Update app.py

app.py

@@ -87,8 +87,7 @@ def compare_texts(reference_text, transcribed_text):
     sm = difflib.SequenceMatcher(None, reference_text, transcribed_text)
     similarity_score = round(sm.ratio() * 100, 2)
 
-    # Construct HTML output
-   # html_output = f"<strong>Fidelity Class:</strong> # Tạo output HTML với các mức đánh giá chi tiết hơn
+    # Construct HTML output with detailed fidelity class
     html_output = f"<strong>Fidelity Class:</strong> "
     if similarity_score >= 85:
         html_output += f"<strong>GOOD (>=85%)</strong><br>"
@@ -98,18 +97,17 @@ def compare_texts(reference_text, transcribed_text):
         html_output += f"<strong>NEEDS IMPROVEMENT (50% - 70%)</strong><br>"
     else:
         html_output += f"<strong>POOR (<50%)</strong><br>"
-    
+
     html_output += f"<strong>Quality Score:</strong> {similarity_score}%<br>"
     html_output += f"<strong>Transcribed Text:</strong> {transcribed_text}<br>"
     html_output += "<strong>Word Score List:</strong><br>"
 
-
     # Generate colored word score list
     for i, word in enumerate(reference_words):
         try:
             if word.lower() == transcribed_words[i].lower():
                 html_output += f'<span style="color: green;">{word}</span> '  # Correct words in green
-            elif difflib.get_close_matches(word, transcribed_words):
+            elif difflib.get_close_matches(word, [transcribed_words[i]]):
                 html_output += f'<span style="color: yellow;">{word}</span> '  # Close matches in yellow
             else:
                 # Incorrect words in red
@@ -117,21 +115,20 @@ def compare_texts(reference_text, transcribed_text):
                 # Create pronunciation audio for the incorrect word
                 audio_file_path = create_pronunciation_audio(word)
                 incorrect_words_audios.append((word, audio_file_path))
-                incorrect_words.append( audio_file_path)
         except IndexError:
-            html_output += f'<span style="color: red;">{word}</span> '  # Words in reference that were not transcribed
+            # Word in reference that was not transcribed
+            html_output += f'<span style="color: red;">{word}</span> '
 
     # Provide audio for incorrect words
     if incorrect_words_audios:
         html_output += "<br><strong>Pronunciation for Incorrect Words:</strong><br>"
-            for word, audio in incorrect_words_audios:
-                suggestion = difflib.get_close_matches(word, reference_words, n=1)
-                suggestion_text = f" (Did you mean: <em>{suggestion[0]}</em>?)" if suggestion else ""
-                html_output += f'{word}: '
-                html_output += f'<audio controls><source src="{audio}" type="audio/wav">Your browser does not support the audio tag.</audio>{suggestion_text}<br>'
-                
-
-    return [html_output, incorrect_words]
+        for word, audio in incorrect_words_audios:
+            suggestion = difflib.get_close_matches(word, reference_words, n=1)
+            suggestion_text = f" (Did you mean: <em>{suggestion[0]}</em>?)" if suggestion else ""
+            html_output += f'{word}: '
+            html_output += f'<audio controls><source src="{audio}" type="audio/wav">Your browser does not support the audio tag.</audio>{suggestion_text}<br>'
+
+    return [html_output, [audio for _, audio in incorrect_words_audios]]
 
 # Step 4: Text-to-Speech Function
 def text_to_speech(paragraph):