Update app.py

app.py

@@ -1,102 +1,75 @@
 import os
 import gradio as gr
-import pdfplumber
-import docx
-import pandas as pd
-from PIL import Image
-from io import BytesIO
-import base64
-import whisper
 from openai import OpenAI
+import tempfile
+import torch
+import whisper
 
-# Load Whisper model
-whisper_model = whisper.load_model("base")
-
-
-
-
-# Load Groq API key
+# Load API key from environment
 GROQ_API_KEY = os.environ.get("GROQ_API_KEY")
 
-# Initialize OpenAI client with Groq base URL
+# Initialize client for Groq-compatible OpenAI API
 client = OpenAI(
     api_key=GROQ_API_KEY,
     base_url="https://api.groq.com/openai/v1"
 )
 
+# Load Whisper model
+whisper_model = whisper.load_model("base")
+
+# Chat history storage
+chat_history = []
 
-def extract_text_from_file(file):
-    if file.name.endswith(".pdf"):
-        with pdfplumber.open(file.name) as pdf:
-            text = "\n".join(page.extract_text() for page in pdf.pages if page.extract_text())
-    elif file.name.endswith(".docx"):
-        doc = docx.Document(file.name)
-        text = "\n".join(p.text for p in doc.paragraphs)
-    elif file.name.endswith(".xlsx"):
-        df = pd.read_excel(file.name)
-        text = df.to_string()
-    elif file.name.endswith((".png", ".jpg", ".jpeg")):
-        img = Image.open(file.name)
-        buffer = BytesIO()
-        img.save(buffer, format="PNG")
-        encoded = base64.b64encode(buffer.getvalue()).decode("utf-8")
-        text = f"[Image uploaded: data:image/png;base64,{encoded[:100]}... (truncated)]"
-    else:
-        with open(file.name, "r", encoding="utf-8", errors="ignore") as f:
-            text = f.read()
-    return text
-
-def transcribe_audio(audio_path):
-    result = whisper_model.transcribe(audio_path)
-    return result["text"]
-
-def generate_reply(history):
-    messages = [{"role": "system", "content": "You are a helpful assistant."}]
-    for user_msg, bot_msg in history:
-        messages.append({"role": "user", "content": user_msg})
-        messages.append({"role": "assistant", "content": bot_msg})
+def chat_with_bot(message, history):
+    global chat_history
+    chat_history = history or []
 
+    # Append user message to chat history
+    chat_history.append({"role": "user", "content": message})
+
+    # Call Groq LLM
     response = client.chat.completions.create(
         model="llama3-8b-8192",
-        messages=messages,
-        temperature=0.7
+        messages=chat_history
     )
+
     reply = response.choices[0].message.content
-    return reply
 
-def respond(message, history):
-    reply = generate_reply(history + [[message, ""]])
-    history.append([message, reply])
-    return history, ""
+    # Append assistant reply to chat history
+    chat_history.append({"role": "assistant", "content": reply})
+
+    # Prepare output format for Gradio (list of tuples)
+    formatted_history = [(m['content'], chat_history[i+1]['content']) for i, m in enumerate(chat_history[:-1]) if m['role'] == 'user']
+    return formatted_history, chat_history
 
-def handle_file_upload(file, message):
-    if file is None:
-        return message
-    file_content = extract_text_from_file(file)
-    return f"{message}\n\n--- File Content Start ---\n{file_content}\n--- File Content End ---"
+def transcribe_audio(audio_file):
+    if audio_file is None:
+        return ""
+    audio = whisper.load_audio(audio_file)
+    audio = whisper.pad_or_trim(audio)
+    mel = whisper.log_mel_spectrogram(audio).to(whisper_model.device)
+    options = whisper.DecodingOptions(fp16=torch.cuda.is_available())
+    result = whisper.decode(whisper_model, mel, options)
+    return result.text
 
-def handle_audio_upload(audio, message):
-    if audio is None:
-        return message
-    transcription = transcribe_audio(audio)
-    return f"{message}\n\n--- Transcription ---\n{transcription}"
+with gr.Blocks() as demo:
+    gr.Markdown("# 🤖 Neobot - Chat with Voice, File & Text")
 
-with gr.Blocks(css="body { background-color: white; color: black }") as demo:
-    gr.Markdown("<h1 style='text-align: center;'>Neobot</h1>")
-    chatbot = gr.Chatbot(label="Chat", elem_id="chatbox", height=450, type="messages")
+    chatbot = gr.Chatbot()
+    state = gr.State([])
 
     with gr.Row():
-        txt = gr.Textbox(placeholder="Type a message or edit transcribed/file content here...", scale=5, show_label=False)
-        send_btn = gr.Button("Send", scale=1)
+        txt = gr.Textbox(placeholder="Type a message or upload audio/file...", show_label=False)
+        send_btn = gr.Button("Send")
 
     with gr.Row():
-        upload_btn = gr.File(label="📎 Upload File", file_types=[".pdf", ".docx", ".txt", ".xlsx", ".png", ".jpg", ".jpeg"])
-        audio_in = gr.Audio(label="🎙️ Upload Audio", type="filepath")
+        audio_upload = gr.Audio(source="upload", type="filepath", label="Upload Audio")
+        transcribe_btn = gr.Button("Transcribe Audio")
 
-    history = gr.State([])
+    # Chat event
+    send_btn.click(chat_with_bot, inputs=[txt, state], outputs=[chatbot, state])
 
-    send_btn.click(respond, [txt, history], [chatbot, txt])
-    upload_btn.change(handle_file_upload, [upload_btn, txt], txt)
-    audio_in.change(handle_audio_upload, [audio_in, txt], txt)
+    # Audio transcription event
+    transcribe_btn.click(transcribe_audio, inputs=audio_upload, outputs=txt)
 
 demo.launch()