taraskevizer 10.4.8 → 10.4.9

package/dist/bin/index.js

@@ -0,0 +1,158 @@
+#!/usr/bin/env node
+import { cpus } from 'node:os';
+import { Worker } from 'node:worker_threads';
+import { pipelines } from '../index.js';
+import { parseArgs } from './parse-args.js';
+const printWithPrefix = (msg) => {
+    process.stdout.write("[taraskevizer]" + ' ' + msg + '\n');
+};
+const splitIntoChunks = (text, n) => {
+    const size = Math.ceil(text.length / n);
+    const chunks = [];
+    let start = 0;
+    for (let i = 0; i < n; i++) {
+        let end = start + size;
+        if (end < text.length) {
+            const forward = text.indexOf('\n', end);
+            const backward = text.lastIndexOf('\n', end);
+            if (forward === -1 && backward === -1) {
+                // no-op, use raw end
+            }
+            else if (forward === -1) {
+                end = backward;
+            }
+            else if (backward === -1) {
+                end = forward;
+            }
+            else {
+                end = forward - end < end - backward ? forward : backward;
+            }
+        }
+        if (end > text.length)
+            end = text.length;
+        chunks.push(text.slice(start, end));
+        start = end;
+        if (start >= text.length)
+            break;
+    }
+    return chunks;
+};
+process.argv.splice(0, 2);
+const firstArg = process.argv[0];
+if (firstArg) {
+    if (firstArg === '-v' || firstArg === '--version') {
+        printWithPrefix("10.4.9");
+        process.exit(0);
+    }
+    if (firstArg === '-h' || firstArg === '--help') {
+        printWithPrefix(`Usage: tarask [options] text
+If text is not passed, interactive mode is enabled
+
+EXAMPLES
+
+Convert and latinize a word
+	tarask --latin 'планета'
+Prints "planeta"
+
+Read from one file and write converted text to another
+	tarask < ./cyr-text.txt > ./lat-text.txt
+
+Enter interactive mode
+	tarask
+Prints "[taraskevizer] Enter the text:" and waits until you enter a new line
+
+OPTIONS
+
+General:
+  -h --help
+  -v --version
+
+Alphabet:
+  -l  --latin
+  -lj --latin-ji
+  -a  --arabic
+
+When to replace і(i) by й(j) after vowels:
+  -jr --jrandom
+  -ja --jalways
+
+Replace ґ(g) by г(h) in cyrillic alphabet:
+  --h
+
+Variations:
+  -nv --no-variations
+  -fv --first-variation
+
+Mode (only one can be used):
+  -html --html
+  -abc  --alphabet-only
+
+Other:
+  -nec --not-escape-caps
+  -nc  --no-color
+  -st  --single-thread
+`);
+        process.exit(0);
+    }
+}
+const { mode, cfg, doForceSingleThread } = parseArgs(process.argv);
+let text = '';
+if (process.argv.length) {
+    text = process.argv.reverse().join(' ');
+}
+else {
+    const chunks = [];
+    let length = 0;
+    if (process.stdin.isTTY) {
+        printWithPrefix('Enter the text');
+        for await (const chunk of process.stdin) {
+            chunks.push(chunk);
+            length += chunk.length;
+            if (chunk.includes('\n'))
+                break;
+        }
+    }
+    else {
+        for await (const chunk of process.stdin) {
+            chunks.push(chunk);
+            length += chunk.length;
+        }
+    }
+    text = Buffer.concat(chunks, length).toString();
+}
+let result = '';
+if (text.length > 50_000 && !doForceSingleThread) {
+    const cpuCount = Math.max(1, cpus()?.length || 1);
+    const chunks = splitIntoChunks(text, cpuCount);
+    const WORKER_CODE = `
+const { parentPort, workerData } = require('node:worker_threads');
+const { pipelines } = require('./dist');
+const { parseArgs } = require('./dist/bin/parse-args');
+const { argv, chunk } = workerData;
+const { mode, cfg } = parseArgs(argv);
+parentPort.postMessage(pipelines[mode](chunk, cfg));`;
+    const results = await Promise.all(chunks.map((chunk) => new Promise((resolve, reject) => {
+        const worker = new Worker(WORKER_CODE, {
+            eval: true,
+            workerData: { argv: process.argv, chunk },
+        });
+        worker.on('message', resolve);
+        worker.on('error', reject);
+        worker.on('exit', (code) => {
+            if (code !== 0)
+                reject(new Error('Worker exit code ' + code));
+        });
+    })));
+    result = results.join('\n') + '\n';
+}
+else {
+    result = pipelines[mode](text, cfg) + '\n';
+}
+if (process.stdout.write(result)) {
+    process.exit(0);
+}
+else {
+    process.stdout.once('drain', () => {
+        process.exit(0);
+    });
+}

package/dist/bin/parse-args.d.ts

@@ -0,0 +1,6 @@
+import { TaraskConfig } from '..';
+export declare const parseArgs: (argv: string[]) => {
+    cfg: Partial<TaraskConfig>;
+    mode: "tarask";
+    doForceSingleThread: boolean;
+};

package/dist/bin/parse-args.js

@@ -0,0 +1,123 @@
+import { dicts, htmlConfigOptions, TaraskConfig, wrappers, } from '../index.js';
+const toHashTable = (dict) => {
+    const result = {};
+    for (const { 0: options, 1: callback } of dict)
+        for (const option of options)
+            result[option] = callback;
+    return result;
+};
+export const parseArgs = (argv) => {
+    let cfg = {
+        g: true,
+        variations: 'all',
+        wrappers: wrappers.ansiColor,
+    };
+    let mode = 'tarask';
+    let isHtml = false;
+    let doForceSingleThread = false;
+    const optionDict = toHashTable([
+        [
+            ['--latin', '-l'],
+            () => {
+                cfg.abc = dicts.alphabets.latin;
+            },
+        ],
+        [
+            ['--latin-ji', '-lj'],
+            () => {
+                cfg.abc = dicts.alphabets.latinJi;
+            },
+        ],
+        [
+            ['--arabic', '-a'],
+            () => {
+                cfg.abc = dicts.alphabets.arabic;
+            },
+        ],
+        [
+            ['--jrandom', '-jr'],
+            () => {
+                cfg.j = 'random';
+            },
+        ],
+        [
+            ['--jalways', '-ja'],
+            () => {
+                cfg.j = 'always';
+            },
+        ],
+        [
+            ['--no-escape-caps', '-nec'],
+            () => {
+                cfg.doEscapeCapitalized = false;
+            },
+        ],
+        [
+            ['--h'],
+            () => {
+                cfg.g = false;
+            },
+        ],
+        [
+            ['--no-variations', '-nv'],
+            () => {
+                cfg.variations = 'no';
+            },
+        ],
+        [
+            ['--first-variation', '-fv'],
+            () => {
+                cfg.variations = 'first';
+            },
+        ],
+        [
+            ['--no-color', '-nc'],
+            () => {
+                cfg.wrappers = null;
+            },
+        ],
+        [
+            ['--html', '-html'],
+            () => {
+                isHtml = true;
+                cfg.wrappers = htmlConfigOptions.wrappers;
+            },
+        ],
+        [
+            ['--alphabet-only', '-abc'],
+            () => {
+                mode = 'alphabetic';
+            },
+        ],
+        [
+            ['--phonetic', '-ph'],
+            () => {
+                mode = 'phonetic';
+            },
+        ],
+        [
+            ['--single-thread', '-st'],
+            () => {
+                doForceSingleThread = true;
+            },
+        ],
+    ]);
+    let currOption;
+    argv.reverse();
+    while ((currOption = argv.pop())) {
+        if (currOption in optionDict) {
+            optionDict[currOption]();
+        }
+        else {
+            argv.push(currOption);
+            break;
+        }
+    }
+    cfg = new TaraskConfig(isHtml
+        ? {
+            ...htmlConfigOptions,
+            ...cfg,
+        }
+        : cfg);
+    return { cfg, mode, doForceSingleThread };
+};

package/dist/config.d.ts

@@ -1,5 +1,5 @@
-import type { DeepReadonly } from './types';
 import type { Alphabet } from './dict/alphabets';
+import type { DeepReadonly } from './types';
 import { type Wrappers } from './wrappers';
 export type Variation = 'no' | 'first' | 'all';
 export type OptionJ = 'never' | 'random' | 'always';

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "taraskevizer",
-	"version": "10.4.8",
+	"version": "10.4.9",
 	"author": "GooseOb",
 	"repository": {
 		"type": "git",
@@ -22,7 +22,7 @@
 		"typescript-eslint": "^8.42.0"
 	},
 	"bin": {
-		"tarask": "dist/bin.js"
+		"tarask": "dist/bin/index.js"
 	},
 	"description": "Канвэртацыя акадэмічнага правапісу ў клясычны",
 	"files": [

package/dist/bin.js

@@ -1,206 +0,0 @@
-#!/usr/bin/env node
-import { dicts, TaraskConfig, pipelines, htmlConfigOptions, wrappers } from './index.js';
-const printWithPrefix = (msg) => {
-    process.stdout.write("[taraskevizer]" + ' ' + msg + '\n');
-};
-process.argv.splice(0, 2);
-const firstArg = process.argv[0];
-if (firstArg) {
-    if (firstArg === '-v' || firstArg === '--version') {
-        printWithPrefix("10.4.8");
-        process.exit(0);
-    }
-    if (firstArg === '-h' || firstArg === '--help') {
-        printWithPrefix(`Usage: tarask [options] text
-If text is not passed, interactive mode will be enabled
-
-EXAMPLES
-
-Convert and latinize a word
-	tarask --latin 'планета'
-Will print "planeta"
-
-Read from one file and write converted text to another
-	tarask < ./cyr-text.txt > ./lat-text.txt
-
-Enter interactive mode
-	tarask
-Will print "[taraskevizer] Enter the text:" and wait until you press Enter
-
-OPTIONS
-
-General:
-  -h --help
-  -v --version
-
-Alphabet:
-  -l --latin
-  -lj --latin-ji
-  -a --arabic
-
-When to replace і(i) by й(j) after vowels:
-  -jr --jrandom
-  -ja --jalways
-
-Replace ґ(g) by г(h) in cyrillic alphabet:
-  --h
-
-Variations:
-  -nv --no-variations
-  -fv --first-variation
-
-Mode (only one can be used):
-  -html --html
-  -abc --alphabet-only
-
-Other:
-  -nec --not-escape-caps
-  -nc --no-color
-`);
-        process.exit(0);
-    }
-}
-let cfg = {
-    g: true,
-    variations: 'all',
-    wrappers: wrappers.ansiColor,
-};
-let mode = 'tarask';
-const toHashTable = (dict) => {
-    const result = {};
-    for (const { 0: options, 1: callback } of dict)
-        for (const option of options)
-            result[option] = callback;
-    return result;
-};
-let isHtml = false;
-const optionDict = toHashTable([
-    [
-        ['--latin', '-l'],
-        () => {
-            cfg.abc = dicts.alphabets.latin;
-        },
-    ],
-    [
-        ['--latin-ji', '-lj'],
-        () => {
-            cfg.abc = dicts.alphabets.latinJi;
-        },
-    ],
-    [
-        ['--arabic', '-a'],
-        () => {
-            cfg.abc = dicts.alphabets.arabic;
-        },
-    ],
-    [
-        ['--jrandom', '-jr'],
-        () => {
-            cfg.j = 'random';
-        },
-    ],
-    [
-        ['--jalways', '-ja'],
-        () => {
-            cfg.j = 'always';
-        },
-    ],
-    [
-        ['--no-escape-caps', '-nec'],
-        () => {
-            cfg.doEscapeCapitalized = false;
-        },
-    ],
-    [
-        ['--h'],
-        () => {
-            cfg.g = false;
-        },
-    ],
-    [
-        ['--no-variations', '-nv'],
-        () => {
-            cfg.variations = 'no';
-        },
-    ],
-    [
-        ['--first-variation', '-fv'],
-        () => {
-            cfg.variations = 'first';
-        },
-    ],
-    [
-        ['--no-color', '-nc'],
-        () => {
-            cfg.wrappers = null;
-        },
-    ],
-    [
-        ['--html', '-html'],
-        () => {
-            isHtml = true;
-            cfg.wrappers = htmlConfigOptions.wrappers;
-        },
-    ],
-    [
-        ['--alphabet-only', '-abc'],
-        () => {
-            mode = 'alphabetic';
-        },
-    ],
-    [
-        ['--phonetic', '-ph'],
-        () => {
-            mode = 'phonetic';
-        },
-    ],
-]);
-let currOption;
-process.argv.reverse();
-while ((currOption = process.argv.pop())) {
-    if (currOption in optionDict) {
-        optionDict[currOption]();
-    }
-    else {
-        process.argv.push(currOption);
-        break;
-    }
-}
-let text = '';
-if (process.argv.length) {
-    text = process.argv.reverse().join(' ');
-}
-else {
-    const chunks = [];
-    let length = 0;
-    if (process.stdin.isTTY) {
-        printWithPrefix('Enter the text');
-        for await (const chunk of process.stdin) {
-            chunks.push(chunk);
-            length += chunk.length;
-            if (chunk.includes('\n'))
-                break;
-        }
-    }
-    else {
-        for await (const chunk of process.stdin) {
-            chunks.push(chunk);
-            length += chunk.length;
-        }
-    }
-    text = Buffer.concat(chunks, length).toString();
-}
-cfg = new TaraskConfig(isHtml
-    ? {
-        ...htmlConfigOptions,
-        ...cfg,
-    }
-    : cfg);
-if (process.stdout.write(pipelines[mode](text, cfg) + '\n')) {
-    process.exit(0);
-}
-else {
-    process.stdout.once('drain', () => {
-        process.exit(0);
-    });
-}