change start offset logic

NationalSecurityAgency · Apr 16, 2024 · 2f1f8b2 · 2f1f8b2
1 parent 5448dec
commit 2f1f8b2
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 9 deletions.
diff --git a/.../query-core/src/main/java/datawave/query/iterator/logic/TermFrequencyExcerptIterator.java b/.../query-core/src/main/java/datawave/query/iterator/logic/TermFrequencyExcerptIterator.java
@@ -359,8 +359,9 @@ protected String generatePhrase(WordsAndScores[] terms) {
             if (terms[i] == null) {
                 termsToOutput[i] = null;
             } else {
+                String outputWord = terms[i].getWordToOutput();
                 // if the term to put in this position is in the stop list, increment the corresponding *skippedWords and put nothing in this position
-                if (terms[i].getWordToOutput() == null) {
+                if (outputWord == null) {
                     if (i <= (terms.length / 2)) {
                         leftSkippedWords++;
                     } else {
@@ -369,7 +370,7 @@ protected String generatePhrase(WordsAndScores[] terms) {
                     termsToOutput[i] = null;
                 } else {
                     // if the term to output is valid, put it in the same position in the new array
-                    termsToOutput[i] = terms[i].getWordToOutput();
+                    termsToOutput[i] = outputWord;
                     if ((bef || aft) && (terms[i].getHasHitTerm())) {
                         if (!lock) {
                             afterIndex = i;

diff --git a/warehouse/query-core/src/main/java/datawave/query/transformer/ExcerptTransform.java b/warehouse/query-core/src/main/java/datawave/query/transformer/ExcerptTransform.java
@@ -313,27 +313,33 @@ private Set<Excerpt> getExcerpts(PhraseIndexes phraseIndexes) {
      * @return the excerpt
      */
     private String getExcerpt(String field, int start, int end, Range range, ArrayList<String> hitTermValues) {
+        if (start < 0) {
+            start = 0;
+        }
         int prevLeftWordsSkipped = 0;
         int currLeftWordsSkipped = 0;
         boolean leftLock = false;
         int prevRightWordsSkipped = 0;
         int currRightWordsSkipped = 0;
         boolean rightLock = false;
         int timesToTry = 1;
-        float origHalfSize;
-        if (start < 0) {
-            origHalfSize = (float) end / 2;
-        } else {
-            origHalfSize = (float) (end - start) / 2;
-        }
+        float origHalfSize = (float) (end - start) / 2;
         excerptIteratorOptions.put(TermFrequencyExcerptIterator.FIELD_NAME, field);
         for (int i = 0; i <= timesToTry; i++) {
             // if we still need to process the left half...
             if (!leftLock) {
                 // save the value from the last run (or setting to zero if this is the first run)
                 prevLeftWordsSkipped = currLeftWordsSkipped;
                 // set the start of the range to the one passed in minus the amount of skipped words in the left half in the previous run
-                excerptIteratorOptions.put(TermFrequencyExcerptIterator.START_OFFSET, String.valueOf(start - (i * 20)));
+                if (i == 0) {
+                    excerptIteratorOptions.put(TermFrequencyExcerptIterator.START_OFFSET, String.valueOf(start));
+                } else {
+                    if (start - (i * 20) > 0) {
+                        excerptIteratorOptions.put(TermFrequencyExcerptIterator.START_OFFSET, String.valueOf(start - (i * 20)));
+                    } else {
+                        excerptIteratorOptions.put(TermFrequencyExcerptIterator.START_OFFSET, String.valueOf(0));
+                    }
+                }
             }
             // if we still need to process the right half...
             if (!rightLock) {