Update app.py

app.py

@@ -1,33 +1,25 @@
 # app.py
 """
 Multi-Mode AI Assistant (Voice, PDF, Image)
-- Fixed Gradio v4+ Audio usage (no source=...).
-- Chatbot uses type="messages" (openai-style {"role","content"} dicts).
-- Voice tab: single mic + three buttons (Ask General / Ask PDF / Ask Image).
-- PDF tab: upload + text questions only (no voice controls).
-- PDF & Image summary download fixed (now outputs same text as Answer box).
-- OCR uses OCR.space (OCR_SPACE_API_KEY).
-- Uses Groq endpoints for transcription + chat completions (GROQ_API_KEY).
-- Embeddings via sentence-transformers (all-MiniLM-L6-v2).
-- Improved interactive UI with attractive layout.
+- Fixed gr.File issue: removed unsupported 'file_types_preview' argument
+- Interactive, compact, visually appealing UI
 """
 import os
 import uuid
 import tempfile
 import requests
+from datetime import datetime
 from dotenv import load_dotenv
 from gtts import gTTS
 from PyPDF2 import PdfReader
 import gradio as gr
 from sentence_transformers import SentenceTransformer, util
 from fpdf import FPDF
-from datetime import datetime
 
 # ------------------ Load API Keys ------------------
 load_dotenv()
 GROQ_API_KEY = os.getenv("GROQ_API_KEY", "").strip()
 OCR_SPACE_API_KEY = os.getenv("OCR_SPACE_API_KEY", "").strip()
-
 if not GROQ_API_KEY:
     raise ValueError("❌ GROQ_API_KEY missing. Set it in env / Hugging Face Secrets.")
 if not OCR_SPACE_API_KEY:
@@ -35,22 +27,36 @@ if not OCR_SPACE_API_KEY:
 
 HEADERS = {"Authorization": f"Bearer {GROQ_API_KEY}"}
 
-# ------------------ Global State ------------------
+# ------------------ Global States ------------------
 SESSION_HISTORY = {}
+CHAT_DISPLAY = {}
 PDF_CONTENT = {}
 PDF_EMBEDS = {}
 IMAGE_TEXT = {}
 IMAGE_EMBEDS = {}
 CHUNK_SIZE = 1500
 
-# Load embedding model once
 embed_model = SentenceTransformer("all-MiniLM-L6-v2")
 
-
 # ------------------ Helpers ------------------
-def chunk_text(text, size=CHUNK_SIZE):
-    return [text[i:i + size] for i in range(0, len(text), size)]
+def _get_path_from_gr_file(gr_file):
+    if not gr_file:
+        return None
+    if isinstance(gr_file, str) and os.path.exists(gr_file):
+        return gr_file
+    try:
+        if hasattr(gr_file, "name") and os.path.exists(gr_file.name):
+            return gr_file.name
+    except:
+        pass
+    if isinstance(gr_file, dict):
+        for key in ("name", "file_name", "filepath"):
+            if key in gr_file and os.path.exists(gr_file[key]):
+                return gr_file[key]
+    return None
 
+def chunk_text(text, size=CHUNK_SIZE):
+    return [text[i:i+size] for i in range(0, len(text), size)]
 
 def synthesize_speech(text, lang="en"):
     try:
@@ -63,7 +69,6 @@ def synthesize_speech(text, lang="en"):
         print("TTS error:", e)
         return None
 
-
 def select_relevant_chunk(question, chunks, chunk_embeds):
     if not chunks or chunk_embeds is None:
         return ""
@@ -72,7 +77,6 @@ def select_relevant_chunk(question, chunks, chunk_embeds):
     top_idx = int(scores.argmax().item())
     return chunks[top_idx]
 
-
 def _chat_display_to_messages(chat_display):
     msgs = []
     for user, assistant in chat_display:
@@ -80,27 +84,12 @@ def _chat_display_to_messages(chat_display):
         msgs.append({"role": "assistant", "content": assistant})
     return msgs
 
+def _append_chat_display(session_id, user_text, assistant_text):
+    if session_id not in CHAT_DISPLAY:
+        CHAT_DISPLAY[session_id] = []
+    CHAT_DISPLAY[session_id].append((user_text, assistant_text))
 
-def _get_path_from_gr_file(gr_file):
-    if not gr_file:
-        return None
-    if isinstance(gr_file, str) and os.path.exists(gr_file):
-        return gr_file
-    try:
-        if hasattr(gr_file, "name") and isinstance(gr_file.name, str) and os.path.exists(gr_file.name):
-            return gr_file.name
-    except Exception:
-        pass
-    if isinstance(gr_file, dict):
-        for key in ("name", "file_name", "filepath"):
-            if key in gr_file:
-                candidate = gr_file.get(key)
-                if isinstance(candidate, str) and os.path.exists(candidate):
-                    return candidate
-    return None
-
-
-# ------------------ Transcription & LLM ------------------
+# ------------------ Voice & LLM ------------------
 def transcribe_audio(audio_path):
     if not audio_path or not os.path.exists(audio_path):
         return "Error: audio file missing."
@@ -116,7 +105,6 @@ def transcribe_audio(audio_path):
         print("transcription error:", e)
         return f"Error transcribing audio: {e}"
 
-
 def generate_response(session_id, user_text):
     if session_id not in SESSION_HISTORY:
         SESSION_HISTORY[session_id] = []
@@ -133,7 +121,6 @@ def generate_response(session_id, user_text):
         print("generate_response error:", e)
         return f"Error generating response: {e}"
 
-
 # ------------------ PDF Handling ------------------
 def handle_pdf_upload(pdf_file, session_id):
     path = _get_path_from_gr_file(pdf_file)
@@ -141,20 +128,17 @@ def handle_pdf_upload(pdf_file, session_id):
         return "No file uploaded or file unreadable."
     try:
         reader = PdfReader(path)
-        text = ""
-        for page in reader.pages:
-            text += (page.extract_text() or "") + "\n"
+        text = "".join([page.extract_text() or "" for page in reader.pages])
         if not text.strip():
             return "No extractable content found in PDF."
         chunks = chunk_text(text)
         PDF_CONTENT[session_id] = chunks
         PDF_EMBEDS[session_id] = embed_model.encode(chunks, convert_to_tensor=True)
-        return f"PDF processed: {len(chunks)} chunks ready."
+        return f"PDF uploaded: {len(chunks)} chunks ready."
     except Exception as e:
         print("PDF upload error:", e)
         return f"Error processing PDF: {e}"
 
-
 def handle_text_pdf(question, session_id):
     if session_id not in PDF_CONTENT:
         return "Document not found. Upload first."
@@ -172,7 +156,56 @@ def handle_text_pdf(question, session_id):
         print("PDF question error:", e)
         return f"Error generating response: {e}"
 
+# ------------------ Image OCR ------------------
+def ocr_space_file(image_path, api_key, language="eng"):
+    if not image_path or not os.path.exists(image_path):
+        return ""
+    try:
+        with open(image_path, "rb") as f:
+            payload = {"apikey": api_key, "language": language}
+            files = {"file": f}
+            r = requests.post("https://api.ocr.space/parse/image", files=files, data=payload, timeout=60)
+        r.raise_for_status()
+        j = r.json()
+        if j.get("IsErroredOnProcessing"):
+            print("OCR.space error:", j)
+            return ""
+        parsed = [pr.get("ParsedText", "") for pr in j.get("ParsedResults", [])]
+        return "\n".join(parsed)
+    except Exception as e:
+        print("ocr_space_file error:", e)
+        return ""
+
+def handle_image_upload(image_file, session_id):
+    path = _get_path_from_gr_file(image_file)
+    if not path:
+        return "No image uploaded.", ""
+    parsed = ocr_space_file(path, OCR_SPACE_API_KEY)
+    if not parsed.strip():
+        return "No extractable text found in the image.", ""
+    chunks = chunk_text(parsed)
+    IMAGE_TEXT[session_id] = chunks
+    IMAGE_EMBEDS[session_id] = embed_model.encode(chunks, convert_to_tensor=True)
+    return f"Image processed: {len(chunks)} chunks ready.", ""
+
+def handle_text_image(question, session_id):
+    if session_id not in IMAGE_TEXT:
+        return "Image not found. Upload first."
+    chunk = select_relevant_chunk(question, IMAGE_TEXT[session_id], IMAGE_EMBEDS[session_id])
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant summarizing image text."},
+        {"role": "user", "content": f"Image chunk:\n{chunk}\n\nQuestion: {question}"}
+    ]
+    body = {"model": "llama-3.1-8b-instant", "messages": messages}
+    try:
+        resp = requests.post("https://api.groq.com/openai/v1/chat/completions", headers=HEADERS, json=body, timeout=60)
+        resp.raise_for_status()
+        return resp.json()["choices"][0]["message"]["content"]
+    except Exception as e:
+        print("Image question error:", e)
+        return f"Error generating response: {e}"
 
+# ------------------ PDF Generation ------------------
 def generate_pdf_file(text, filename_prefix="summary"):
     pdf = FPDF()
     pdf.add_page()
@@ -180,24 +213,58 @@ def generate_pdf_file(text, filename_prefix="summary"):
     pdf.set_font("Arial", size=12)
     for line in text.split("\n"):
         pdf.multi_cell(0, 6, line)
-    file_path = f"/tmp/{filename_prefix}_{uuid.uuid4()}.pdf"
-    pdf.output(file_path)
-    return file_path
-
+    tmp_path = f"/tmp/{filename_prefix}_{uuid.uuid4()}.pdf"
+    pdf.output(tmp_path)
+    return tmp_path
 
 def download_pdf_summary(session_pdf_id):
-    summary_text = "\n".join([m["content"] for m in SESSION_HISTORY.get(session_pdf_id, []) if m["role"]=="assistant"])
+    summary_text = "\n".join([msg["content"] for msg in SESSION_HISTORY.get(session_pdf_id, []) if msg["role"]=="assistant"])
     if not summary_text:
         summary_text = "No summary available."
     return generate_pdf_file(summary_text, "pdf_summary")
 
-
 def download_image_summary(session_image_id):
-    summary_text = "\n".join([m["content"] for m in SESSION_HISTORY.get(session_image_id, []) if m["role"]=="assistant"])
+    summary_text = "\n".join([msg["content"] for msg in SESSION_HISTORY.get(session_image_id, []) if msg["role"]=="assistant"])
     if not summary_text:
         summary_text = "No summary available."
     return generate_pdf_file(summary_text, "image_summary")
 
+# ------------------ Voice Handlers ------------------
+def handle_voice_general(audio_file, session_id, tts_lang="en"):
+    path = _get_path_from_gr_file(audio_file)
+    if not path:
+        return "No audio provided.", None, []
+    user_text = transcribe_audio(path)
+    assistant_text = generate_response(session_id, user_text)
+    _append_chat_display(session_id, user_text, assistant_text)
+    audio_path = synthesize_speech(assistant_text, lang=tts_lang)
+    return assistant_text, audio_path, _chat_display_to_messages(CHAT_DISPLAY[session_id])
+
+def handle_voice_pdf(audio_file, session_id, tts_lang="en"):
+    path = _get_path_from_gr_file(audio_file)
+    if not path:
+        return "No audio provided.", None, []
+    user_text = transcribe_audio(path)
+    assistant_text = handle_text_pdf(user_text, session_id)
+    _append_chat_display(session_id, user_text, assistant_text)
+    audio_path = synthesize_speech(assistant_text, lang=tts_lang)
+    return assistant_text, audio_path, _chat_display_to_messages(CHAT_DISPLAY[session_id])
+
+def handle_voice_image(audio_file, session_id, tts_lang="en"):
+    path = _get_path_from_gr_file(audio_file)
+    if not path:
+        return "No audio provided.", None, []
+    user_text = transcribe_audio(path)
+    assistant_text = handle_text_image(user_text, session_id)
+    _append_chat_display(session_id, user_text, assistant_text)
+    audio_path = synthesize_speech(assistant_text, lang=tts_lang)
+    return assistant_text, audio_path, _chat_display_to_messages(CHAT_DISPLAY[session_id])
+
+# ------------------ Text Handlers ------------------
+def handle_text_general(user_text, session_id):
+    assistant = generate_response(session_id, user_text)
+    _append_chat_display(session_id, user_text, assistant)
+    return assistant, _chat_display_to_messages(CHAT_DISPLAY[session_id])
 
 # ------------------ Gradio UI ------------------
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
@@ -207,51 +274,78 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     session_pdf = gr.State(str(uuid.uuid4()))
     session_image = gr.State(str(uuid.uuid4()))
 
-    # --- Voice ---
+    # ---------------- Voice Tab ----------------
     with gr.Tab("🎤 Voice Chat"):
-        chat_voice = gr.Chatbot(type="messages", height=380)
+        chat_voice = gr.Chatbot(type="messages", height=350)
         with gr.Row():
-            mic = gr.Audio(label="Hold & speak", type="filepath")
-            tts_lang = gr.Dropdown(choices=["en", "ur"], value="en", label="TTS Language")
+            mic = gr.Audio(type="filepath", label="🎤 Record Voice (hold & speak)", show_download_button=False)
+            tts_lang = gr.Dropdown(choices=["en", "ur"], value="en", label="TTS Language", interactive=True, scale=1)
         with gr.Row():
-            btn_general = gr.Button("Ask General")
-            btn_pdf = gr.Button("Ask PDF")
-            btn_image = gr.Button("Ask Image")
-        audio_output = gr.Audio(label="Assistant Voice Output", type="filepath")
+            btn_general = gr.Button("Ask General 🎯", scale=1)
+            btn_pdf = gr.Button("Ask PDF 📄", scale=1)
+            btn_image = gr.Button("Ask Image 🖼", scale=1)
+        answer_voice = gr.Textbox(label="Assistant Answer (text)", lines=4)
+        audio_output = gr.Audio(label="Assistant Voice Output", type="filepath", interactive=False)
 
-    # --- PDF ---
+        # Text-only general chat
+        with gr.Row():
+            text_input = gr.Textbox(label="Or type a question (General)", placeholder="Type message here...", lines=2)
+            btn_send_text = gr.Button("Send (Text General)", scale=1)
+
+        btn_general.click(
+            fn=handle_voice_general,
+            inputs=[mic, session_voice, tts_lang],
+            outputs=[answer_voice, audio_output, chat_voice],
+        )
+        btn_pdf.click(
+            fn=handle_voice_pdf,
+            inputs=[mic, session_pdf, tts_lang],
+            outputs=[answer_voice, audio_output, chat_voice],
+        )
+        btn_image.click(
+            fn=handle_voice_image,
+            inputs=[mic, session_image, tts_lang],
+            outputs=[answer_voice, audio_output, chat_voice],
+        )
+        btn_send_text.click(
+            fn=handle_text_general,
+            inputs=[text_input, session_voice],
+            outputs=[answer_voice, chat_voice],
+        )
+
+    # ---------------- PDF Tab ----------------
     with gr.Tab("📄 PDF Summarizer"):
-        pdf_output = gr.Textbox(label="Answer (Text Only)", lines=8)
-        pdf_summary_file = gr.File(label="Download Summary PDF")
+        pdf_output = gr.Textbox(label="Answer (Text Only)", lines=6)
+        pdf_summary_file = gr.File(label="📥 Download PDF Summary")
         with gr.Row():
             pdf_upload_btn = gr.File(label="Upload PDF", file_types=[".pdf"], interactive=True)
-            pdf_upload_msg = gr.Textbox(label="Upload Status", interactive=False)
+            pdf_upload_msg = gr.Textbox(label="Upload Status", interactive=False, lines=1)
         pdf_question = gr.Textbox(label="Ask a question about PDF (text)", lines=2)
-        pdf_send_btn = gr.Button("Ask (Text)")
-        pdf_reset_btn = gr.Button("♻ Reset PDF")
-        pdf_download_btn = gr.Button("📥 Download Summary")
+        pdf_send_btn = gr.Button("Ask (Text)", scale=1)
+        pdf_reset_btn = gr.Button("♻ Reset PDF", scale=1)
+        pdf_download_btn = gr.Button("📥 Download Summary", scale=1)
 
         pdf_upload_btn.upload(handle_pdf_upload, inputs=[pdf_upload_btn, session_pdf], outputs=[pdf_upload_msg])
         pdf_send_btn.click(handle_text_pdf, inputs=[pdf_question, session_pdf], outputs=[pdf_output])
         pdf_reset_btn.click(lambda: (str(uuid.uuid4()), ""), outputs=[session_pdf, pdf_output])
         pdf_download_btn.click(download_pdf_summary, inputs=[session_pdf], outputs=[pdf_summary_file])
 
-    # --- Image ---
+    # ---------------- Image Tab ----------------
     with gr.Tab("🖼 Image OCR"):
-        image_output = gr.Textbox(label="Answer (Text Only)", lines=8)
-        img_summary_file = gr.File(label="Download Summary PDF")
+        image_output = gr.Textbox(label="Answer (Text Only)", lines=6)
+        img_summary_file = gr.File(label="📥 Download PDF Summary")
         with gr.Row():
             image_upload_btn = gr.File(label="Upload Image", file_types=[".png", ".jpg", ".jpeg"], interactive=True)
-            image_upload_msg = gr.Textbox(label="Upload Status", interactive=False)
+            image_upload_msg = gr.Textbox(label="Upload Status", interactive=False, lines=1)
         image_question = gr.Textbox(label="Ask a question about Image (text)", lines=2)
-        image_send_btn = gr.Button("Ask (Text)")
-        image_reset_btn = gr.Button("♻ Reset Image")
-        img_download_btn = gr.Button("📥 Download Summary")
+        image_send_btn = gr.Button("Ask", scale=1)
+        image_reset_btn = gr.Button("♻ Reset Image", scale=1)
+        img_download_btn = gr.Button("📥 Download Summary", scale=1)
 
-        image_upload_btn.upload(handle_image_upload, inputs=[image_upload_btn, session_image], outputs=[image_upload_msg])
+        image_upload_btn.upload(handle_image_upload, inputs=[image_upload_btn, session_image], outputs=[image_upload_msg, image_output])
         image_send_btn.click(handle_text_image, inputs=[image_question, session_image], outputs=[image_output])
         image_reset_btn.click(lambda: (str(uuid.uuid4()), ""), outputs=[session_image, image_output])
         img_download_btn.click(download_image_summary, inputs=[session_image], outputs=[img_summary_file])
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=7860, ssr_mode=False)
+    demo.launch(server_name="0.0.0.0", server_port=7860)