From fa335564eda13e793403912740f3d47c80867ccb Mon Sep 17 00:00:00 2001
From: Chi Vinh Le <vinh@wikiwi.io>
Date: Thu, 14 Sep 2017 00:10:29 +0700
Subject: [PATCH] Generate RegExp from phrases

---
 .../src/components/CommentBodyHighlighter.js  | 100 ++++--------------
 1 file changed, 21 insertions(+), 79 deletions(-)

diff --git a/client/coral-admin/src/components/CommentBodyHighlighter.js b/client/coral-admin/src/components/CommentBodyHighlighter.js
index fe5bcff96..64b53b829 100644
--- a/client/coral-admin/src/components/CommentBodyHighlighter.js
+++ b/client/coral-admin/src/components/CommentBodyHighlighter.js
@@ -1,90 +1,32 @@
 import React from 'react';
 import {matchLinks} from '../utils';
 
-const capturingWordSeparator = /([.\s'"?!])/;
-const wordSeparator = /[.\s'"?!]/;
+function escapeRegExp(string) {
+  return string.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
+}
+
+function generateRegExp(phrases) {
+  const inner = phrases
+    .map((phrase) => {
+      return phrase.split(/\s+/)
+        .map((word) => escapeRegExp(word))
+        .join('[\\s"?!.]+');
+    }).join('|');
+
+  return `(^|[^\\w])(${inner})(?=[^\\w]|$)`;
+}
 
 // markPhrases looks for `phrases` inside `body` and highlights them by returning
 // an array of React Elements.
 function markPhrases(body, phrases, keyPrefix) {
-  const tokens = body.split(capturingWordSeparator);
-  const phraseWords = phrases.map((phrase) => phrase.toLowerCase().split(wordSeparator));
-  const content = [];
-  let tmp = [];
-
-  for (let l = 0; l < tokens.length; l++) {
-
-    // matchedWords is > 0 when a full match was found and contains
-    // the range length from this index to the end of the match.
-    let matchedWords = 0;
-
-    // Skip word separators and ''.
-    if (tokens[l] !== '' && !tokens[l].match(wordSeparator)) {
-      for (let m = 0; m < phraseWords.length; m++) {
-        const words = phraseWords[m];
-
-        // We try to match the full phrase, index keeps track
-        // of where we are now on the tokens array while matching
-        // the words of the phrase.
-        let index = l;
-        for (let n = 0; n < words.length; n++, index++) {
-
-          // Skip word separators and ''.
-          while (index < tokens.length && (tokens[index].match(wordSeparator) || tokens[index] === '')) {
-            index++;
-          }
-
-          // No more tokens left.
-          if (index >= tokens.length) {
-            break;
-          }
-
-          const token = tokens[index].toLowerCase();
-          const word = words[n];
-          if (token !== word) {
-            break;
-          }
-
-          // Full match!
-          if (n === words.length - 1) {
-
-            // Save the matched range length into matched words.
-            matchedWords = index - l + 1;
-            break;
-          }
-        }
-
-        // We matched a word so break out the loop.
-        if (matchedWords) {
-          break;
-        }
-      }
+  const regexp = new RegExp(generateRegExp(phrases), 'iu');
+  const tokens = body.split(regexp);
+  return tokens.map((token, i) => {
+    if (i % 3 === 2) {
+      return <mark key={`${keyPrefix}_${i}`}>{token}</mark>;
     }
-
-    // We have a match!
-    if (matchedWords) {
-      const match = tokens.slice(l, l + matchedWords).join('');
-
-      // Append whatever we have in `tmp` and clear it.
-      content.push(tmp.join(''));
-      tmp = [];
-
-      content.push(<mark key={`${keyPrefix}_${l}`}>{match}</mark>);
-
-      // Move index further if we matched more than one word.
-      l += matchedWords - 1;
-
-      continue;
-    }
-
-    // No match, we just push this into `tmp`.
-    tmp.push(tokens[l]);
-  }
-
-  // Append any non matched tokens currently in `tmp`.
-  content.push(tmp.join(''));
-
-  return content;
+    return token;
+  });
 }
 
 // markLinks looks for links inside `body` and highlights them by returning