From db0216936eddbbcc902e9d089b1cacc5e00caaf7 Mon Sep 17 00:00:00 2001
From: Viktor Lofgren <vlofgren@marginalia.nu>
Date: Wed, 16 Aug 2023 15:48:34 +0200
Subject: [PATCH] (summary) Reduce the chance of expensive operations

---
 .../summary/heuristic/SummarizingDOMFilter.java      | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/code/features-convert/summary-extraction/src/main/java/nu/marginalia/summary/heuristic/SummarizingDOMFilter.java b/code/features-convert/summary-extraction/src/main/java/nu/marginalia/summary/heuristic/SummarizingDOMFilter.java
index f2137a60..f72b0eae 100644
--- a/code/features-convert/summary-extraction/src/main/java/nu/marginalia/summary/heuristic/SummarizingDOMFilter.java
+++ b/code/features-convert/summary-extraction/src/main/java/nu/marginalia/summary/heuristic/SummarizingDOMFilter.java
@@ -101,9 +101,15 @@ public class SummarizingDOMFilter implements NodeFilter {
 
         for (var stats : in) {
             // text() is expensive, we don't mind sifting through superfluous whitespace
-            int cnt = stats.score(tn ->
-                    countOccurrencesOfAnyWord(tn.getWholeText(), importantWords)
-                        - countOccurrencesOfAnyWord(tn.getWholeText(), badWords));
+            int cnt = stats.score(tn -> {
+                String wholeText = tn.getWholeText();
+
+                if (wholeText.length() > 128)
+                    return 0;
+
+                return countOccurrencesOfAnyWord(wholeText, importantWords)
+                        - countOccurrencesOfAnyWord(wholeText, badWords);
+            });
 
             if (cnt > 0) {
                 ret.put(stats, -cnt);