ericdwkim/fluent_python

Code snippets from Fluent Python

Reference Notes

Concepts & Takeaways

• Leverage built-in dunder methods instead of re-creating the wheel as new methods
• Abstract Base Classes (ABCs) --> enforces its subclasses to implement specific methods; ensures consistency and reliability of the interface
• Polymorphism --> provides flexibility and extensibility to parent class methods, thus ensuring interface uniformity

Terminology

• infix operators -> create new objs and not touching/modifying the operands

Example: This addition function simply reads the operands self and other , but does not modify them and ultimately returns the new object.

class Vector:
                ...
           
    def __add__(self, other):
        x = self.x + other.x
        y = self.y + other.y
        return Vector(x, y)

• Container sequences -> versatile; can be more memory intensive as it must hold both a ref and the obj in memory, but flexible in that it can hold memory of any dtype
• Flat sequences -> homogenous data (only single dtype allowed); less memory intensive; stores the value of each item

Special methods

__repr__ --> provides a helpful, string representation of an object for inspection

For example, the output Vector(4, 5) is descriptive of how the Vector class is defined. This output can also now be used as an input too. Without this custom method, we would see <Vector object at 0x...> which isn't helpful or descriptive.

    def __repr__(self):
        return f'Vector({self.x!r}, {self.y!r})'
    
    v1 = Vector(2, 4)
    v2 = Vector(2, 1)
    v3 = v1 + v2
    
>> Vector(4, 5)

Syntax

!r in the __repr__ allows Vector(2, 4) and Vector('2', '4') to be both valid instead of expecting the x, y values to be ints, so the returned string would output the more accurate dtypes of the args

Data Structures

• Dynamic arrays = lists (mutable)
• Fixed-size arrays = tuples (immutable)