Update app.py

app.py

@@ -1,12 +1,11 @@
 import os
-import requests
 import speech_recognition as sr
 import difflib
 import gradio as gr
 from gtts import gTTS
-import io
-from pydub import AudioSegment
 import time
+from pydub import AudioSegment
+
 # Create audio directory if it doesn't exist
 if not os.path.exists('audio'):
     os.makedirs('audio')
@@ -49,7 +48,6 @@ def transcribe_audio(audio):
 def create_pronunciation_audio(word):
     time.sleep(5)  # Chờ 5 giây
     tts = gTTS(word)
-    main_url="https://mr2along-speech-recognize.hf.space/gradio_api/file="
     audio_file_path = f"audio/{word}.mp3"  # Save the audio to a file
     tts.save(audio_file_path)
     return audio_file_path  # Return the file path of the saved audio
@@ -64,8 +62,6 @@ def compare_texts(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
-   # html_output = f"<strong>Fidelity Class:</strong> # Tạo output HTML với các mức đánh giá chi tiết hơn
     html_output = f"<strong>Fidelity Class:</strong> "
     if similarity_score >= 85:
         html_output += f"<strong>GOOD (>=85%)</strong><br>"
@@ -76,9 +72,7 @@ def compare_texts(reference_text, transcribed_text):
     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 += f"<strong>Quality Score:</strong> {similarity_score}<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>"
 
@@ -94,15 +88,15 @@ html_output += f"<strong>Transcribed Text:</strong> {transcribed_text}<br>"
                 html_output += f'<span style="color: red;">{word}</span> '
                 # Create pronunciation audio for the incorrect word
                 audio_file_path = create_pronunciation_audio(word)
-                #incorrect_words_audios.append((word, audio_file_path))
-                incorrect_words_audios.append( audio_file_path)
+                incorrect_words_audios.append(audio_file_path)
         except IndexError:
             html_output += f'<span style="color: red;">{word}</span> '  # Words in reference that were not transcribed
 
     # Provide audio for incorrect words
     if incorrect_words_audios:
         html_output += "<br><strong>Pronunciation for Incorrect Words:</strong><br>"
-
+        for audio_path in incorrect_words_audios:
+            html_output += f'<audio controls><source src="{audio_path}" type="audio/mpeg">Your browser does not support the audio tag.</audio><br>'
 
     return [html_output, incorrect_words_audios]
 
@@ -120,12 +114,13 @@ def text_to_speech(paragraph):
 def gradio_function(paragraph, audio):
     # Transcribe the audio
     transcribed_text = transcribe_audio(audio)
+
     # Compare the original paragraph with the transcribed text
     comparison_result = compare_texts(paragraph, transcribed_text)
 
     # Return comparison result
     return comparison_result
-    
+
 # Gradio Interface using the updated API
 interface = gr.Interface(
     fn=gradio_function, 
@@ -133,7 +128,7 @@ interface = gr.Interface(
         gr.Textbox(lines=5, label="Input Paragraph"),
         gr.Audio(type="filepath", label="Record Audio")
     ], 
-    outputs=["html","files"],
+    outputs=["html", "files"],
     title="Speech Recognition Comparison",
     description="Input a paragraph, record your audio, and compare the transcription to the original text."
 )