6. 在 iOS 上使用 Python¶
- 作者:
Russell Keith-Magee (2024-03)
Python on iOS is unlike Python on desktop platforms. On a desktop platform, Python is generally installed as a system resource that can be used by any user of that computer. Users then interact with Python by running a python executable and entering commands at an interactive prompt, or by running a Python script.
On iOS, there is no concept of installing as a system resource. The only unit of software distribution is an "app". There is also no console where you could run a python executable, or interact with a Python REPL.
As a result, the only way you can use Python on iOS is in embedded mode - that
is, by writing a native iOS application, and embedding a Python interpreter
using libPython
, and invoking Python code using the Python embedding
API. The full Python interpreter, the standard library, and all
your Python code is then packaged as a standalone bundle that can be
distributed via the iOS App Store.
If you're looking to experiment for the first time with writing an iOS app in Python, projects such as BeeWare and Kivy will provide a much more approachable user experience. These projects manage the complexities associated with getting an iOS project running, so you only need to deal with the Python code itself.
6.1. iOS 上的 Python 运行时¶
6.1.1. 平台识别¶
When executing on iOS, sys.platform
will report as ios
. This value will
be returned on an iPhone or iPad, regardless of whether the app is running on
the simulator or a physical device.
Information about the specific runtime environment, including the iOS version,
device model, and whether the device is a simulator, can be obtained using
platform.ios_ver()
. platform.system()
will report iOS
or
iPadOS
, depending on the device.
os.uname()
reports kernel-level details; it will report a name of
Darwin
.
6.1.2. 标准库可用性¶
The Python standard library has some notable omissions and restrictions on iOS. See the API availability guide for iOS for details.
6.1.3. 二进制扩展模块¶
One notable difference about iOS as a platform is that App Store distribution imposes hard requirements on the packaging of an application. One of these requirements governs how binary extension modules are distributed.
The iOS App Store requires that all binary modules in an iOS app must be
dynamic libraries, contained in a framework with appropriate metadata, stored
in the Frameworks
folder of the packaged app. There can be only a single
binary per framework, and there can be no executable binary material outside
the Frameworks
folder.
This conflicts with the usual Python approach for distributing binaries, which
allows a binary extension module to be loaded from any location on
sys.path
. To ensure compliance with App Store policies, an iOS project must
post-process any Python packages, converting .so
binary modules into
individual standalone frameworks with appropriate metadata and signing. For
details on how to perform this post-processing, see the guide for adding
Python to your project.
To help Python discover binaries in their new location, the original .so
file on sys.path
is replaced with a .fwork
file. This file is a text
file containing the location of the framework binary, relative to the app
bundle. To allow the framework to resolve back to the original location, the
framework must contain a .origin
file that contains the location of the
.fwork
file, relative to the app bundle.
例如,考虑导入 from foo.bar import _whiz
的情况,其中 _whiz
是使用二进制模块 sources/foo/bar/_whiz.abi3.so
实现的,这里 sources
是在 sys.path
中注册的相对于 app 包的位置。 此模块 必须 发布为 Frameworks/foo.bar._whiz.framework/foo.bar._whiz
(根据模块的完整导入路径创建框架名称),并通过 .framework
目录中的 Info.plist
文件将二进制文件标识为一个框架。 foo.bar._whiz
模块在原始位置中以一个 sources/foo/bar/_whiz.abi3.fwork
标记文件来代表,其中包含路径 Frameworks/foo.bar._whiz/foo.bar._whiz
。 该框架还要包含 Frameworks/foo.bar._whiz.framework/foo.bar._whiz.origin
,其中包含 .fwork
文件的路径。
When running on iOS, the Python interpreter will install an
AppleFrameworkLoader
that is able to read and
import .fwork
files. Once imported, the __file__
attribute of the
binary module will report as the location of the .fwork
file. However, the
ModuleSpec
for the loaded module will report the
origin
as the location of the binary in the framework folder.
6.1.4. 编译器存根二进制文件¶
Xcode doesn't expose explicit compilers for iOS; instead, it uses an xcrun
script that resolves to a full compiler path (e.g., xcrun --sdk iphoneos
clang
to get the clang
for an iPhone device). However, using this script
poses two problems:
The output of
xcrun
includes paths that are machine specific, resulting in a sysconfig module that cannot be shared between users; andIt results in
CC
/CPP
/LD
/AR
definitions that include spaces. There is a lot of C ecosystem tooling that assumes that you can split a command line at the first space to get the path to the compiler executable; this isn't the case when usingxcrun
.
To avoid these problems, Python provided stubs for these tools. These stubs are
shell script wrappers around the underingly xcrun
tools, distributed in a
bin
folder distributed alongside the compiled iOS framework. These scripts
are relocatable, and will always resolve to the appropriate local system paths.
By including these scripts in the bin folder that accompanies a framework, the
contents of the sysconfig
module becomes useful for end-users to compile
their own modules. When compiling third-party Python modules for iOS, you
should ensure these stub binaries are on your path.
6.2. 在 iOS 上安装 Python¶
6.2.1. 构建 iOS 应用程序的工具¶
Building for iOS requires the use of Apple's Xcode tooling. It is strongly recommended that you use the most recent stable release of Xcode. This will require the use of the most (or second-most) recently released macOS version, as Apple does not maintain Xcode for older macOS versions. The Xcode Command Line Tools are not sufficient for iOS development; you need a full Xcode install.
If you want to run your code on the iOS simulator, you'll also need to install an iOS Simulator Platform. You should be prompted to select an iOS Simulator Platform when you first run Xcode. Alternatively, you can add an iOS Simulator Platform by selecting from the Platforms tab of the Xcode Settings panel.
6.2.2. 在 iOS 项目中添加 Python¶
Python can be added to any iOS project, using either Swift or Objective C. The following examples will use Objective C; if you are using Swift, you may find a library like PythonKit to be helpful.
要在 iOS Xcode 项目中添加 Python:
Build or obtain a Python
XCFramework
. See the instructions in iOS/README.rst (in the CPython source distribution) for details on how to build a PythonXCFramework
. At a minimum, you will need a build that supportsarm64-apple-ios
, plus one of eitherarm64-apple-ios-simulator
orx86_64-apple-ios-simulator
.Drag the
XCframework
into your iOS project. In the following instructions, we'll assume you've dropped theXCframework
into the root of your project; however, you can use any other location that you want by adjusting paths as needed.Drag the
iOS/Resources/dylib-Info-template.plist
file into your project, and ensure it is associated with the app target.Add your application code as a folder in your Xcode project. In the following instructions, we'll assume that your user code is in a folder named
app
in the root of your project; you can use any other location by adjusting paths as needed. Ensure that this folder is associated with your app target.Select the app target by selecting the root node of your Xcode project, then the target name in the sidebar that appears.
In the "General" settings, under "Frameworks, Libraries and Embedded Content", add
Python.xcframework
, with "Embed & Sign" selected.在"构建设置"选项卡中,修改以下内容:
构建选项
用户脚本沙盒:否
启用可测试性:是
搜索路径
框架搜索路径:
$(PROJECT_DIR)
头文件搜索路径:
"$(BUILT_PRODUCTS_DIR)/Python.framework/Headers"
Apple Clang - 警告 - 所有语言
引用包含在框架头文件中:否
Add a build step that copies the Python standard library into your app. In the "Build Phases" tab, add a new "Run Script" build step before the "Embed Frameworks" step, but after the "Copy Bundle Resources" step. Name the step "Install Target Specific Python Standard Library", disable the "Based on dependency analysis" checkbox, and set the script content to:
set -e mkdir -p "$CODESIGNING_FOLDER_PATH/python/lib" if [ "$EFFECTIVE_PLATFORM_NAME" = "-iphonesimulator" ]; then echo "Installing Python modules for iOS Simulator" rsync -au --delete "$PROJECT_DIR/Python.xcframework/ios-arm64_x86_64-simulator/lib/" "$CODESIGNING_FOLDER_PATH/python/lib/" else echo "Installing Python modules for iOS Device" rsync -au --delete "$PROJECT_DIR/Python.xcframework/ios-arm64/lib/" "$CODESIGNING_FOLDER_PATH/python/lib/" fi
Note that the name of the simulator "slice" in the XCframework may be different, depending the CPU architectures your
XCFramework
supports.Add a second build step that processes the binary extension modules in the standard library into "Framework" format. Add a "Run Script" build step directly after the one you added in step 8, named "Prepare Python Binary Modules". It should also have "Based on dependency analysis" unchecked, with the following script content:
set -e install_dylib () { INSTALL_BASE=$1 FULL_EXT=$2 # The name of the extension file EXT=$(basename "$FULL_EXT") # The location of the extension file, relative to the bundle RELATIVE_EXT=${FULL_EXT#$CODESIGNING_FOLDER_PATH/} # The path to the extension file, relative to the install base PYTHON_EXT=${RELATIVE_EXT/$INSTALL_BASE/} # The full dotted name of the extension module, constructed from the file path. FULL_MODULE_NAME=$(echo $PYTHON_EXT | cut -d "." -f 1 | tr "/" "."); # A bundle identifier; not actually used, but required by Xcode framework packaging FRAMEWORK_BUNDLE_ID=$(echo $PRODUCT_BUNDLE_IDENTIFIER.$FULL_MODULE_NAME | tr "_" "-") # The name of the framework folder. FRAMEWORK_FOLDER="Frameworks/$FULL_MODULE_NAME.framework" # If the framework folder doesn't exist, create it. if [ ! -d "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER" ]; then echo "Creating framework for $RELATIVE_EXT" mkdir -p "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER" cp "$CODESIGNING_FOLDER_PATH/dylib-Info-template.plist" "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER/Info.plist" plutil -replace CFBundleExecutable -string "$FULL_MODULE_NAME" "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER/Info.plist" plutil -replace CFBundleIdentifier -string "$FRAMEWORK_BUNDLE_ID" "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER/Info.plist" fi echo "Installing binary for $FRAMEWORK_FOLDER/$FULL_MODULE_NAME" mv "$FULL_EXT" "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER/$FULL_MODULE_NAME" # Create a placeholder .fwork file where the .so was echo "$FRAMEWORK_FOLDER/$FULL_MODULE_NAME" > ${FULL_EXT%.so}.fwork # Create a back reference to the .so file location in the framework echo "${RELATIVE_EXT%.so}.fwork" > "$CODESIGNING_FOLDER_PATH/$FRAMEWORK_FOLDER/$FULL_MODULE_NAME.origin" } PYTHON_VER=$(ls -1 "$CODESIGNING_FOLDER_PATH/python/lib") echo "Install Python $PYTHON_VER standard library extension modules..." find "$CODESIGNING_FOLDER_PATH/python/lib/$PYTHON_VER/lib-dynload" -name "*.so" | while read FULL_EXT; do install_dylib python/lib/$PYTHON_VER/lib-dynload/ "$FULL_EXT" done # Clean up dylib template rm -f "$CODESIGNING_FOLDER_PATH/dylib-Info-template.plist" echo "Signing frameworks as $EXPANDED_CODE_SIGN_IDENTITY_NAME ($EXPANDED_CODE_SIGN_IDENTITY)..." find "$CODESIGNING_FOLDER_PATH/Frameworks" -name "*.framework" -exec /usr/bin/codesign --force --sign "$EXPANDED_CODE_SIGN_IDENTITY" ${OTHER_CODE_SIGN_FLAGS:-} -o runtime --timestamp=none --preserve-metadata=identifier,entitlements,flags --generate-entitlement-der "{}" \;
Add Objective C code to initialize and use a Python interpreter in embedded mode. You should ensure that:
UTF-8 模式
被 启用;
带缓冲的 stdio
被 禁用;
写入字节码
被 禁用;
信号处理器
被 启用;
PYTHONHOME
for the interpreter is configured to point at thepython
subfolder of your app's bundle; andThe
PYTHONPATH
for the interpreter includes:
the
python/lib/python3.X
subfolder of your app's bundle,the
python/lib/python3.X/lib-dynload
subfolder of your app's bundle, andthe
app
subfolder of your app's bundleYour app's bundle location can be determined using
[[NSBundle mainBundle] resourcePath]
.
Steps 8, 9 and 10 of these instructions assume that you have a single folder of
pure Python application code, named app
. If you have third-party binary
modules in your app, some additional steps will be required:
You need to ensure that any folders containing third-party binaries are either associated with the app target, or copied in as part of step 8. Step 8 should also purge any binaries that are not appropriate for the platform a specific build is targeting (i.e., delete any device binaries if you're building an app targeting the simulator).
Any folders that contain third-party binaries must be processed into framework form by step 9. The invocation of
install_dylib
that processes thelib-dynload
folder can be copied and adapted for this purpose.If you're using a separate folder for third-party packages, ensure that folder is included as part of the
PYTHONPATH
configuration in step 10.