Merge pull request #142 from hcodes/fix_repeated_words

Print row:col for repeated words
hcodes · Mar 27, 2020 · 5d8175f · 5d8175f
2 parents fc084ef + 6091f1d
commit 5d8175f
Show file tree

Hide file tree

Showing 6 changed files with 37 additions and 33 deletions.
diff --git a/lib/dictionary.js b/lib/dictionary.js
@@ -1,6 +1,7 @@
 'use strict';
 
 const chalk = require('chalk');
+const { ERROR_TOO_MANY_ERRORS } = require('yandex-speller');
 const debug = require('./debug');
 const exitCodes = require('./exit-codes');
 const utils = require('./utils');
@@ -132,8 +133,7 @@ module.exports = {
         const dictionary = this.get();
 
         data.forEach(function(typo) {
-            // ERROR_TOO_MANY_ERRORS = 4
-            if (typo.code === 4 || this.isTypo(typo.word, dictionary)) {
+            if (typo.code === ERROR_TOO_MANY_ERRORS || this.isTypo(typo.word, dictionary)) {
                 result.push(typo);
             }
         }, this);

diff --git a/lib/report/index.js b/lib/report/index.js
@@ -11,22 +11,23 @@ const stats = {
     total: 0,
 };
 const reports = [];
-const reportNames = {};
+const reportNames = new Set();
 const buffer = [];
 
 module.exports = {
     addReports(names) {
         names.forEach(function(name) {
-            const moduleName = pth.extname(name) === '.js' ? name : './report/' + name;
-            if (reportNames[moduleName]) {
+            const moduleName = pth.extname(name) === '.js' ? name : './' + name;
+
+            if (reportNames.has(moduleName)) {
                 return;
             }
 
             try {
-                reportNames[moduleName] = true;
                 reports.push(require(moduleName));
+                reportNames.add(moduleName);
             } catch (e) {
-                console.error(chalk.red('Can\'t load report module "' + moduleName + '".'));
+                console.error(chalk.red(`Can't load report module "${moduleName}".`));
                 console.error(chalk.red(e));
             }
         });
@@ -37,17 +38,17 @@ module.exports = {
         });
     },
     oneach(hasError, data, dictionary) {
+        stats.total++;
+
         const hasTypos = Boolean(data && data.data && data.data.length);
 
-        stats.total++;
+        if (hasTypos) {
+            stats.hasTypos = true;
+        }
 
         if (hasError || hasTypos) {
             stats.errors++;
 
-            if (hasTypos) {
-                stats.hasTypos = true;
-            }
-
             buffer.push([hasError, data]);
         } else {
             stats.ok++;

diff --git a/lib/utils.js b/lib/utils.js
@@ -5,6 +5,7 @@ const isutf8 = require('isutf8');
 const minimatch = require('minimatch');
 const pth = require('path');
 const stripJsonComments = require('strip-json-comments');
+const { ERROR_TOO_MANY_ERRORS } = require('yandex-speller');
 const isWin = process.platform === 'win32';
 
 module.exports = {
@@ -177,7 +178,7 @@ module.exports = {
      * @returns {boolean}
      */
     hasManyErrors(data) {
-        return data.some(el => el.code === 4); // ERROR_TOO_MANY_ERRORS
+        return data.some(el => el.code === ERROR_TOO_MANY_ERRORS);
     },
     /**
      * Get uptime in sec.

diff --git a/lib/yaspeller.js b/lib/yaspeller.js
@@ -15,7 +15,6 @@ const yaspellerApi = require('yandex-speller');
 const md = new MarkdownIt();
 const printDebug = require('../lib/debug').print;
 const MAX_LEN_TEXT = 10000; // Max length of text for Yandex.Speller API
-const TOO_MANY_ERRORS = 4;
 
 function getMaxRequest(settings) {
     return settings.maxRequest || 2;
@@ -327,14 +326,19 @@ function checkSitemap(url, commonCallback, settings, callback) {
  */
 function addPositions(text, data) {
     data.forEach(function(item) {
-        if (item.code === TOO_MANY_ERRORS || item.position) {
+        if (item.code === yaspellerApi.ERROR_TOO_MANY_ERRORS || item.position) {
             return;
         }
 
         const result = [];
         const letters = '[^a-zA-Zа-яА-ЯЁёҐґЄєІіЇї]';
+        let word = item.word;
 
-        text.replace(new RegExp(item.word + '(?:' + letters + '|$)', 'g'), function($0, index) {
+        if (item.code === yaspellerApi.ERROR_REPEATED_WORD) {
+            word = item.word + '\\s+' + item.word;
+        }
+
+        text.replace(new RegExp(word + '(?:' + letters + '|$)', 'mg'), function($0, index) {
             const prevSymbol = text[index - 1];
             if (prevSymbol && prevSymbol.search(letters) === -1) {
                 return;
@@ -473,15 +477,15 @@ function prepareText(text) {
     text = fixLineEndings(text);
     text = removeSpecialSymbols(text);
 
-    return text.trim();
+    return text;
 }
 
 function fixLineEndings(text) {
-    return text.replace(/\r\n/g, '\n') // Fix Windows
+    return text
+        .replace(/\r\n/g, '\n') // Fix Windows
         .replace(/\r/g, '\n') // Fix MacOS
         .replace(/\s+\n/g, '\n') // Trailling spaces
-        .replace(/\s+/g, ' ') // Repeat spaces
-        .replace(/\n+/g, '\n'); // Repeat line endings
+        .trimRight();
 }
 
 function removeSpecialSymbols(text) {
@@ -495,17 +499,15 @@ function removeSpecialSymbols(text) {
 }
 
 function getErrors() {
-    return yaspellerApi.errorCodes.filter(function(el) {
-        return el.code !== TOO_MANY_ERRORS;
-    }).map(function(el) {
-        return {
+    return yaspellerApi.errors
+        .filter(el => el.code !== yaspellerApi.ERROR_TOO_MANY_ERRORS)
+        .map(el => ({
             code: el.code,
             title: el.text
-        };
-    }).concat({
-        code: 100, // ERROR_EYO
-        title: 'Letter Ё (Yo)'
-    });
+        })).concat({
+            code: 100, // ERROR_EYO
+            title: 'Letter Ё (Yo)'
+        });
 }
 
 module.exports = {

diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -46,7 +46,7 @@
     "node-fetch": "^2.6.0",
     "strip-json-comments": "^3.0.1",
     "xml2js": "^0.4.23",
-    "yandex-speller": "^4.0.0"
+    "yandex-speller": "^4.1.0"
   },
   "devDependencies": {
     "chai": "^4.2.0",