🍰 The little DI framework that tastes like a cake. 🍰
Documentation: https://fresh-bakery.readthedocs.io/en/latest/
Fresh bakery is a lightweight [Dependency Injection][DI] framework/toolkit, which is ideal for building object dependencies in Python.
It is [fully] production-ready, and gives you the following:
asyncio
, trio
).Mypy
compatible (no probably need for # type: ignore
).FastAPI
fully compatible.Litestar
compatible.Pytest
fully compatible (Fresh Bakery encourages the use of pytest
).Python 3.8+
$ pip3 install fresh-bakery
This example is intended to show the nature of Dependency Injection and the ease of use the library. Many of us work 8 hours per day on average, 5 days a week, i.e. ~ 40 hours per week. Let's describe it using DI and bakery:
from bakery import Bakery, Cake
def full_days_in(hours: int) -> float:
return hours / 24
def average(total: int, num: int) -> float:
return total / num
class WorkingBakery(Bakery):
average_hours: int = Cake(8)
week_hours: int = Cake(sum, [average_hours, average_hours, 7, 9, average_hours])
full_days: float = Cake(full_days_in, week_hours)
async def main() -> None:
async with WorkingBakery() as bakery:
assert bakery.week_hours == 40
assert bakery.full_days - 0.00001 < full_days_in(40)
assert int(bakery.average_hours) == 8
You can see it's as simple as it can be.
Let's suppose we have a thin wrapper around file object.
from typing import ClassVar, Final
from typing_extensions import Self
class FileWrapper:
file_opened: bool = False
write_lines: ClassVar[list[str]] = []
def __init__(self, filename: str) -> None:
self.filename: Final = filename
def write(self, line: str) -> int:
type(self).write_lines.append(line)
return len(line)
def __enter__(self) -> Self:
type(self).file_opened = True
return self
def __exit__(self, *_args: object) -> None:
type(self).file_opened = False
type(self).write_lines.clear()
This wrapper acts exactly like a file object: it can be opened, closed, and can write line to file.
Let's open file hello.txt
, write 2 lines into it and close it. Let's do all this with the bakery syntax:
from bakery import Bakery, Cake
class FileBakery(Bakery):
_file_obj: FileWrapper = Cake(FileWrapper, "hello.txt")
file_obj: FileWrapper = Cake(_file_obj)
write_1_bytes: int = Cake(file_obj.write, "hello, ")
write_2_bytes: int = Cake(file_obj.write, "world")
async def main() -> None:
assert FileWrapper.file_opened is False
assert FileWrapper.write_lines == []
async with FileBakery() as bakery:
assert bakery.file_obj.filename == "hello.txt"
assert FileWrapper.file_opened is True
assert FileWrapper.write_lines == ["hello, ", "world"]
assert FileWrapper.file_opened is False
assert FileWrapper.write_lines == []
Maybe you noticed some strange things concerning FileBakery
bakery:
_file_obj
and file_obj
objects. Do we need them both?write_1_bytes
and write_2_bytes
objects. Do we need them?Let's try to fix both cases. First, let's figure out why do we need _file_obj
and file_obj
objects?
Cake
for _file_obj
initiates FileWrapper
object, i.e. calls __init__
method;Cake
for file_obj
calls context-manager, i.e. calls __enter__
method on enter and __exit__
method on exit.Actually, we can merge these two statements into single one:
# class FileBakery(Bakery):
file_obj: FileWrapper = Cake(Cake(FileWrapper, "hello.txt"))
So, what about unused arguments? OK, let's re-write this gist a little bit. First, let's declare the list of strings we want to write:
# class FileBakery(Bakery):
strs_to_write: list[str] = Cake(["hello, ", "world"])
How to apply function to every string in this list? There are several ways to do it. One of them is built-in map
function.
map_cake = Cake(map, file_obj.write, strs_to_write)
But map
function returns iterator and we need to get elements from it. Built-in list
function will do the job.
list_cake = Cake(list, map_cake)
In the same manner as we did for file_obj
let's merge these two statements into one. The final FileBakery
will look like this:
class FileBakeryMap(Bakery):
file_obj: FileWrapper = Cake(Cake(FileWrapper, "hello.txt"))
strs_to_write: list[str] = Cake(["hello, ", "world"])
_: list[int] = Cake(list, Cake(map, file_obj.write, strs_to_write))
The last thing nobody likes is hard-coded strings! In this case such strings are:
hello.txt
hello,
and world
What if we've got another filename or other strings to write? Let's define filename and list of strings as FileBakery
parameters:
from bakery import Bakery, Cake, __Cake__
class FileBakery(Bakery):
filename: str = __Cake__()
strs_to_write: list[str] = __Cake__()
file_obj: FileWrapper = Cake(Cake(FileWrapper, filename))
_: list[int] = Cake(list, Cake(map, file_obj.write, strs_to_write))
To define parameters you can use dunder-cake construction: __Cake__()
.
To pass arguments into FileBakery
you can use native python syntax:
# async def main() -> None:
async with FileBakeryMapWithParams(
filename="hello.txt", strs_to_write=["hello, ", "world"]
) as bakery:
...
And the whole example will look like this:
from typing import ClassVar, Final
from typing_extensions import Self
from bakery import Bakery, Cake, __Cake__
# class FileWrapper: ...
class FileBakery(Bakery):
filename: str = __Cake__()
strs_to_write: list[str] = __Cake__()
file_obj: FileWrapper = Cake(Cake(FileWrapper, filename))
_: list[int] = Cake(list, Cake(map, file_obj.write, strs_to_write))
async def main() -> None:
assert FileWrapper.file_opened is False
assert FileWrapper.write_lines == []
async with FileBakeryMapWithParams(
filename="hello.txt", strs_to_write=["hello, ", "world"]
) as bakery:
assert bakery.file_obj.filename == "hello.txt"
assert FileWrapper.file_opened is True
assert FileWrapper.write_lines == ["hello, ", "world"]
assert FileWrapper.file_opened is False
assert FileWrapper.write_lines == []
More examples are presented in section bakery examples.
No dependencies ;)
You can see the release history here: https://github.com/Mityuha/fresh-bakery/releases/
Fresh Bakery is MIT licensed code.