@scrabble-solver/word-definitions 2.9.3 → 2.10.0

package/build/crawl/crawl.js

@@ -5,6 +5,7 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
 Object.defineProperty(exports, "__esModule", { value: true });
 const types_1 = require("@scrabble-solver/types");
 const crawlEnglish_1 = __importDefault(require("./crawlEnglish"));
+const crawlFarsi_1 = __importDefault(require("./crawlFarsi"));
 const crawlFrench_1 = __importDefault(require("./crawlFrench"));
 const crawlGerman_1 = __importDefault(require("./crawlGerman"));
 const crawlPolish_1 = __importDefault(require("./crawlPolish"));
@@ -14,6 +15,7 @@ const crawlPerLocale = {
     [types_1.Locale.EN_GB]: crawlEnglish_1.default,
     [types_1.Locale.EN_US]: crawlEnglish_1.default,
     [types_1.Locale.ES_ES]: crawlSpanish_1.default,
+    [types_1.Locale.FA_IR]: crawlFarsi_1.default,
     [types_1.Locale.FR_FR]: crawlFrench_1.default,
     [types_1.Locale.PL_PL]: crawlPolish_1.default,
 };

package/build/crawl/crawlFarsi.d.ts

@@ -0,0 +1,2 @@
+declare const crawlFarsi: (word: string) => Promise<string>;
+export default crawlFarsi;

package/build/crawl/crawlFarsi.js

@@ -0,0 +1,11 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+const lib_1 = require("../lib");
+const crawlFarsi = (word) => {
+    return (0, lib_1.request)({
+        protocol: 'https',
+        hostname: 'www.vajehyab.com',
+        path: `/moein/${encodeURIComponent(word)}`,
+    });
+};
+exports.default = crawlFarsi;

package/build/lib/normalizeDefinition.js

@@ -15,7 +15,8 @@ const normalizeQuotes = (definition) => definition.replace(/\."/g, '".');
 const normalizeMarkers = (definition) => definition.replace(/^\(\d+\.\d+\)\s+/g, '');
 const normalizeTrailingSymbols = (definition) => definition.trim().replace(/:$/, '');
 const normalizeLeadingSymbols = (definition) => definition.trim().replace(/^:/, '');
-const normalizeNonWords = (definition) => (/\w/.test(definition) ? definition : '');
+// https://stackoverflow.com/a/40732368
+const normalizeNonWords = (definition) => (/[\w\u0621-\u064A]+/.test(definition) ? definition : '');
 const normalizeCommas = (definition) => {
     return definition
         .replace(/\s+,\s+/g, ', ')

package/build/parse/index.d.ts

@@ -1,5 +1,6 @@
 export { default as parse } from './parse';
 export { default as parseEnglish } from './parseEnglish';
+export { default as parseFarsi } from './parseFarsi';
 export { default as parseFrench } from './parseFrench';
 export { default as parseGerman } from './parseGerman';
 export { default as parsePolish } from './parsePolish';

package/build/parse/index.js

@@ -3,11 +3,13 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
 };
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.parseSpanish = exports.parsePolish = exports.parseGerman = exports.parseFrench = exports.parseEnglish = exports.parse = void 0;
+exports.parseSpanish = exports.parsePolish = exports.parseGerman = exports.parseFrench = exports.parseFarsi = exports.parseEnglish = exports.parse = void 0;
 var parse_1 = require("./parse");
 Object.defineProperty(exports, "parse", { enumerable: true, get: function () { return __importDefault(parse_1).default; } });
 var parseEnglish_1 = require("./parseEnglish");
 Object.defineProperty(exports, "parseEnglish", { enumerable: true, get: function () { return __importDefault(parseEnglish_1).default; } });
+var parseFarsi_1 = require("./parseFarsi");
+Object.defineProperty(exports, "parseFarsi", { enumerable: true, get: function () { return __importDefault(parseFarsi_1).default; } });
 var parseFrench_1 = require("./parseFrench");
 Object.defineProperty(exports, "parseFrench", { enumerable: true, get: function () { return __importDefault(parseFrench_1).default; } });
 var parseGerman_1 = require("./parseGerman");

package/build/parse/parse.js

@@ -6,6 +6,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const types_1 = require("@scrabble-solver/types");
 const lib_1 = require("../lib");
 const parseEnglish_1 = __importDefault(require("./parseEnglish"));
+const parseFarsi_1 = __importDefault(require("./parseFarsi"));
 const parseFrench_1 = __importDefault(require("./parseFrench"));
 const parseGerman_1 = __importDefault(require("./parseGerman"));
 const parsePolish_1 = __importDefault(require("./parsePolish"));
@@ -15,6 +16,7 @@ const parsePerLocale = {
     [types_1.Locale.EN_GB]: parseEnglish_1.default,
     [types_1.Locale.EN_US]: parseEnglish_1.default,
     [types_1.Locale.ES_ES]: parseSpanish_1.default,
+    [types_1.Locale.FA_IR]: parseFarsi_1.default,
     [types_1.Locale.FR_FR]: parseFrench_1.default,
     [types_1.Locale.PL_PL]: parsePolish_1.default,
 };

package/build/parse/parseFarsi.d.ts

@@ -0,0 +1,3 @@
+import { ParseResult } from '../types';
+declare const parseFarsi: (html: string) => ParseResult;
+export default parseFarsi;

package/build/parse/parseFarsi.js

@@ -0,0 +1,13 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+const cheerio_1 = require("cheerio");
+const DOES_NOT_EXIST_MESSAGE = '404 Page Not Found';
+const parseFarsi = (html) => {
+    const $ = (0, cheerio_1.load)(html);
+    const $definitions = $('[itemprop=articleBody]');
+    return {
+        definitions: Array.from($definitions).map((definition) => $(definition).text()),
+        exists: $('#container h1').text() !== DOES_NOT_EXIST_MESSAGE,
+    };
+};
+exports.default = parseFarsi;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@scrabble-solver/word-definitions",
-  "version": "2.9.3",
+  "version": "2.10.0",
   "description": "Scrabble Solver 2 - Word definitions",
   "main": "build/index.js",
   "types": "build/index.d.ts",
@@ -23,7 +23,7 @@
     "clean": "rimraf build/ node_modules/ package-lock.json"
   },
   "dependencies": {
-    "@scrabble-solver/types": "^2.9.3",
+    "@scrabble-solver/types": "^2.10.0",
     "cheerio": "^1.0.0-rc.12",
     "follow-redirects": "^1.15.2",
     "striptags": "^3.2.0"
@@ -31,5 +31,5 @@
   "devDependencies": {
     "@types/follow-redirects": "^1.14.1"
   },
-  "gitHead": "ef8d030cbb76256a377f747f7ebe4b85b7eba61e"
+  "gitHead": "7fd0f6bf536fca88c18ce02e320da71b0c6d6f61"
 }

package/src/crawl/crawl.ts

@@ -1,6 +1,7 @@
 import { Locale } from '@scrabble-solver/types';
 
 import crawlEnglish from './crawlEnglish';
+import crawlFarsi from './crawlFarsi';
 import crawlFrench from './crawlFrench';
 import crawlGerman from './crawlGerman';
 import crawlPolish from './crawlPolish';
@@ -11,6 +12,7 @@ const crawlPerLocale: Record<Locale, (word: string) => Promise<string>> = {
   [Locale.EN_GB]: crawlEnglish,
   [Locale.EN_US]: crawlEnglish,
   [Locale.ES_ES]: crawlSpanish,
+  [Locale.FA_IR]: crawlFarsi,
   [Locale.FR_FR]: crawlFrench,
   [Locale.PL_PL]: crawlPolish,
 };

package/src/crawl/crawlFarsi.ts

@@ -0,0 +1,11 @@
+import { request } from '../lib';
+
+const crawlFarsi = (word: string): Promise<string> => {
+  return request({
+    protocol: 'https',
+    hostname: 'www.vajehyab.com',
+    path: `/moein/${encodeURIComponent(word)}`,
+  });
+};
+
+export default crawlFarsi;

package/src/lib/normalizeDefinition.ts

@@ -21,7 +21,8 @@ const normalizeTrailingSymbols: Normalize = (definition) => definition.trim().re
 
 const normalizeLeadingSymbols: Normalize = (definition) => definition.trim().replace(/^:/, '');
 
-const normalizeNonWords: Normalize = (definition) => (/\w/.test(definition) ? definition : '');
+// https://stackoverflow.com/a/40732368
+const normalizeNonWords: Normalize = (definition) => (/[\w\u0621-\u064A]+/.test(definition) ? definition : '');
 
 const normalizeCommas: Normalize = (definition) => {
   return definition

package/src/parse/index.ts

@@ -1,5 +1,6 @@
 export { default as parse } from './parse';
 export { default as parseEnglish } from './parseEnglish';
+export { default as parseFarsi } from './parseFarsi';
 export { default as parseFrench } from './parseFrench';
 export { default as parseGerman } from './parseGerman';
 export { default as parsePolish } from './parsePolish';

package/src/parse/parse.ts

@@ -4,6 +4,7 @@ import { normalizeDefinition, unique } from '../lib';
 import { ParseResult } from '../types';
 
 import parseEnglish from './parseEnglish';
+import parseFarsi from './parseFarsi';
 import parseFrench from './parseFrench';
 import parseGerman from './parseGerman';
 import parsePolish from './parsePolish';
@@ -14,6 +15,7 @@ const parsePerLocale: Record<Locale, (html: string) => ParseResult> = {
   [Locale.EN_GB]: parseEnglish,
   [Locale.EN_US]: parseEnglish,
   [Locale.ES_ES]: parseSpanish,
+  [Locale.FA_IR]: parseFarsi,
   [Locale.FR_FR]: parseFrench,
   [Locale.PL_PL]: parsePolish,
 };

package/src/parse/parseFarsi.ts

@@ -0,0 +1,17 @@
+import { load } from 'cheerio';
+
+import { ParseResult } from '../types';
+
+const DOES_NOT_EXIST_MESSAGE = '404 Page Not Found';
+
+const parseFarsi = (html: string): ParseResult => {
+  const $ = load(html);
+  const $definitions = $('[itemprop=articleBody]');
+
+  return {
+    definitions: Array.from($definitions).map((definition) => $(definition).text()),
+    exists: $('#container h1').text() !== DOES_NOT_EXIST_MESSAGE,
+  };
+};
+
+export default parseFarsi;