Update app.py

app.py

@@ -1,160 +1,87 @@
-import os
 import gradio as gr
-import json, re
 import openai
-from datetime import datetime
-from pydub import AudioSegment
-import pdfplumber, docx, pandas as pd
+import os
+import pdfplumber
+import docx
+import pandas as pd
 from PIL import Image
-import whisper  # Local whisper for transcription
-
-openai.api_key = os.environ.get("GROQ_API_KEY")
-openai.api_base = "https://api.groq.com/openai/v1"
-
-model_whisper = whisper.load_model("base")
-
-# Chat with Groq
-def chat_with_groq(message, state):
-    if state is None:
-        state = {"oai_history": [], "chatbot_ui": []}
-
-    oai_history = state["oai_history"]
-    chatbot_ui = state["chatbot_ui"]
-
-    if not oai_history:
-        oai_history.append({"role": "system", "content": "You are Neobot 🤖 – helpful and concise."})
-
-    oai_history.append({"role": "user", "content": message})
-    try:
-        response = openai.ChatCompletion.create(
-            model="llama3-70b-8192",
-            messages=oai_history
-        )
-        reply = response["choices"][0]["message"]["content"]
-    except Exception as e:
-        reply = f"❌ Error: {e}"
-
-    oai_history.append({"role": "assistant", "content": reply})
-    chatbot_ui.append({"role": "user", "content": message})
-    chatbot_ui.append({"role": "assistant", "content": reply})
-
-    return "", chatbot_ui, {"oai_history": oai_history, "chatbot_ui": chatbot_ui}
-
-# Save chat
-def save_session(state):
-    if not state or not state.get("oai_history"):
-        return "❌ No chat to save"
-
-    oai_history = state["oai_history"]
-    prompt = next((m["content"] for m in oai_history if m["role"] == "user"), "chat")
-    title = re.sub(r"[^\w\s]", "", prompt).strip()
-    title = " ".join(title.split()[:6]) or "Chat"
-    timestamp = datetime.now().strftime("%b %d %Y %H-%M")
-    filename = f"{title} - {timestamp}.json"
-    with open(filename, "w", encoding="utf-8") as f:
-        json.dump(oai_history, f, indent=2, ensure_ascii=False)
-    return f"✅ Chat saved as: {filename[:-5]}"
-
-def list_saved_files():
-    return sorted([f[:-5] for f in os.listdir() if f.endswith(".json")])
-
-def load_chat(name):
-    filename = f"{name}.json"
-    try:
-        with open(filename, "r", encoding="utf-8") as f:
-            oai_history = json.load(f)
-        chatbot_ui = []
-        for i in range(0, len(oai_history)):
-            if oai_history[i]["role"] == "user":
-                chatbot_ui.append({"role": "user", "content": oai_history[i]["content"]})
-            elif oai_history[i]["role"] == "assistant":
-                chatbot_ui.append({"role": "assistant", "content": oai_history[i]["content"]})
-        return chatbot_ui, {"oai_history": oai_history, "chatbot_ui": chatbot_ui}, f"✅ Loaded {name}"
-    except Exception as e:
-        return [], {"oai_history": [], "chatbot_ui": []}, f"❌ Could not load {name}: {e}"
-
-# Transcription using local Whisper
-def transcribe_audio(file):
-    if not file:
-        return ""
+import pytesseract
+from pydub import AudioSegment
+import tempfile
+import sounddevice as sd
+import scipy.io.wavfile as wav
+
+openai.api_key = os.getenv("OPENAI_API_KEY")
+
+def extract_text(file):
+    ext = file.name.split(".")[-1].lower()
+    if ext == "pdf":
+        with pdfplumber.open(file.name) as pdf:
+            return "\n".join([page.extract_text() for page in pdf.pages if page.extract_text()])
+    elif ext in ["doc", "docx"]:
+        doc = docx.Document(file.name)
+        return "\n".join([p.text for p in doc.paragraphs])
+    elif ext in ["xls", "xlsx", "csv"]:
+        df = pd.read_excel(file.name) if ext != "csv" else pd.read_csv(file.name)
+        return df.to_string()
+    elif ext in ["jpg", "jpeg", "png"]:
+        image = Image.open(file.name)
+        text = pytesseract.image_to_string(image)
+        return text or "❌ No text found in image."
+    else:
+        return "❌ Unsupported file format."
+
+def transcribe_audio(audio_path):
     try:
-        audio = whisper.load_audio(file)
-        audio = whisper.pad_or_trim(audio)
-        mel = whisper.log_mel_spectrogram(audio).to(model_whisper.device)
-        result = model_whisper.decode(mel)
-        return result.text
+        with open(audio_path, "rb") as f:
+            transcript = openai.Audio.transcribe("whisper-1", f)
+        return transcript["text"]
     except Exception as e:
-        return f"❌ Transcription error: {e}"
+        return f"❌ Transcription error: {str(e)}"
 
-# File processor – send file content to textbox for user to ask about
-def file_to_message(file, state):
-    if file is None:
-        return "", state
-    filename = file.name
-    ext = filename.split(".")[-1].lower()
+def generate_response(messages):
     try:
-        if ext == "pdf":
-            with pdfplumber.open(file.name) as pdf:
-                text = "\n".join([p.extract_text() for p in pdf.pages if p.extract_text()])
-        elif ext == "docx":
-            doc = docx.Document(file.name)
-            text = "\n".join([p.text for p in doc.paragraphs])
-        elif ext in ["csv", "xlsx"]:
-            df = pd.read_csv(file.name) if ext == "csv" else pd.read_excel(file.name)
-            text = df.to_string()
-        elif ext in ["png", "jpg", "jpeg"]:
-            text = "[Image uploaded: please describe what you'd like me to do with it.]"
-        else:
-            text = file.read().decode("utf-8")
-
-        if len(text) > 8000:
-            text = text[:8000]
-
-        return text, state
+        response = openai.ChatCompletion.create(
+            model="gpt-3.5-turbo",
+            messages=messages
+        )
+        return response.choices[0].message["content"]
     except Exception as e:
-        return f"❌ Error processing file: {e}", state
+        return f"❌ Error: {str(e)}"
 
-# UI
-with gr.Blocks(css="""
-body { background: #fff; color: #000; }
-#title { font-size: 2rem; text-align: center; margin-top: 1rem; }
-.gr-chatbot { height: 70vh !important; }
-textarea, input[type='text'] { background: #f0f0f0; border-radius: 8px; }
-.gr-button { border-radius: 8px; background: #000; color: white; }
-""") as demo:
+chat_history = []
 
-    state = gr.State({"oai_history": [], "chatbot_ui": []})
+def chat(user_message, file=None, image=None, mic_audio=None):
+    if mic_audio:
+        user_message = transcribe_audio(mic_audio)
+    elif file:
+        user_message = extract_text(file)
+    elif image:
+        user_message = extract_text(image)
 
-    gr.Markdown("# 🤖 Neobot - Always Listening", elem_id="title")
+    chat_history.append({"role": "user", "content": user_message})
+    bot_response = generate_response(chat_history)
+    chat_history.append({"role": "assistant", "content": bot_response})
+    return bot_response
 
-    chatbot = gr.Chatbot(type="messages")
+with gr.Blocks() as demo:
+    gr.Markdown("### 🎧 **Neobot - Always Listening**")
+    chatbot = gr.Chatbot(height=300)
 
     with gr.Row():
-        chat_input = gr.Textbox(placeholder="Type here or use mic...", scale=6, show_label=False)
+        txt = gr.Textbox(placeholder="Type here or use mic...", scale=4)
         send_btn = gr.Button("🚀", scale=1)
 
     with gr.Row():
-        mic_audio = gr.Audio(type="filepath", label="🎙️ Record Voice", interactive=True)
-        mic_audio.change(transcribe_audio, [mic_audio], [chat_input])
-
-    with gr.Row():
-        file_upload = gr.File(label="📎 Upload file", file_types=[".pdf", ".docx", ".txt", ".csv", ".xlsx", ".jpg", ".png"])
-        file_upload.change(file_to_message, [file_upload, state], [chat_input, state])
-
-    with gr.Row():
-        new_btn = gr.Button("🆕 New Chat")
-        save_btn = gr.Button("💾 Save Chat")
-        dropdown = gr.Dropdown(label="📂 Saved Chats", choices=list_saved_files(), interactive=True)
-        load_btn = gr.Button("📥 Load Chat")
-
-    status = gr.Markdown()
+        mic_audio = gr.Audio(type="filepath", label="🎤 Record Voice", interactive=True)
+        upload_file = gr.File(label="📎 Upload File")
+        upload_img = gr.Image(type="filepath", label="🖼️ Upload Image")
 
-    send_btn.click(chat_with_groq, [chat_input, state], [chat_input, chatbot, state])
-    chat_input.submit(chat_with_groq, [chat_input, state], [chat_input, chatbot, state])
+    def handle_input(message, file, image, mic_audio):
+        reply = chat(message, file, image, mic_audio)
+        return chatbot.update(chatbot.value + [[message, reply]])
 
-    new_btn.click(lambda: ("", [], {"oai_history": [], "chatbot_ui": []}), [], [chat_input, chatbot, state])
-    save_btn.click(save_session, [state], [status])
-    save_btn.click(lambda: gr.update(choices=list_saved_files()), [], [dropdown])
-    load_btn.click(load_chat, [dropdown], [chatbot, state, status])
+    send_btn.click(handle_input, inputs=[txt, upload_file, upload_img, mic_audio], outputs=chatbot)
+    txt.submit(handle_input, inputs=[txt, upload_file, upload_img, mic_audio], outputs=chatbot)
 
 demo.launch()