Add labels vector processing, that prevent an segfault if label index…

…es are not sorted
mwydmuch · Jan 10, 2025 · 5e9ae56 · 5e9ae56
1 parent 824ec68
commit 5e9ae56
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 1 deletion.
diff --git a/src/read_data.cpp b/src/read_data.cpp
@@ -100,7 +100,10 @@ bool DataReader::readData(SRMatrix& labels, SRMatrix& features, Args& args, int
             continue;
         }
 
-        if(args.processData) processFeaturesVector(lFeatures, args.norm, args.hash, args.featuresThreshold);
+        if(args.processData) {
+            processFeaturesVector(lFeatures, args.norm, args.hash, args.featuresThreshold);
+            processLabelsVector(lLabels);
+        }
 
         labels.appendRow(lLabels);
         features.appendRow(lFeatures);
@@ -193,4 +196,9 @@ void DataReader::processFeaturesVector(std::vector<IRVPair> &lFeatures, bool nor
     if (!std::is_sorted(lFeatures.begin(), lFeatures.end(), IRVPairIndexComp())) sort(lFeatures.begin(), lFeatures.end(), IRVPairIndexComp());
 }
 
+void DataReader::processLabelsVector(std::vector<IRVPair> &lLabels) {
+    // Check if it requires sorting
+    if (!std::is_sorted(lLabels.begin(), lLabels.end(), IRVPairIndexComp())) sort(lLabels.begin(), lLabels.end(), IRVPairIndexComp());
+}
+
 
diff --git a/src/read_data.h b/src/read_data.h
@@ -39,6 +39,7 @@ class DataReader {
 
     static void prepareFeaturesVector(std::vector<IRVPair> &lFeatures, Real bias = 1.0);
     static void processFeaturesVector(std::vector<IRVPair> &lFeatures, bool norm = true, size_t hashSize = 0, Real featuresThreshold = 0);
+    static void processLabelsVector(std::vector<IRVPair> &lLabels);
 
 private:
     std::ifstream in;