feat: implement the Formatter API for python lib

[latonis]

• exposed the format API via the Python lib
• updated the typing stub
• documented the functions in Py
• added __pycache__ to be ignored
• test for python formatting

[plusvic]

I think the Python API for the formatter should mimic the Rust API as much as possible. For instance, the Formatter::format method should receive two file objects, one for reading the unformatted code and the other for writing the formatted code. This API allows using the Formatter::format method with in-memory data, but also with file objects, which is what most people will probably want, as they are formatting files in most cases. Additionally, this way the Python API doesn't need to make copies of the input or output data, it simply passes the file objects down to the yara-x-fmt API.

The serialize_into and deserialize_from methods can be used as examples of how to convert a Python's file object into a Rust reader or writer, that can be passed to the Rust API.

yara-x/py/src/lib.rs

Lines 528 to 548 in 4d59017

	fn serialize_into(&self, file: PyObject) -> PyResult<()> {
	let f = PyFileLikeObject::with_requirements(
	file, false, true, false, false,
	)?;
	self.inner
	.rules
	.serialize_into(f)
	.map_err(|err| PyIOError::new_err(err.to_string()))
	}
	/// Deserializes rules from a file-like object.
	#[staticmethod]
	fn deserialize_from(file: PyObject) -> PyResult<Py<Rules>> {
	let f = PyFileLikeObject::with_requirements(
	file, true, false, false, false,
	)?;
	let rules = yrx::Rules::deserialize_from(f)
	.map_err(|err| PyIOError::new_err(err.to_string()))?;
	Python::with_gil(|py| Py::new(py, Rules::new(rules)))
	}

[latonis]

I think the Python API for the formatter should mimic the Rust API as much as possible. For instance, the Formatter::format method should receive two file objects, one for reading the unformatted code and the other for writing the formatted code. This API allows using the Formatter::format method with in-memory data, but also with file objects, which is what most people will probably want, as they are formatting files in most cases. Additionally, this way the Python API doesn't need to make copies of the input or output data, it simply passes the file objects down to the yara-x-fmt API.

The serialize_into and deserialize_from methods can be used as examples of how to convert a Python's file object into a Rust reader or writer, that can be passed to the Rust API.

yara-x/py/src/lib.rs

Lines 528 to 548 in 4d59017

	fn serialize_into(&self, file: PyObject) -> PyResult<()> {
	let f = PyFileLikeObject::with_requirements(
	file, false, true, false, false,
	)?;
	self.inner
	.rules
	.serialize_into(f)
	.map_err(|err| PyIOError::new_err(err.to_string()))
	}
	/// Deserializes rules from a file-like object.
	#[staticmethod]
	fn deserialize_from(file: PyObject) -> PyResult<Py<Rules>> {
	let f = PyFileLikeObject::with_requirements(
	file, true, false, false, false,
	)?;
	let rules = yrx::Rules::deserialize_from(f)
	.map_err(|err| PyIOError::new_err(err.to_string()))?;
	Python::with_gil(|py| Py::new(py, Rules::new(rules)))
	}

Ahh good point😸 , fixed in 53bae8f