From 5fd04fa470e5438d3afba43ac01f853cca977e5a Mon Sep 17 00:00:00 2001
From: Dan Brown <ssddanbrown@googlemail.com>
Date: Sat, 30 Sep 2017 14:14:23 +0100
Subject: [PATCH] Updated search indexer to split words better
Will now split up words based on more chars than just spaces.
Not takes into account newlines, tabs, periods & commas.
Fixed #531
---
app/Services/SearchService.php | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/app/Services/SearchService.php b/app/Services/SearchService.php
index bb92a1d7c..aebeace1e 100644
--- a/app/Services/SearchService.php
+++ b/app/Services/SearchService.php
@@ -382,11 +382,13 @@ class SearchService
protected function generateTermArrayFromText($text, $scoreAdjustment = 1)
{
$tokenMap = []; // {TextToken => OccurrenceCount}
- $splitText = explode(' ', $text);
- foreach ($splitText as $token) {
- if ($token === '') continue;
+ $splitChars = " \n\t.,";
+ $token = strtok($text, $splitChars);
+
+ while ($token !== false) {
if (!isset($tokenMap[$token])) $tokenMap[$token] = 0;
$tokenMap[$token]++;
+ $token = strtok($splitChars);
}
$terms = [];