Created
October 18, 2024 04:22
-
-
Save bpevs/5c3fdfdda4758aaa7698a55864d5d36c to your computer and use it in GitHub Desktop.
Parse Dict
This file contains hidden or 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
/** | |
* @description | |
* This is a modified version of parse-dictd to be used with Deno, using only | |
* std dependencies. | |
* | |
* @reference https://github.com/nvdnkpr/parse-dictd | |
*/ | |
import { TextLineStream } from "jsr:@std/streams/text-line-stream"; | |
import { toTransformStream } from "jsr:@std/streams/to-transform-stream"; | |
const az = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/"; | |
const codes = {}; | |
for (let i = 0; i < 64; i++) codes[az.charAt(i)] = i; | |
const dzUtf8Stream = (await Deno.open("spa-eng/spa-eng.dict.dz")).readable | |
.pipeThrough(new DecompressionStream("gzip")) | |
.pipeThrough(new TextDecoderStream()); | |
const indexUtf8Stream = (await Deno.open("spa-eng/spa-eng.index")).readable | |
.pipeThrough(new TextDecoderStream()); | |
const r = await parse(dzUtf8Stream, indexUtf8Stream); | |
await Deno.writeTextFile("dict.json", JSON.stringify(r, null, "\t")); | |
async function parse( | |
dstream: ReadableStream<string>, | |
istream: ReadableStream<string>, | |
) { | |
const offsets: Record<number, { word: string }> = {}; | |
let pos = 0; | |
// Transform the index stream into objects by processing each line | |
const indexStream = istream | |
.pipeThrough(new TextLineStream()) // Split the index stream by lines | |
.pipeThrough(toTransformStream(async function* (src) { | |
for await (const chunk of src) { | |
const fields = chunk.trim().split("\t"); | |
if (fields.length !== 3) continue; | |
offsets[decode(fields[1])] = { word: fields[0] }; | |
yield; | |
} | |
})); | |
await Array.fromAsync(indexStream); | |
const outputStream = dstream | |
.pipeThrough(new TextLineStream()) | |
.pipeThrough(toTransformStream(async function* (src) { | |
for await (const line of src) { | |
const to = line.trim() | |
.replace(/[^\x20-\x7E]/g, "") | |
.replace(/^\w+:/, "") | |
.trim().split(","); | |
if (offsets[pos]) { | |
yield { from: offsets[pos].word, to }; | |
} | |
pos += new TextEncoder().encode(line).length; | |
} | |
})); | |
const results = {}; | |
for await (const output of outputStream) { | |
if (output) results[output.from] = output.to; | |
} | |
return results; | |
} | |
function decode(s) { | |
let index = 0; | |
const len = s.length; | |
for (let i = 0; i < len; i++) { | |
index += codes[s.charAt(i)] * Math.pow(64, len - i - 1); | |
} | |
return index; | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment