Finding The Biggest Key In A Python Dictionary

Question

General:

I need help finding a way in python to get the max N items in a multi-dimensional python dictionary. For example:

 things = {
          "car": { "weight": 100 },
          "apple": { "weight": 1 },
          "spanner": { "weight": 10 } 
          }

In this case, I would want to find the 2 highest-weighted items in the dictionary, specifically the keys of these items. So in this case, it should return ["car", "spanner"]

Actual Problem:

Note: This is my first attempt at a genetic algorithm, so I might not be doing it correctly. At all.

As I am British, I am searching for the best cup of tea I can imagine, so I am writing a python program that generates 10 random cups of tea, then uses natural selection to find the top 5 in that ten and so on.

A cup of tea is modelled as a python dictionary, with 5 keys:

{
 "brew_time": Some Number,
 "milk": Some Number,
 "sweeteners": Some Number,
 "fitness": Some Number (This is what I'm interested in),
 "name": Some randomly generated name (Doesn't really matter)
}

A cup of tea my program will spit out will look something like this:

{'brew_time': 2.0, 'milk': 0.5, 'sweeteners': 3.0, 'name': 'bold cup', 'fitness': 0}

It then generates 10 cups of tea, stored in the teas variable. This is an example of an output of that:

{0: {'brew_time': 2.0, 'milk': 0.4, 'sweeteners': 1.0, 'name': 'unafraid brew', 'fitness': 0}, 1: {'brew_time': 3.0, 'milk': 0.5, 'sweeteners': 3.0, 'name': 'fire-eating blend', 'fitness': 0}, 2: {'brew_time': 2.0, 'milk': 0.6, 'sweeteners': 2.0, 'name': 'fearless drink', 'fitness': 0}, 3: {'brew_time': 2.0, 'milk': 0.9, 'sweeteners': 3.0, 'name': 'fire-eating blend', 'fitness': 0}, 4: {'brew_time': 2.0, 'milk': 0.8, 'sweeteners': 2.0, 'name': 'fire-eating cuppa', 'fitness': 0}, 5: {'brew_time': 3.0, 'milk': 0.3, 'sweeteners': 1.0, 'name': 'fire-eating drink', 'fitness': 0}, 6: {'brew_time': 4.0, 'milk': 0.7, 'sweeteners': 2.0, 'name': 'dauntless medley', 'fitness': 0}, 7: {'brew_time': 3.0, 'milk': 0.3, 'sweeteners': 2.0, 'name': 'dauntless cuppa', 'fitness': 0}, 8: {'brew_time': 3.0, 'milk': 0.9, 'sweeteners': 2.0, 'name': 'epic drink', 'fitness': 0}, 9: {'brew_time': 2.0, 'milk': 0.4, 'sweeteners': 2.0, 'name': 'gusty drink', 'fitness': 0}}

I'm now trying to code a function called selection() that will remove the 5 least fit teas from the dictionary. (The fitness of a tea is set by me, using the rank_tea() function, which takes an array and sets all the teas fitnesses, which is a number between 0 - 1 that represents the quality of the tea)

This is what I've got so far, but it doesn't work:

def selection():
    teaCopy = teas.copy()
    fitnesses = []
    for i in range(0, len(teaCopy)):
        fitnesses.append(teas[i]["fitness"])

    print(fitnesses)

    max_fitnesses_indicies = sorted(range(len(fitnesses)), key=lambda x: fitnesses[x])
    print(max_fitnesses_indicies)

    len_array = []
    print(len_array)
    for i in range(0, len(teas)):
        len_array.append(i)

    to_be_del = list( set(max_fitnesses_indicies) - set(len_array) )
    print(to_be_del)

This is the full code. Sorry for the length of the question, I just didn't want to miss anything.

Any help would be appreciated

Answer 1

You can simply use:

>>> sorted(things.keys(),key=lambda x:things[x]['weight'],reverse=True)
['car', 'spanner', 'apple']

To obtain a list of items sorted by their weight (here in reversed order such that the more heavy things are sorted first). So if you call:

>>> sorted(things.keys(),key=lambda x:things[x]['weight'],reverse=True)[:2]
['car', 'spanner']

you get the two heaviest. But this will run in O(n log n). In case the number of values k you wish to obtain is small (compared to the total number). You can use heapq:

from heapq import nlargest

result = nlargest(k,things.keys(),key=lambda x:things[x]['weight'])

which will - as far as I know - run in O(n log k) (k the numbers of items you want to pick).