knapsack: implement exercise

From metadata:

Given a knapsack that can only carry a certain weight, determine which items to put in the knapsack in order to maximize their combined value.

Canonical Data

Contributor Checklist:

• Clone exercism/problem-specifications to local $SPEC_DIRECTORY
• Make new directory $PYTHON_TRACK_DIRECTORY/exercises/knapsack
• In $PYTHON_TRACK_DIRECTORY, run bin/get-configlet
• Add entry for knapsack in $PYTHON_TRACK_DIRECTORY/config.json (use bin/configlet uuid to generate a new UUID for entry)
• Run $PYTHON_TRACK_DIRECTORY/bin/configlet generate -p $SPEC_DIRECTORY -O knapsack $PYTHON_TRACK_DIRECTORY to generate the exercise README.md
• Implement test suite $PYTHON_TRACK_DIRECTORY/exercises/knapsack/knapsack_test.py (use existing exercise ~coins~ change as an example)
• Create a solution stub $PYTHON_TRACK_DIRECTORY/exercises/knapsack/knapsack.py
• Implement example solution $PYTHON_TRACK_DIRECTORY/exercises/knapsack/example.py. This solution may be tested using $PYTHON_TRACK_DIRECTORY/test/check-exercises.py knapsack

[cmccandless]

When I wrote my implementation, I used a list of dictionaries. I tend to favor mappings for various reasons.

As in {"weight": 2, "value": 3}? That is fine by me.

In addition, they are an allowed json type, so it makes automatic test generation easier.

Nested lists are allowed in JSON; they just have to be homogeneous. That said, I think the dict approach is clearer.

Also, another note from the bug text, but I don’t think coins is an exercise in this track. Perhaps you meant change?

Yes, my mistake. I updated the issue description.

[AnAccountForReportingBugs]

When I wrote my implementation, I used a list of dictionaries. I tend to favor mappings for various reasons. In addition, they are an allowed json type, so it makes automatic test generation easier.

Also, another note from the bug text, but I don’t think coins is an exercise in this track. Perhaps you meant change?