "java_encoding",
     "bedrock_encoding",
 ]
-
-from ._tag.abc import (
-    AbstractBaseTag,
-    AbstractBaseImmutableTag,
-    AbstractBaseMutableTag,
-)
-from ._tag.numeric import AbstractBaseNumericTag
-
-# Types
-from ._tag.int import (
-    AbstractBaseIntTag,
-    ByteTag,
-    ShortTag,
-    IntTag,
-    LongTag,
-    ByteTag as TAG_Byte,
-    ShortTag as TAG_Short,
-    IntTag as TAG_Int,
-    LongTag as TAG_Long,
-)
-from ._tag.float import (
-    AbstractBaseFloatTag,
-    FloatTag,
-    DoubleTag,
-    FloatTag as TAG_Float,
-    DoubleTag as TAG_Double,
-)
-from ._tag.array import (
-    AbstractBaseArrayTag,
-    ByteArrayTag,
-    IntArrayTag,
-    LongArrayTag,
-    ByteArrayTag as TAG_Byte_Array,
-    IntArrayTag as TAG_Int_Array,
-    LongArrayTag as TAG_Long_Array,
-)
-from ._tag.string import (
-    StringTag,
-    StringTag as TAG_String,
-)
-
-from ._tag.named_tag import NamedTag
-
-from ._tag.list import (
-    ListTag,
-    ListTag as TAG_List,
-)
-from ._tag.compound import (
-    CompoundTag,
-    CompoundTag as TAG_Compound,
-)
-
-# Load functions
-from amulet_nbt._nbt_encoding._binary import load, load_array, ReadOffset
-from amulet_nbt._nbt_encoding._string import from_snbt
-
-from ._errors import NBTError, NBTLoadError, NBTFormatError, SNBTParseError
-
-from ._string_encoding.encoding import (
-    mutf8_encoding,
-    utf8_encoding,
-    utf8_escape_encoding,
-    StringEncoding,
-)
-from ._nbt_encoding._binary.encoding_preset import (
-    java_encoding,
-    bedrock_encoding,
-    EncodingPreset,
-)
-
-SNBTType = str
-
-IntType = Union[ByteTag, ShortTag, IntTag, LongTag]
-
-FloatType = Union[FloatTag, DoubleTag]
-
-NumberType = Union[ByteTag, ShortTag, IntTag, LongTag, FloatTag, DoubleTag]
-
-ArrayType = Union[ByteArrayTag, IntArrayTag, LongArrayTag]
-
-AnyNBT = Union[
-    ByteTag,
-    ShortTag,
-    IntTag,
-    LongTag,
-    FloatTag,
-    DoubleTag,
-    ByteArrayTag,
-    StringTag,
-    ListTag,
-    CompoundTag,
-    IntArrayTag,
-    LongArrayTag,
-]