150 coding challenges in 25 days

What I learned from attempting Advent of Code 2017, 2016 & 2015 all at one go.

Yong Jun

24 Jan 2018

About me

Advent of Code a series of small programming puzzles for a variety of skill levels.

 

They are self-contained and are just as appropriate for an expert who wants to stay sharp as they are for a beginner who is just learning to code.

 

Each puzzle calls upon different skills and has two parts that build on a theme.

Leaderboard

Personal Stat

Global Stat

If you need a little help

Some things I've learned

Generators

function* generator () {
  // do something
  yield 'someValue'

  // do some more thing
  yield 'anotherValue'
}


const iterator = generator()

console.log(iterator.next())
// {value: 'someValue', done: false}

console.log(iterator.next())
// {value: 'anotherValue', done: false}

console.log(iterator.next())
// {value: undefined, done: true}

AoC 2017 Day 3: Spiral Memory

Right x 1

Up x 1

Left x 2

Down x 2

Right x 3

Up x 3

Left x 4

Down x 4

function * traverse () {
  let step = 1
  let x = 0
  let y = 0
  yield [x, y]
  while (true) {
    for (let right = step; right > 0; right--) {
      yield [++x, y]
    }
    for (let up = step; up > 0; up--) {
      yield [x, ++y]
    }
    step++
    for (let left = step; left > 0; left--) {
      yield [--x, y]
    }
    for (let down = step; down > 0; down--) {
      yield [x, --y]
    }
    step++
  }
}

function manhattanDistance (target) {
  let coordinates
  const location = traverse()

  for (let i = 0; i < target; i++) {
    coordinates = location.next().value
  }
  return Math.abs(coordinates[0]) + Math.abs(coordinates[1])
}

Solution

Why use a generator?

  • Rather than create an object that maintain internal state by variables assignment (explicit)

  • Internal state can be implicit in the position of the yield statement
  • Behave sort of like the goto statement from old school imperative programming

Another example

AoC 2015 Day 14:

Reindeer Olympics

function * reindeer (speed, fly, rest) {
  let distance = 0
  while (true) {
    for (let i = 0; i < fly; i++) {
      distance += speed
      yield distance
    }
    for (let i = 0; i < rest; i++) {
      yield distance
    }
  }
}

RegEx

AoC 2016 Day 7:

Internet Protocol Version 7

Task:

  Find ABBA

 

Solution:

function containsAbba (str) {
  const match = str.match(/([a-z])([a-z])\2\1/)
  if (!match) return false
  if (match[1] !== match[2]) return true
  return containsAbba(str.slice(match.index + 1))
}

Tail Call

Optimization

AoC 2015 Day 10:

Elves Look, Elves Say

Solution

// no proper tail call

function lookNsay (str) {
  const match = str.match(/(.)(\1*)/)
  if (match[0].length === str.length) return match[0].length + match[1]
  return match[0].length + match[1] + lookNsay(str.slice(match[0].length))
}

// with proper tail call

function lookNsay (str, prefix = '') {
  const match = str.match(/(.)(\1*)/)
  if (match[0].length === str.length) return prefix + match[0].length + match[1]
  return lookNsay(str.slice(match[0].length), prefix + match[0].length + match[1])
}

Limitations

  • 'use strict'
  • --harmony
  • Only node 6 & 7

Workaround

  • rewrite recursion into iteration
  • use a stack
  • eg. Depth-first-search

Enumeration

AoC 2015 Day 17:

No Such Thing as Too Much

Task:

  Given containers of various sizes (eg. 20, 15, 10, 5, and 5).

  Find how many different combinations total up to a given capacity.

  For example if required total is 25, four ways are possible:

  • 15 and 10
  • 20 and 5 (the first 5)
  • 20 and 5 (the second 5)
  • 15, 5, and 5

Solution:

  Enumerate all possible assignment combination

  Eg. 000, 001, 010, 011, 100, 101, 110, 111

AoC 2015 Day 9:

All in a Single Night

Task:

  Given distances between each location pair, find the

  shortest route where every location is visited exactly once.

 

Solution:

  Enumerate all possible permutation

  Eg. ABC, ACB, BAC, BCA, CAB, CBA

AoC 2015 Day 13:

Knights of the Dinner Table

Task:

  Find the ideal seating arrangement around a round

  dining table

 

Solution:

  Fix a start point and enumerate all possible

  permutation that start and end with that point. Eg.

  • A > B > C > D > A
  • A > B > D > C > A
  • A > C > B > D > A
  • A > C > D > B > A

AoC 2015 Day 15:

Science for Hungry People

Task:

  With exactly 100 teaspoons of ingredients. Find

  the ideal recipe

 

Solution:

  Enumeration all possible split combinations. Eg.

0/0/100
0/1/99 1/0/99
0/2/98 1/2/98 2/1/98
0/3/97 1/2/97 2/1/97 3/0/97

Enumerate Assignment Combinations

  • Count up from 0 to
  • Convert number to binary representation using Number.prototype.toString(2)
  • Left pad string with zeros
  • Convert to array using String.prototype.split('')
2^n - 1
function getAssignments (items, groups = 2) {
  const combinations = []
  const nCombinations = Math.pow(groups, items)
  const zeroPad = '0'.repeat(items)
  for (let i = 0; i < nCombinations; i++) {
    const combiString = (zeroPad + i.toString(groups)).slice(-items)
    const combination = combiString.split('')
    if (groups === 2) combinations.push(combination.map(v => +v))
    else combinations.push(combination)
  }
  return combinations
}

Implementation

Enumerate Split Combination

  • Instead of enumerating units, enumerate cuts. Eg.                        

 

 

 

 

 

  • If you have k different ingredients, the number of cuts to enumerate will be k - 1
  • Pick any position that is between last cut's position and N inclusive
  • Cuts can overlap
  • Subtract adjacent cut index to get units of ingredient
  • i.e.
1 1 1 1 1 1 1 1 1 1

First cut is always at 0

Last cut is always at N

Second cut at 2

Third cut also at 2

Fourth cut at 5

unit_i = cut_i - cut_{i-1}

Implementation

function getSplitCombinations (total, items) {
  const combinations = []
  function recurse (division, i) {
    if (division.length < items - 1) {
      while (i <= total) {
        recurse(division.concat(i), i)
        i++
      }
    } else {
      division = [0, ...division, total]
      const combination = []
      for (let i = 1; i < division.length; i++) {
        combination.push(division[i] - division[i - 1])
      }
      combinations.push(combination)
    }
  }
  recurse([], 0)
  return combinations
}

More enumeration helpers

How about lazy enumeration?

AoC 2015 Day 24:

It Hangs in the Balance

Task:

  Split cargo (28 pieces) into 3 groups of equal weight.

  Pick combination with least items in the first group.

 

Analysis:

  Total possible combinations is                                                       !!!

  Takes too long to enumerate every possible combination.

  Need a way to exit early once a match is found.

3^{28} = 22,876,792,454,961

Iterables

const iterable = {
  [Symbol.iterator]: function* () {
    let n = 0
    while (true) {
      yield n++
    }
  }
}

for (let n of iterable) {
  if (n >= 10) break
}

Example

// lazy version of assignment enumerator

function getAssignments (items, groups = 2) {
  const nCombinations = Math.pow(groups, items)
  const zeroPad = '0'.repeat(items)
  return {
    [Symbol.iterator]: function* () {
      for (let i = 0; i < nCombinations; i++) {
        const combiString = (zeroPad + i.toString(groups)).slice(-items)
        const combination = combiString.split('')
        if (groups === 2) yield combination.map(v => +v)
        else yield combination
      }
    }
  }
}
// wrong use

const iterable = getAssignment(n)
Array.from(iterable)
[...iterable]
// correct use

const iterable = getAssignment(n)
for (let item of iterable) {
  if (/* condition met */) return item
}

Graph search

Breadth-First-Search

Depth-First-Search

function bfs (root) {
  const visited = {}
  const unvisited = []
  unvisited.push([root, 0])
  
  while (unvisited.length > 0) {
    const [next, steps] = unvisited.shift()
    
    if (next in visited) continue
    visited[next] = 1

    if (found) return steps

    next.children
      .forEach(child => {
        unvisited.push([child, steps + 1]
      })
  }
}
function dfs (root) {
  const visited = {}
  const unvisited = []
  unvisited.push([root, 0])
  
  while (unvisited.length > 0) {
    const [next, steps] = unvisited.pop()
    
    if (next in visited) continue
    visited[next] = 1

    if (found) return steps

    next.children
      .sort(sortFunc)
      .reverse()
      .forEach(child => {
        unvisited.push([child, steps + 1]
      })
  }
}

Dealing with coding task of the assembly language kind

  • Use Chrome inspector
  • node --inspect-brk script.js
  • Add debugger statement

Wrapping up

  • Work it out on paper
  • Premature optimization
  • Readability & Extensibility > Efficiency & Speed
  • Some knowledge of computation complexity is helpful

All solutions

Made with Slides.com