Upload 3 files

app.py

@@ -0,0 +1,206 @@
+import os
+import json
+import requests
+import gradio as gr
+from bs4 import BeautifulSoup
+from groq import Groq
+from youtube_transcript_api import YouTubeTranscriptApi
+from dotenv import load_dotenv
+
+load_dotenv()
+
+# --- API KEYS ---
+BRIGHTDATA_API_KEY = os.getenv("BRIGHTDATA_API_KEY")
+GROQ_API_KEY = os.getenv("GROQ_API_KEY")
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+
+# --- Clients ---
+client = Groq(api_key=GROQ_API_KEY)
+openai_client = None
+if OPENAI_API_KEY:
+    from openai import OpenAI
+    openai_client = OpenAI(api_key=OPENAI_API_KEY)
+
+# --- Persistent Storage ---
+HISTORY_FILE = "chat_history.json"
+
+if os.path.exists(HISTORY_FILE):
+    try:
+        with open(HISTORY_FILE, "r") as f:
+            conversation_history = json.load(f)
+            if not isinstance(conversation_history, list):
+                conversation_history = []
+    except (json.JSONDecodeError, Exception):
+        conversation_history = []
+else:
+    conversation_history = []
+
+# ----------------------
+# LLM Wrapper
+# ----------------------
+def ask_llm(query, context=None):
+    system_prompt = """
+You are a helpful AI assistant.
+Use ONLY the provided context and conversation history to answer the question.
+If the answer is not found in the context, respond clearly that you don't know based on the provided info.
+"""
+
+    messages = [{"role": "system", "content": system_prompt}]
+    
+    # Add context if available
+    if context:
+        messages.append({"role": "system", "content": f"CONTEXT:\n{context}"})
+    
+    messages.extend(conversation_history)
+    messages.append({"role": "user", "content": query})
+
+    try:
+        response = client.chat.completions.create(
+            model="llama-3.1-8b-instant",
+            messages=messages,
+            temperature=0.3
+        )
+        answer = response.choices[0].message.content
+        
+        # Update conversation history
+        conversation_history.append({"role": "user", "content": query})
+        conversation_history.append({"role": "assistant", "content": answer})
+
+        # Save to persistent storage
+        with open(HISTORY_FILE, "w") as f:
+            json.dump(conversation_history, f, indent=2)
+
+        return answer
+    except Exception as e:
+        return f"Error communicating with LLM: {str(e)}"
+
+# ----------------------
+# Website Scraper
+# ----------------------
+def scrape_website(url, question):
+    try:
+        headers = {"Authorization": f"Bearer {BRIGHTDATA_API_KEY}"}
+        payload = {"zone": "web_unlocker1", "url": url, "format": "raw"}
+
+        response = requests.post(
+            "https://api.brightdata.com/request",
+            headers=headers,
+            json=payload,
+            timeout=60
+        )
+        if response.status_code != 200:
+            return f"Bright Data Error: {response.status_code}"
+
+        soup = BeautifulSoup(response.text, "html.parser")
+        text = soup.get_text(separator=" ", strip=True)
+
+        if not text:
+            return "⚠️ Could not extract content from the website."
+
+        return ask_llm(question, context=text[:12000])
+
+    except Exception as e:
+        return f"Error scraping website: {str(e)}"
+
+# ----------------------
+# YouTube Transcript Q&A
+# ----------------------
+def youtube_qa(video_id, question):
+    try:
+        transcript = YouTubeTranscriptApi.get_transcript(video_id)
+        full_text = " ".join([entry["text"] for entry in transcript])
+
+        if not full_text.strip():
+            return "⚠️ No transcript text found for this video."
+
+        return ask_llm(question, context=full_text[:12000])
+
+    except Exception:
+        return "❌ Could not retrieve transcript. Invalid video ID or no transcript available."
+
+# ----------------------
+# Voice Chat (STT + TTS)
+# ----------------------
+def voice_chat(audio_file):
+    if not audio_file:
+        return "", "⚠️ No audio provided.", None
+
+    # Transcribe audio using Groq (since model is whisper-large-v3)
+    try:
+        with open(audio_file, "rb") as f:
+            transcription = client.audio.transcriptions.create(
+                file=f,
+                model="whisper-large-v3"
+            )
+        user_text = transcription.text
+    except Exception as e:
+        return "", f"❌ Could not transcribe audio: {e}", None
+
+    # Ask LLM
+    answer_text = ask_llm(user_text)
+
+    # Convert answer to speech using OpenAI TTS
+    audio_path = "temp_audio/output.mp3"
+    try:
+        if not openai_client:
+            return user_text, f"{answer_text}\n\n(Voice output unavailable - OpenAI key missing)", None
+            
+        tts_response = openai_client.audio.speech.create(
+            model="tts-1",
+            voice="alloy",
+            input=answer_text[:4096] # Limit input for TTS
+        )
+        
+        os.makedirs("temp_audio", exist_ok=True)
+        with open(audio_path, "wb") as f:
+            f.write(tts_response.content)
+    except Exception as e:
+        return user_text, f"{answer_text}\n\n❌ Could not generate audio: {e}", None
+
+    return user_text, answer_text, audio_path
+
+# ----------------------
+# Gradio Interface
+# ----------------------
+with gr.Blocks() as demo:
+    gr.Markdown("# 🤖 Multimodal AI Assistant (Voice + Text)")
+
+    with gr.Tabs():
+        # Tab 1: Website Q&A
+        with gr.Tab("🌐 Website Q&A"):
+            url_input = gr.Textbox(label="Enter Website URL")
+            website_question = gr.Textbox(label="Ask a Question")
+            website_output = gr.Textbox(label="Answer")
+            website_btn = gr.Button("Ask")
+            website_btn.click(
+                scrape_website,
+                inputs=[url_input, website_question],
+                outputs=website_output
+            )
+
+        # Tab 2: YouTube Transcript Q&A
+        with gr.Tab("🎥 YouTube Transcript Q&A"):
+            video_id_input = gr.Textbox(label="Enter YouTube Video ID")
+            youtube_question = gr.Textbox(label="Ask a Question")
+            youtube_output = gr.Textbox(label="Answer")
+            youtube_btn = gr.Button("Ask")
+            youtube_btn.click(
+                youtube_qa,
+                inputs=[video_id_input, youtube_question],
+                outputs=youtube_output
+            )
+
+        # Tab 3: Voice Chat
+        with gr.Tab("🎤 Voice Chat"):
+            audio_input = gr.Audio(sources=["microphone"], type="filepath", label="Record your question")
+            voice_text_output = gr.Textbox(label="Transcribed Text")
+            voice_answer_output = gr.Textbox(label="AI Answer")
+            voice_audio_output = gr.Audio(label="AI Voice Response", autoplay=True)
+            voice_btn = gr.Button("Ask")
+            voice_btn.click(
+                voice_chat,
+                inputs=[audio_input],
+                outputs=[voice_text_output, voice_answer_output, voice_audio_output]
+            )
+
+demo.launch()

chat_history.json

@@ -0,0 +1,10 @@
+[
+  {
+    "role": "user",
+    "content": " Hey, how are you? I'm looking to meet you soon."
+  },
+  {
+    "role": "assistant",
+    "content": "I'm just a computer program, so I don't have feelings like humans do, but I'm functioning properly and ready to help. As for meeting you in person, I'm a large language model, I don't have a physical presence, so we can only interact through text-based conversations like this one. How can I assist you today?"
+  }
+]

requirements.txt

@@ -0,0 +1,9 @@
+gradio
+requests>=2.31
+beautifulsoup4>=4.12
+pandas>=2.0
+groq>=0.0.3
+youtube-transcript-api>=0.6.0
+python-dotenv>=1.0.0
+openai>=0.28
+soundfile>=0.12.1