Update app.py

app.py

@@ -29,6 +29,8 @@ import gradio as gr
 
 import shutil
 
+import speech_recognition as sr
+
 
 # 오디오 변환 mp4 --> wav
 def extract_audio_from_video(video_file_path, audio_file_path):
@@ -134,6 +136,28 @@ def real_fake_check(list_dir, path, model):
     return {'real':r_cnt, 'fake':f_cnt, 'prob': prob}
 
 
+# 음성에서 text 추출
+def convert_wav_to_text(wav_file_path):
+    recognizer = sr.Recognizer()
+
+    # WAV 파일 열기
+    with sr.AudioFile(wav_file_path) as source:
+        print("WAV 파일에서 오디오를 로드 중...")
+        audio_data = recognizer.record(source)  # 전체 오디오를 녹음
+
+        try:
+            # Google Web Speech API로 텍스트 변환
+
+            text = recognizer.recognize_google(audio_data)
+
+        except sr.UnknownValueError:
+            text = 'error'
+        except sr.RequestError as e:
+            text = 'error'
+
+    return text
+
+
 def main(file_name):
     
     if os.path.exists('/tmp/wav'):
@@ -171,11 +195,19 @@ def main(file_name):
     fake = os.listdir(fake_path)
 
     rf_check = real_fake_check(fake, fake_path,model)       #fake dataset\
-    return rf_check
+
+    text_list =[]
+    to_text = os.listdir('/tmp/wav')
+    for i in to_text:
+        text = convert_wav_to_text(os.path.join(path, i))
+        text_list.append(text)
+
+        
+    return rf_check, text_list
 
 def deepvoice_check(video_file):
-    results = main(video_file)
-    return results
+    results,text = main(video_file)
+    return results,text
 
 # Gradio 인터페이스 생성
 deepfake = gr.Interface(