Update app.py

app.py

@@ -1,98 +1,196 @@
-# app.py
 import os
+import base64
+import re
+import requests
+import pytz
+import json
+from io import BytesIO
+from datetime import datetime
 import gradio as gr
 import openai
+import fitz  # pymupdf
+from bs4 import BeautifulSoup
+from moviepy.video.io.VideoFileClip import VideoFileClip
 
-# ─── CONFIG ────────────────────────────────────────────────────────────────────
+# 🔐 CONFIG
 KEY_FILE = "openai_api_key.txt"
-MODEL    = "gpt-4o-2024-05-13"
+MODEL = "gpt-4o-2024-05-13"
 
-# Try to load a cached key (global for all users)
+# 🍿 Default key load
 if os.path.exists(KEY_FILE):
-    with open(KEY_FILE, "r") as f:
+    with open(KEY_FILE, 'r') as f:
         DEFAULT_KEY = f.read().strip()
 else:
-    DEFAULT_KEY = ""
+    DEFAULT_KEY = ''
 
-# ─── HELPERS ───────────────────────────────────────────────────────────────────
-
-def save_api_key(api_key: str) -> str:
-    """Save the key to disk so it's remembered for all future sessions."""
-    with open(KEY_FILE, "w") as f:
+# 🔧 HELPERS
+def save_api_key(api_key):
+    with open(KEY_FILE, 'w') as f:
         f.write(api_key.strip())
-    return "✅ API key saved and will be used for all users."
+    return "🔑 Key saved!"
+
+# 🗒️ Chat
 
-def chat_with_openai(api_key: str, user_message: str, history: list) -> list:
-    """Invoke GPT-4o with the supplied API key and append to chat history."""
+def chat_with_openai(api_key, user_message, history):
     openai.api_key = api_key.strip()
-    # ensure history is a list of tuples [(user, bot), ...]
-    history = history or []
-    # Build messages in OpenAI format
     messages = []
     for u, a in history:
-        messages.append({"role":"user",    "content": u})
-        messages.append({"role":"assistant","content": a})
-    messages.append({"role":"user", "content": user_message})
-    # Call the API
-    resp = openai.ChatCompletion.create(
-        model=MODEL,
-        messages=messages,
-    )
+        messages.append({"role": "user", "content": u})
+        messages.append({"role": "assistant", "content": a})
+    messages.append({"role": "user", "content": user_message})
+    resp = openai.ChatCompletion.create(model=MODEL, messages=messages)
     answer = resp.choices[0].message.content
     history.append((user_message, answer))
     return history
 
-# ─── UI LAYOUT ────────────────────────────────────────────────────────────────
-
-with gr.Blocks(title="🌐 GPT-4o Multimodal (Skeleton)") as demo:
+# 🖼️ Image analysis
+def image_to_base64(file):
+    return base64.b64encode(file.read()).decode()
 
-    gr.Markdown(
-        """
-        # 🤖 GPT-4o Client  
-        Enter your OpenAI API key once below.  
-        It will be cached on the server for all future sessions.
-        """
+def analyze_image(api_key, file, prompt):
+    data_uri = f"data:image/png;base64,{image_to_base64(file)}"
+    openai.api_key = api_key.strip()
+    resp = openai.ChatCompletion.create(
+        model=MODEL,
+        messages=[
+            {"role": "system", "content": "You are a helpful assistant that responds in Markdown."},
+            {"role": "user", "content": [
+                    {"type":"text","text":prompt},
+                    {"type":"image_url","image_url":{"url":data_uri}}
+            ]}
+        ]
     )
+    return resp.choices[0].message.content
 
-    with gr.Row():
-        api_key_input = gr.Textbox(
-            label="🔑 OpenAI API Key",
-            placeholder="sk-…",
-            value=DEFAULT_KEY,
-            type="password",
-        )
-        save_button = gr.Button("💾 Save API Key")
-        save_status = gr.Textbox(
-            label="Status",
-            interactive=False,
-            placeholder="–"
-        )
-
-    # Wire the save button
-    save_button.click(
-        fn=save_api_key,
-        inputs=api_key_input,
-        outputs=save_status,
-    )
+# 🎤 Audio transcription + chat
+def transcribe_audio_file(api_key, file):
+    openai.api_key = api_key.strip()
+    resp = openai.Audio.transcriptions.create(model="whisper-1", file=file)
+    return resp.text
 
-    gr.Markdown("---
+# 🎥 Video summarize
+def summarize_video(api_key, file, prompt, seconds=2):
+    # save tmp
+    with open("tmp_vid.mp4", 'wb') as f: f.write(file.read())
+    clip = VideoFileClip("tmp_vid.mp4")
+    frames = []
+    step = int(clip.fps * seconds)
+    for t in range(0, int(clip.duration), seconds):
+        frame = clip.get_frame(t)
+        buf = BytesIO()
+        from PIL import Image
+        Image.fromarray(frame).save(buf, format='JPEG')
+        frames.append(base64.b64encode(buf.getvalue()).decode())
+    transcript = transcribe_audio_file(api_key, open("tmp_vid.mp4", 'rb'))
+    openai.api_key = api_key.strip()
+    messages = [{"role":"system","content":"You are a helpful assistant."},
+                {"role":"user","content": prompt}]
+    for f64 in frames:
+        messages.append({"role":"user","content": {"type":"image_url","image_url":{"url":f"data:image/jpeg;base64,{f64}"}}})
+    messages.append({"role":"user","content": f"Transcript: {transcript}"})
+    resp = openai.ChatCompletion.create(model=MODEL, messages=messages)
+    return resp.choices[0].message.content
 
-## 💬 Chat with GPT-4o")
-    chatbot = gr.Chatbot(label="Chat History")
-    msg = gr.Textbox(
-        label="Your message",
-        placeholder="Type something and press Enter…",
-    )
+# 📄 PDF->Markdown
+def pdf_to_markdown(path):
+    doc = fitz.open(path)
+    md = ''
+    for page in doc:
+        md += page.get_text('markdown') + '\n'
+    return md
+
+# 🔍 ArXiv RAG
+from gradio_client import Client
+def arxiv_search(query):
+    client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
+    refs = client.predict(query, 10, "Semantic Search - up to 10 Mar 2024", "mistralai/Mixtral-8x7B-Instruct-v0.1", api_name="/update_with_rag_md")
+    ans = client.predict(query, "mistralai/Mixtral-8x7B-Instruct-v0.1", True, api_name="/ask_llm")
+    return refs + "\n" + ans
+
+# 🔈 TTS
+from gtts import gTTS
+def tts_bytes(text):
+    buf = BytesIO()
+    gTTS(text=text, lang='en').write_to_fp(buf)
+    buf.seek(0)
+    return buf.read()
 
-    # When you hit Enter in the textbox, call the chat fn
-    msg.submit(
-        fn=chat_with_openai,
-        inputs=[api_key_input, msg, chatbot],
-        outputs=chatbot,
+# UI CONFIG
+ui_config = {
+    "chat": {"label":"💬 ChatGPT-4o", "placeholder":"Say something..."},
+    "image_prompt": {"label":"🖼️ Image Prompt", "default":"Describe this image..."},
+    "audio_prompt": {"label":"🎤 Audio Prompt", "default":"Transcribe and summarize..."},
+    "video_prompt": {"label":"🎥 Video Prompt", "default":"Summarize this video..."},
+    "pdf_prompt": {"label":"📄 PDF Prompt", "default":"Convert PDF to markdown..."},
+    "arxiv_prompt": {"label":"🔍 Arxiv Query", "default":"Search papers..."}
+}
+
+with gr.Blocks(title="🔬🧠 ScienceBrain.Gradio") as demo:
+    gr.Markdown("# 🔬🧠 ScienceBrain Gradio
+Enter API key below."
     )
+    with gr.Row():
+        api_key = gr.Textbox(label="🔑 OpenAI Key", value=DEFAULT_KEY, type="password")
+        save_btn = gr.Button("💾 Save Key")
+        status = gr.Textbox(label="Status", interactive=False)
+    save_btn.click(save_api_key, inputs=api_key, outputs=status)
+
+    # Tabs for each modality
+    with gr.Tab("💬 Chat"):
+        chatbot = gr.Chatbot(label=ui_config['chat']['label'], value=[])
+        msg = gr.Textbox(label=ui_config['chat']['label'], placeholder=ui_config['chat']['placeholder'])
+        msg.submit(chat_with_openai, inputs=[api_key, msg, chatbot], outputs=chatbot)
+
+    with gr.Tab("🖼️ Image"):
+        img_in = gr.File(file_types=['png','jpg','jpeg'])
+        img_prompt = gr.Textbox(label=ui_config['image_prompt']['label'], value=ui_config['image_prompt']['default'])
+        img_btn = gr.Button("🔍 Analyze Image")
+        img_out = gr.Markdown()
+        img_btn.click(analyze_image, inputs=[api_key, img_in, img_prompt], outputs=img_out)
+
+    with gr.Tab("🎤 Audio"):
+        aud_in = gr.File(file_types=['wav','mp3'])
+        aud_btn = gr.Button("🔊 Transcribe + Chat")
+        aud_out = gr.Markdown()
+        def audio_pipeline(key, f):
+            text = transcribe_audio_file(key, f)
+            reply = chat_with_openai(key, text, [])[-1][1]
+            return f"**Transcript:** {text}\n\n**Reply:** {reply}"
+        aud_btn.click(audio_pipeline, inputs=[api_key, aud_in], outputs=aud_out)
+
+    with gr.Tab("🎥 Video"):
+        vid_in = gr.File(file_types=['mp4'])
+        vid_prompt = gr.Textbox(label=ui_config['video_prompt']['label'], value=ui_config['video_prompt']['default'])
+        vid_btn = gr.Button("🎞️ Summarize Video")
+        vid_out = gr.Markdown()
+        vid_btn.click(summarize_video, inputs=[api_key, vid_in, vid_prompt], outputs=vid_out)
+
+    with gr.Tab("📄 PDF"):
+        pdf_in = gr.File(file_types=['pdf'])
+        pdf_btn = gr.Button("📝 Convert PDF")
+        pdf_out = gr.Markdown()
+        pdf_btn.click(lambda f: pdf_to_markdown(f.name), inputs=[pdf_in], outputs=pdf_out)
+
+    with gr.Tab("🔍 ArXiv"):
+        arxiv_in = gr.Textbox(label=ui_config['arxiv_prompt']['label'], value=ui_config['arxiv_prompt']['default'])
+        arxiv_btn = gr.Button("🔎 Search ArXiv")
+        arxiv_out = gr.Markdown()
+        arxiv_btn.click(arxiv_search, inputs=[arxiv_in], outputs=arxiv_out)
+
+    with gr.Tab("⚙️ Quick Tests"):
+        tests = [
+            ("📝 Text","What is 2+2?"),
+            ("🖼️ Image","Analyze image https://via.placeholder.com/150.png"),
+            ("🎤 Audio","Transcribe https://www2.cs.uic.edu/~i101/SoundFiles/gettysburg10.wav"),
+            ("🎥 Video","Summarize video https://sample-videos.com/video123/mp4/240/big_buck_bunny_240p_1mb.mp4"),
+            ("🖼️+📝 Img+Txt","Given image https://via.placeholder.com/150.png list 3 facts."),
+            ("🎤+📝 Aud+Txt","Given audio https://www2.cs.uic.edu/~i101/SoundFiles/gettysburg10.wav summarize."),
+            ("🎥+📝 Vid+Txt","Given video https://sample-videos.com/video123/mp4/240/big_buck_bunny_240p_1mb.mp4 transcript+summary.")
+        ]
+        for idx, (e,p) in enumerate(tests,1):
+            btn = gr.Button(f"{idx}. {e} Test")
+            btn.click(chat_with_openai, inputs=[api_key, gr.State(p), gr.State([])], outputs=chatbot)
 
-# ─── RUN ───────────────────────────────────────────────────────────────────────
 
 if __name__ == "__main__":
-    # Use 0.0.0.0 if you want external access; port can be adjusted as needed
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+    demo.launch()