pyo3/Architecture.md

166 lines
7.3 KiB
Markdown
Raw Normal View History

<!-- This file contains a rough overview of the PyO3 codebase. -->
2021-02-20 09:37:57 +00:00
<!-- Please do not make descriptions too specific, so that we can easily -->
<!-- keep this file in sync with the codebase. -->
# PyO3: Architecture.md
This document roughly describes the high-level architecture of PyO3.
If you want to become familiar with the codebase, you are in the right place!
2021-02-20 09:37:57 +00:00
## Overview
PyO3 provides a bridge between Rust and Python, based on the [Python C/API].
Thus, PyO3 has low-level bindings of these API as its core.
On top of that, we have higher-level bindings to operate Python objects safely.
Also, to define Python classes and functions in Rust code, we have `trait PyClass<T>` and a set of
2021-02-20 09:37:57 +00:00
protocol traits (e.g., `PyIterProtocol`) for supporting object protocols (i.e., `__dunder__` methods).
Since implementing `PyClass` requires lots of boilerplates, we have a proc-macro `#[pyclass]`.
To summarize, we have mainly four parts in the PyO3 codebase.
1. Low-level bindings of Python C/API.
- [`src/ffi`]
2. Bindings to Python objects.
- [`src/instance.rs`], [`src/types`]
3. `PyClass<T>` and related functionalities
- [`src/pycell.rs`], [`src/pyclass.rs`]
4. Protocol methods like `__getitem__`.
- [`src/class`]
5. Defining a Python class requires lots of glue code, so we provide proc-macros to simplify the procedure.
2021-02-20 09:37:57 +00:00
- [`src/derive_utils.rs`]
- [`pyo3-macros`], [`pyo3-macros-backend`]
## Low-level bindings of CPython API
[`src/ffi`] contains wrappers of [Python C/API](https://docs.python.org/3/c-api/).
We aim to provide straight-forward Rust wrappers resembling the file structure of
2021-02-20 09:37:57 +00:00
[`cpython/Include`](https://github.com/python/cpython/tree/v3.9.2/Include).
However, we still lack some API and continue to refactor the module to completely resemble
the CPython's file structure.
The tracking issue is [#1289](https://github.com/PyO3/pyo3/issues/1289), and contribution is welcome.
## Bindings to Python Objects
[`src/types`] contains bindings to [built-in types](https://docs.python.org/3/library/stdtypes.html)
of Python, such as `dict` and `list`.
Due to historical reasons, Python's `object` is called `PyAny` and placed in [`src/types/any.rs`].
2021-02-20 09:37:57 +00:00
Currently, `PyAny` is a straight-forward wrapper of `ffi::PyObject`, like:
```rust
#[repr(transparent)]
pub struct PyAny(UnsafeCell<ffi::PyObject>);
```
All built-in types are defined as a C struct.
For example, `dict` is defined as:
```c
typedef struct {
/* Base object */
PyObject ob_base;
/* Number of items in the dictionary */
Py_ssize_t ma_used;
/* Dictionary version */
uint64_t ma_version_tag;
PyDictKeysObject *ma_keys;
PyObject **ma_values;
} PyDictObject;
```
However, we cannot access such a specific data structure with `#[cfg(Py_LIMITED_API)]` set.
Thus, all builtin objects are implemented as opaque types by wrapping `PyAny`, like:
```rust
#[repr(transparent)]
pub struct PyDict(PyAny);
```
Note that `PyAny` is not a pointer, and it is usually used as a pointer to the object in the
Python heap, as `&PyAny`.
This design choice can be changed
(see the discussion in [#1056](https://github.com/PyO3/pyo3/issues/1056)).
Since we need lots of boilerplate for implementing common traits for these types
2021-02-20 09:37:57 +00:00
(e.g., `AsPyPointer`, `AsRef<PyAny>`, and `Debug`), we have some macros in
[`src/types/mod.rs`].
## PyClass
[`src/pycell.rs`], [`src/pyclass.rs`], and [`src/type_object.rs`] contains types and
traits to make `#[pyclass]` work.
Also, [`src/pyclass_init.rs`] and [`src/pyclass_slots.rs`] have related functionalities.
To realize object-oriented programming in C, all Python objects must have the following two fields
at the beginning.
```rust
#[repr(C)]
pub struct PyObject {
pub ob_refcnt: usize,
pub ob_type: *mut PyTypeObject,
...
}
```
Thanks to this guarantee, casting `*mut A` to `*mut PyObject` is valid if `A` is a Python object.
2021-02-20 09:37:57 +00:00
To ensure this guarantee, we have a wrapper struct `PyCell<T>` in [`src/pycell.rs`] which is roughly:
2021-02-20 09:37:57 +00:00
```rust
#[repr(C)]
pub struct PyCell<T: PyClass> {
object: crate::ffi::PyObject,
inner: T,
}
```
Thus, when copying a Rust struct to a Python object, we first allocate `PyCell` on the Python heap and then
copy `T`.
2021-02-20 09:37:57 +00:00
Also, `PyCell` provides [RefCell](https://doc.rust-lang.org/std/cell/struct.RefCell.html)-like methods
to ensure Rust's borrow rules.
See [the document](https://docs.rs/pyo3/latest/pyo3/pycell/struct.PyCell.html) for more.
`PyCell<T>` requires that `T` implements `PyClass`.
This trait is somewhat complex and derives many traits, but the most important one is `PyTypeObject`
in [`src/type_object.rs`].
`PyTypeObject` is also implemented for built-in types.
Type objects are singletons, and all Python types have their unique type objects.
2021-02-20 09:37:57 +00:00
For example, you can see `type({})` shows `dict` and `type(type({}))` shows `type` in Python REPL.
`T: PyTypeObject` implies that `T` has a corresponding type object.
2021-02-20 09:37:57 +00:00
## Protocol methods
Python has some built-in special methods called dunder, such as `__iter__`.
They are called [abstract objects layer](https://docs.python.org/3/c-api/abstract.html) in
Python C/API.
We provide a way to implement those protocols by using `#[pyproto]` and specific traits, such
as `PyIterProtocol`.
[`src/class`] defines these traits.
Each protocol method has a corresponding FFI function.
For example, `PyIterProtocol::__iter__` has
2021-02-20 09:37:57 +00:00
`pub unsafe extern "C" fn iter<T>(slf: *mut PyObject) -> *mut PyObject`.
When `#[pyproto]` finds that `T` implements `PyIterProtocol::__iter__`, it automatically
sets `iter<T>` on the type object of `T`.
2021-02-20 09:37:57 +00:00
Also, [`src/class/methods.rs`] has utilities for `#[pyfunction]` and [`src/class/impl_.rs`] has
some internal tricks for making `#[pyproto]` flexible.
## Proc-macros
[`pyo3-macros`] provides six proc-macro APIs: `pymodule`, `pyproto`, `pyfunction`, `pyclass`,
`pymethods`, and `#[derive(FromPyObject)]`.
[`pyo3-macros-backend`] has the actual implementations of these APIs.
2021-02-20 09:37:57 +00:00
[`src/derive_utils.rs`] contains some utilities used in codes generated by these proc-macros,
such as parsing function arguments.
<!-- External Links -->
[Python C/API](https://docs.python.org/3/c-api/).
<!-- Crates -->
[`pyo3-macros`]: (https://github.com/PyO3/pyo3/tree/master/pyo3-macros)
[`pyo3-macros-backend`]: (https://github.com/PyO3/pyo3/tree/master/pyo3-macros-backend)
<!-- Directories -->
[`src/class`]: https://github.com/PyO3/pyo3/tree/master/src/class
[`src/ffi`]: https://github.com/PyO3/pyo3/tree/master/src/ffi
[`src/types`]: https://github.com/PyO3/pyo3/tree/master/src/types
<!-- Files -->
[`src/derive_utils.rs`]: https://github.com/PyO3/pyo3/tree/master/src/derive_utils.rs
[`src/instance.rs`]: https://github.com/PyO3/pyo3/tree/master/src/instance.rs
[`src/pycell.rs`]: https://github.com/PyO3/pyo3/tree/master/src/pycell.rs
[`src/pyclass.rs`]: https://github.com/PyO3/pyo3/tree/master/src/pyclass.rs
[`src/pyclass_init.rs`]: https://github.com/PyO3/pyo3/tree/master/src/pyclass_init.rs
[`src/pyclass_slot.rs`]: https://github.com/PyO3/pyo3/tree/master/src/pyclass_slot.rs
[`src/type_object.rs`]: https://github.com/PyO3/pyo3/tree/master/src/type_object.rs
[`src/class/methods.rs`]: https://github.com/PyO3/pyo3/tree/master/src/class/methods.rs
[`src/class/impl_.rs`]: https://github.com/PyO3/pyo3/tree/master/src/class/impl_.rs
[`src/types/any.rs`]: https://github.com/PyO3/pyo3/tree/master/src/types/any.rs
[`src/types/mod.rs`]: https://github.com/PyO3/pyo3/tree/master/src/types/mod.rs