Last active
July 10, 2016 12:14
-
-
Save SPY/ba618fe5a63eff57ff40edbefd5ff48c to your computer and use it in GitHub Desktop.
TypeScript implementation of Burrows–Wheeler transform. It is more effective than naive implementation and requires O(n) additional memory for direct and inverse transformation.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
const enum Order { | |
LT = -1, | |
EQ = 0, | |
GT = 1 | |
} | |
interface BWTransformed { | |
data: string, | |
start: number, | |
eof: number | |
} | |
function bwt(input: string): BWTransformed { | |
const len = input.length + 2 | |
const rotations = new Int32Array(len) | |
for (let i = 0; i < len; i++) { | |
rotations[i] = i | |
} | |
const EOF = len - 2 | |
const START = -1 | |
rotations.sort((a, b) => { | |
for (let i = 0; i < len; i++) { | |
const x = a + i > len - 1 ? i - (len - a) - 1 : i + a - 1 | |
const y = b + i > len - 1 ? i - (len - b) - 1 : i + b - 1 | |
if (x === EOF) { | |
return Order.GT | |
} | |
else if (x === START) { | |
return y === EOF ? Order.LT : Order.GT | |
} | |
else if (y === EOF || y === START) { | |
return Order.LT | |
} | |
else if (input[x] < input[y]) { | |
return Order.LT | |
} | |
else if (input[x] > input[y]) { | |
return Order.GT | |
} | |
} | |
return Order.EQ | |
}) | |
let start = 0 | |
let eof = 0 | |
const data = new Array(input.length) | |
let added = 0 | |
const lastIdx = len - 1 | |
for (let i = 0; i < len; i++) { | |
const shift = rotations[i] | |
const real = shift + lastIdx > len - 1 ? lastIdx - (len - shift) - 1 : lastIdx + shift - 1 | |
if (real === EOF) { | |
eof = i | |
} | |
else if (real === START) { | |
start = i | |
} | |
else { | |
data[added++] = input[real] | |
} | |
} | |
return { start, data: data.join(''), eof } | |
} | |
function ibwt({start, data, eof}: BWTransformed): string { | |
const len = data.length + 2 | |
const sorted = new Int32Array(len) | |
const permutations = new Int32Array(len) | |
for (let i = 0; i < len; i++) { | |
sorted[i] = i | |
} | |
sorted.sort((a, b) => { | |
if (a === eof) { | |
return Order.GT | |
} | |
else if (a === start) { | |
return b === eof ? Order.LT : Order.GT | |
} | |
else if (b === start || b === eof) { | |
return Order.LT | |
} | |
const x = a + (a >= start ? -1 + (a >= eof ? -1 : 0) : 0) | |
const y = b + (b >= start ? -1 + (b >= eof ? -1 : 0) : 0) | |
return data[x] < data[y] | |
? Order.LT | |
: ( | |
data[x] > data[y] | |
? Order.GT | |
: (a < b ? Order.LT : Order.GT) | |
) | |
}) | |
for (let i = 0; i < len; i++) { | |
permutations[sorted[i]] = i | |
} | |
let current = len - 1 | |
const result = new Array(len - 2) | |
let i = len - 2 | |
while (i--) { | |
const idx = current + (current < start ? 0 : (-1 + (current < eof ? 0 : -1))) | |
result[i] = data[idx] | |
current = permutations[current] | |
} | |
return result.join('') | |
} | |
const WORD = `banana` | |
console.log(ibwt(bwt(WORD)) === WORD) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment