lyrics-transcriber 0.37.0__py3-none-any.whl → 0.40.0__py3-none-any.whl

lyrics_transcriber/correction/handlers/extend_anchor.py

@@ -80,8 +80,19 @@ class ExtendAnchorHandler(GapCorrectionHandler):
                 confidence = len(matching_sources) / len(gap.reference_words)
                 sources = ", ".join(matching_sources)
 
-                # Calculate reference positions for matching sources
-                reference_positions = WordOperations.calculate_reference_positions(gap, matching_sources)
+                # Get base reference positions
+                base_reference_positions = WordOperations.calculate_reference_positions(gap, matching_sources)
+
+                # Adjust reference positions based on the word's position in the reference text
+                reference_positions = {}
+                for source in matching_sources:
+                    if source in base_reference_positions:
+                        # Find this word's position in the reference text
+                        ref_words = gap.reference_words[source]
+                        for ref_idx, ref_word in enumerate(ref_words):
+                            if ref_word.lower() == word.lower():
+                                reference_positions[source] = base_reference_positions[source] + ref_idx
+                                break
 
                 corrections.append(
                     WordOperations.create_word_replacement_correction(

lyrics_transcriber/correction/handlers/word_operations.py

@@ -23,9 +23,15 @@ class WordOperations:
 
             for source in sources_to_check:
                 if source in gap.preceding_anchor.reference_positions:
-                    # Calculate position based on anchor position and offset
+                    # Calculate base position from anchor
                     anchor_pos = gap.preceding_anchor.reference_positions[source]
-                    ref_pos = anchor_pos + len(gap.preceding_anchor.words)
+                    base_ref_pos = anchor_pos + len(gap.preceding_anchor.words)
+
+                    # Calculate word offset within the gap
+                    word_offset = gap.words.index(gap.words[gap.transcription_position - gap.transcription_position])
+
+                    # Add word offset to base position
+                    ref_pos = base_ref_pos + word_offset
                     reference_positions[source] = ref_pos
         return reference_positions