nexus, index source, nexus main

Python

• Recommendation: The Python Data Science Handbook by Jake Vander Plas

In this directory

• Jupyter

Topics covered here

A wish list… topics I’d like to understand

• ‘hashable’ means an object is immutable and comparable (more below)
• iterable
• dictionaries: more than an inconvenient version of a list
• yield
• decorators
• class under what circumstances? (see manim Scene)

hashing

• ‘hashable’ means an object is immutable and comparable
  • First: It is immutable and subject to hashing (object has a __hash__() method)
    • to hash means to create a unique numerical identifier for an object
    • mutable non-hashables include lists, dictionaries and sets
    • hashable objects include strings, floats, ints, tuples (if containing immutables), frozen sets
    • use hash(som_object) to raise an error if it is not hashable
  • Second: It can be compared to other objects (has a __eq__() and/or __cmp__() method)

This code:

print('this string', hash('this string'))
print(10, hash(10))
n = 10
print(n, hash(n))
n = 10.0
print(n, hash(n))
n = 10.1
print(n, hash(n))

produces

this string 8534039767811493564
10 10
10 10
10.0 10
10.1 230584300921368586

So 10^20 is roughly the dynamic range apparent in Python hashing. This is the case as well for larger objects such as a string of 600 characters. There are say (10^2)^20 possible strings of length 20… so surely it is possible to have hash collisions. These do happen (and the birthday paradox is relevant). There are two ideas to note in consequence:

• Hashing is not encoding; it is labeling. There is no intent to reconstruct the value being hashed from the hash
• Hashing collisions can raise issues like how to deal with storage/retrieval; but this is Someone Elses Problem.