Update app.py

app.py

@@ -1,49 +1,38 @@
+import re
 import gradio as gr
-from langchain_community.document_loaders import YoutubeLoader
-from youtube_transcript_api import NoTranscriptFound
-from pytube.exceptions import PytubeError
+from youtube_transcript_api import YouTubeTranscriptApi
 
-def load_youtube_video_info(url):
-    """
-    Load YouTube video information including metadata and transcript.
-    
-    Args:
-        url (str): URL of the YouTube video.
-        
-    Returns:
-        str: Information about the video, including metadata and transcript or error messages.
-    """
+def extract_video_id(youtube_url):
+    # Regex para extrair o ID do vídeo da URL do YouTube
+    match = re.search(r'(?:youtu\.be\/|(?:www\.)?youtube\.com\/(?:watch\?v=|embed\/|v\/|.+\?v=))([^&]{11})', youtube_url)
+    return match.group(1) if match else None
+
+def get_transcript(video_id, language="en"):
     try:
-        # Initialize YouTube loader with preferred language as Portuguese
-        loader = YoutubeLoader.from_youtube_url(url, add_video_info=True, language="pt")
-        
-        # Load the video information and transcript
-        result = loader.load()
+        transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=[language])
+        return "\n".join([f"{t['start']}: {t['text']}" for t in transcript])
+    except Exception as e:
+        return f"Erro ao obter transcrição: {str(e)}"
 
-        if result and len(result) > 0:
-            # Retrieve metadata information
-            author = result[0].metadata.get('author', 'Unknown Author')
-            length = result[0].metadata.get('length', 'Unknown Length')
-            title = result[0].metadata.get('title', 'Unknown Title')
-            
-            # Check if transcript is available
-            transcript = result[0].page_content if hasattr(result[0], 'page_content') else "No transcript available."
+# Função que combina a extração do ID e a obtenção da transcrição
+def gradio_interface(youtube_url, language):
+    video_id = extract_video_id(youtube_url)
+    if not video_id:
+        return "Erro: URL inválida. Por favor, insira um link válido do YouTube."
+    
+    return get_transcript(video_id, language)
 
-            # Format information
-            info = (
-                f"Video Title: {title}\n"
-                f"Author: {author}\n"
-                f"Length: {length}\n\n"
-                f"Transcript:\n{transcript}"
-            )
-        else:
-            info = "No result found."
+# Criação da interface Gradio
+iface = gr.Interface(
+    fn=gradio_interface,
+    inputs=[
+        gr.Textbox(label="URL do Vídeo (YouTube)", placeholder="Ex: https://www.youtube.com/watch?v=tl1jHm0qC_4"),
+        gr.Dropdown(label="Idioma", choices=["en", "pt", "es", "fr"], value="en")
+    ],
+    outputs="text",
+    title="Obter Transcrição de Vídeos do YouTube",
+    description="Insira a URL de um vídeo do YouTube e selecione o idioma da transcrição."
+)
 
-    except NoTranscriptFound:
-        info = "Transcript not found for this video."
-    except PytubeError as e:
-        info = f"Error loading video information: {str(e)}"
-    except Exception as e:
-        info = f"An unexpected error occurred: {str(e)}"
-    
-    return info
+# Execução da interface Gradio
+iface.launch()