Create app.py

app.py

@@ -0,0 +1,409 @@
+import gradio as gr
+import requests
+import os
+import base64
+import io
+import json
+from PIL import Image
+import numpy as np
+
+HF_TOKEN = os.environ.get("HF_TOKEN", "")
+
+HEADERS = {"Authorization": f"Bearer {HF_TOKEN}"}
+
+# ── Model IDs ──────────────────────────────────────────────────────────────────
+TEXT_MODEL        = "Qwen/Qwen2.5-0.5B-Instruct"          # tiny chat LLM
+IMAGE_MODEL       = "black-forest-labs/FLUX.1-schnell"     # fast image gen
+AUDIO_MODEL       = "facebook/musicgen-small"              # audio/music gen
+VIDEO_MODEL       = "ali-vilab/text-to-video-ms-1.7b"     # text-to-video
+MULTIMODAL_MODEL  = "Salesforce/blip2-opt-2.7b"           # image+text (VQA / caption)
+
+# ── HF Inference API helpers ───────────────────────────────────────────────────
+
+def query_text(messages: list, system: str = "") -> str:
+    """Chat via HF serverless inference (text generation)."""
+    url = f"https://api-inference.huggingface.co/models/{TEXT_MODEL}/v1/chat/completions"
+    payload = {
+        "model": TEXT_MODEL,
+        "messages": messages,
+        "max_tokens": 1024,
+        "temperature": 0.7,
+    }
+    r = requests.post(url, headers=HEADERS, json=payload, timeout=60)
+    if r.status_code == 200:
+        return r.json()["choices"][0]["message"]["content"]
+    # fallback plain text-generation endpoint
+    url2 = f"https://api-inference.huggingface.co/models/{TEXT_MODEL}"
+    prompt = "\n".join(m["content"] for m in messages)
+    r2 = requests.post(url2, headers=HEADERS, json={"inputs": prompt, "parameters": {"max_new_tokens": 512}}, timeout=60)
+    if r2.status_code == 200:
+        result = r2.json()
+        if isinstance(result, list):
+            return result[0].get("generated_text", str(result))
+        return str(result)
+    return f"⚠️ Error {r.status_code}: {r.text[:300]}"
+
+
+def query_image(prompt: str) -> Image.Image | str:
+    """Generate image via HF inference."""
+    url = f"https://api-inference.huggingface.co/models/{IMAGE_MODEL}"
+    r = requests.post(url, headers=HEADERS, json={"inputs": prompt}, timeout=120)
+    if r.status_code == 200:
+        return Image.open(io.BytesIO(r.content))
+    return f"⚠️ Error {r.status_code}: {r.text[:300]}"
+
+
+def query_audio(prompt: str) -> str | None:
+    """Generate audio via HF inference, returns a temp file path."""
+    url = f"https://api-inference.huggingface.co/models/{AUDIO_MODEL}"
+    r = requests.post(url, headers=HEADERS, json={"inputs": prompt}, timeout=120)
+    if r.status_code == 200:
+        path = "/tmp/generated_audio.wav"
+        with open(path, "wb") as f:
+            f.write(r.content)
+        return path
+    return None
+
+
+def query_video(prompt: str) -> str | None:
+    """Generate short video via HF inference, returns a temp file path."""
+    url = f"https://api-inference.huggingface.co/models/{VIDEO_MODEL}"
+    r = requests.post(url, headers=HEADERS, json={"inputs": prompt}, timeout=180)
+    if r.status_code == 200:
+        path = "/tmp/generated_video.mp4"
+        with open(path, "wb") as f:
+            f.write(r.content)
+        return path
+    return None
+
+
+def query_multimodal(image: Image.Image | None, text: str) -> tuple[str, Image.Image | None]:
+    """VQA / image captioning with BLIP-2. Also returns the original image."""
+    if image is None:
+        # No image → just caption with a placeholder or echo
+        return "Please upload an image for multimodal analysis.", None
+    # Encode image to base64
+    buf = io.BytesIO()
+    image.save(buf, format="PNG")
+    b64 = base64.b64encode(buf.getvalue()).decode()
+    url = f"https://api-inference.huggingface.co/models/{MULTIMODAL_MODEL}"
+    payload = {"inputs": {"image": b64, "question": text or "Describe this image in detail."}}
+    r = requests.post(url, headers=HEADERS, json=payload, timeout=90)
+    if r.status_code == 200:
+        result = r.json()
+        if isinstance(result, list):
+            answer = result[0].get("answer", str(result[0]))
+        elif isinstance(result, dict):
+            answer = result.get("answer", str(result))
+        else:
+            answer = str(result)
+        return answer, image
+    return f"⚠️ Error {r.status_code}: {r.text[:300]}", image
+
+
+# ── Chat state helper ──────────────────────────────────────────────────────────
+
+def chat_respond(user_msg: str, history: list) -> tuple[str, list]:
+    if not user_msg.strip():
+        return "", history
+    messages = [{"role": "system", "content": "You are NEXUS, an advanced AI assistant. Be helpful, thorough, and thoughtful."}]
+    for human, bot in history:
+        messages.append({"role": "user", "content": human})
+        messages.append({"role": "assistant", "content": bot})
+    messages.append({"role": "user", "content": user_msg})
+    reply = query_text(messages)
+    history.append((user_msg, reply))
+    return "", history
+
+
+def gen_image(prompt: str, progress=gr.Progress()):
+    progress(0.2, desc="Connecting to FLUX…")
+    result = query_image(prompt)
+    progress(1.0, desc="Done")
+    if isinstance(result, str):   # error string
+        return None, result
+    return result, "✅ Image generated successfully!"
+
+
+def gen_audio(prompt: str, progress=gr.Progress()):
+    progress(0.2, desc="Composing music…")
+    path = query_audio(prompt)
+    progress(1.0, desc="Done")
+    if path:
+        return path, "✅ Audio generated!"
+    return None, "⚠️ Audio generation failed. The model may be loading – try again in a moment."
+
+
+def gen_video(prompt: str, progress=gr.Progress()):
+    progress(0.2, desc="Rendering frames…")
+    path = query_video(prompt)
+    progress(1.0, desc="Done")
+    if path:
+        return path, "✅ Video generated!"
+    return None, "⚠️ Video generation failed. The model may be loading – try again in a moment."
+
+
+def gen_multimodal(image, question: str, progress=gr.Progress()):
+    progress(0.3, desc="Analyzing image…")
+    answer, img_out = query_multimodal(image, question)
+    progress(1.0, desc="Done")
+    return answer, img_out
+
+
+# ── CSS ────────────────────────────────────────────────────────────────────────
+CSS = """
+@import url('https://fonts.googleapis.com/css2?family=Space+Mono:wght@400;700&family=DM+Sans:ital,wght@0,300;0,500;0,700;1,300&display=swap');
+
+:root {
+    --bg:        #0a0a0f;
+    --panel:     #111118;
+    --border:    #1e1e2e;
+    --accent:    #7c3aed;
+    --accent2:   #06b6d4;
+    --accent3:   #f59e0b;
+    --text:      #e2e8f0;
+    --muted:     #64748b;
+    --success:   #10b981;
+    --danger:    #ef4444;
+}
+
+body, .gradio-container { background: var(--bg) !important; font-family: 'DM Sans', sans-serif; color: var(--text); }
+
+/* Header */
+.nexus-header {
+    text-align: center;
+    padding: 2.5rem 1rem 1rem;
+    background: linear-gradient(135deg, #0a0a0f 0%, #130d22 50%, #0a0a0f 100%);
+    border-bottom: 1px solid var(--border);
+    margin-bottom: 1.5rem;
+}
+.nexus-title {
+    font-family: 'Space Mono', monospace;
+    font-size: clamp(2rem, 6vw, 3.5rem);
+    font-weight: 700;
+    background: linear-gradient(90deg, var(--accent) 0%, var(--accent2) 50%, var(--accent3) 100%);
+    -webkit-background-clip: text;
+    -webkit-text-fill-color: transparent;
+    letter-spacing: -0.02em;
+    margin: 0;
+}
+.nexus-sub {
+    color: var(--muted);
+    font-size: 0.95rem;
+    margin-top: 0.5rem;
+    letter-spacing: 0.08em;
+    text-transform: uppercase;
+}
+.badge-row { display: flex; justify-content: center; gap: 0.5rem; flex-wrap: wrap; margin-top: 1rem; }
+.badge {
+    font-family: 'Space Mono', monospace;
+    font-size: 0.65rem;
+    padding: 0.25rem 0.75rem;
+    border-radius: 999px;
+    border: 1px solid;
+    letter-spacing: 0.05em;
+}
+.badge-chat   { border-color: var(--accent);  color: var(--accent);  }
+.badge-img    { border-color: var(--accent2); color: var(--accent2); }
+.badge-audio  { border-color: var(--accent3); color: var(--accent3); }
+.badge-video  { border-color: var(--success); color: var(--success); }
+.badge-mm     { border-color: var(--danger);  color: var(--danger);  }
+
+/* Tabs */
+.tab-nav { border-bottom: 1px solid var(--border) !important; }
+.tab-nav button {
+    font-family: 'Space Mono', monospace !important;
+    font-size: 0.8rem !important;
+    letter-spacing: 0.05em !important;
+    color: var(--muted) !important;
+    padding: 0.75rem 1.25rem !important;
+    border-bottom: 2px solid transparent !important;
+    transition: all 0.2s !important;
+}
+.tab-nav button.selected {
+    color: var(--accent2) !important;
+    border-bottom-color: var(--accent2) !important;
+}
+
+/* Chatbot */
+.chatbot .message.user { background: rgba(124,58,237,0.15) !important; border-left: 3px solid var(--accent) !important; }
+.chatbot .message.bot  { background: rgba(6,182,212,0.08)  !important; border-left: 3px solid var(--accent2) !important; }
+
+/* Inputs */
+textarea, input[type=text] {
+    background: var(--panel) !important;
+    border: 1px solid var(--border) !important;
+    color: var(--text) !important;
+    border-radius: 8px !important;
+    font-family: 'DM Sans', sans-serif !important;
+}
+textarea:focus, input:focus {
+    border-color: var(--accent2) !important;
+    box-shadow: 0 0 0 2px rgba(6,182,212,0.15) !important;
+}
+
+/* Buttons */
+.btn-primary {
+    background: linear-gradient(135deg, var(--accent) 0%, var(--accent2) 100%) !important;
+    color: #fff !important;
+    font-family: 'Space Mono', monospace !important;
+    font-size: 0.8rem !important;
+    letter-spacing: 0.05em !important;
+    border: none !important;
+    border-radius: 8px !important;
+    padding: 0.65rem 1.5rem !important;
+    cursor: pointer !important;
+    transition: opacity 0.2s !important;
+}
+.btn-primary:hover { opacity: 0.85 !important; }
+
+/* Status box */
+.status-box {
+    background: var(--panel);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    padding: 0.75rem 1rem;
+    font-family: 'Space Mono', monospace;
+    font-size: 0.75rem;
+    color: var(--muted);
+    min-height: 2.5rem;
+}
+
+/* Section labels */
+.section-label {
+    font-family: 'Space Mono', monospace;
+    font-size: 0.7rem;
+    letter-spacing: 0.1em;
+    color: var(--muted);
+    text-transform: uppercase;
+    margin-bottom: 0.4rem;
+}
+"""
+
+# ── Build UI ───────────────────────────────────────────────────────────────────
+
+with gr.Blocks(css=CSS, title="NEXUS AI Studio", theme=gr.themes.Base()) as demo:
+
+    # Header
+    gr.HTML("""
+    <div class="nexus-header">
+        <h1 class="nexus-title">⬡ NEXUS AI STUDIO</h1>
+        <p class="nexus-sub">Multi-Modal Intelligence Platform · Powered by Hugging Face</p>
+        <div class="badge-row">
+            <span class="badge badge-chat">💬 CHAT</span>
+            <span class="badge badge-img">🖼 IMAGE GEN</span>
+            <span class="badge badge-audio">🎵 AUDIO GEN</span>
+            <span class="badge badge-video">🎬 VIDEO GEN</span>
+            <span class="badge badge-mm">🔮 MULTIMODAL</span>
+        </div>
+    </div>
+    """)
+
+    with gr.Tabs(elem_classes="tab-nav"):
+
+        # ── TAB 1: Chat ────────────────────────────────────────────────────────
+        with gr.Tab("💬  Chat"):
+            gr.HTML('<p class="section-label">Conversational AI · Qwen 2.5</p>')
+            chatbot = gr.Chatbot(
+                label="",
+                height=460,
+                bubble_full_width=False,
+                elem_classes="chatbot",
+                show_label=False,
+                avatar_images=(None, "https://huggingface.co/front/assets/huggingface_logo-noborder.svg"),
+            )
+            with gr.Row():
+                chat_in = gr.Textbox(
+                    placeholder="Ask me anything — code, math, science, creative writing…",
+                    show_label=False,
+                    scale=8,
+                    lines=1,
+                )
+                send_btn = gr.Button("SEND →", elem_classes="btn-primary", scale=1)
+            clear_btn = gr.Button("Clear conversation", variant="secondary", size="sm")
+
+            send_btn.click(chat_respond, [chat_in, chatbot], [chat_in, chatbot])
+            chat_in.submit(chat_respond, [chat_in, chatbot], [chat_in, chatbot])
+            clear_btn.click(lambda: ([], ""), None, [chatbot, chat_in])
+
+        # ── TAB 2: Image Generation ────────────────────────────────────────────
+        with gr.Tab("🖼  Image Gen"):
+            gr.HTML('<p class="section-label">Text → Image · FLUX.1-schnell</p>')
+            with gr.Row():
+                with gr.Column(scale=1):
+                    img_prompt = gr.Textbox(
+                        label="Prompt",
+                        placeholder="A neon-lit cyberpunk city at midnight, rain reflections, ultra-detailed…",
+                        lines=4,
+                    )
+                    img_btn = gr.Button("✦ GENERATE IMAGE", elem_classes="btn-primary")
+                    img_status = gr.Textbox(label="Status", elem_classes="status-box", show_label=False)
+                with gr.Column(scale=1):
+                    img_out = gr.Image(label="Generated Image", show_label=False)
+
+            img_btn.click(gen_image, [img_prompt], [img_out, img_status])
+
+        # ── TAB 3: Audio Generation ────────────────────────────────────────────
+        with gr.Tab("🎵  Audio Gen"):
+            gr.HTML('<p class="section-label">Text → Music · MusicGen Small</p>')
+            with gr.Row():
+                with gr.Column(scale=1):
+                    audio_prompt = gr.Textbox(
+                        label="Describe the music",
+                        placeholder="Lo-fi hip hop beat, warm piano chords, gentle rain ambiance, 80 BPM…",
+                        lines=4,
+                    )
+                    audio_btn = gr.Button("♪ GENERATE AUDIO", elem_classes="btn-primary")
+                    audio_status = gr.Textbox(label="Status", elem_classes="status-box", show_label=False)
+                with gr.Column(scale=1):
+                    audio_out = gr.Audio(label="Generated Audio", show_label=True)
+
+            audio_btn.click(gen_audio, [audio_prompt], [audio_out, audio_status])
+
+        # ── TAB 4: Video Generation ────────────────────────────────────────────
+        with gr.Tab("🎬  Video Gen"):
+            gr.HTML('<p class="section-label">Text → Video · ModelScope 1.7B</p>')
+            with gr.Row():
+                with gr.Column(scale=1):
+                    video_prompt = gr.Textbox(
+                        label="Describe the video",
+                        placeholder="A lone astronaut walking on Mars at sunset, dust swirling around boots…",
+                        lines=4,
+                    )
+                    video_btn = gr.Button("▶ GENERATE VIDEO", elem_classes="btn-primary")
+                    video_status = gr.Textbox(label="Status", elem_classes="status-box", show_label=False)
+                with gr.Column(scale=1):
+                    video_out = gr.Video(label="Generated Video", show_label=True)
+
+            video_btn.click(gen_video, [video_prompt], [video_out, video_status])
+
+        # ── TAB 5: Multimodal ─────────────────────────────────────────────────
+        with gr.Tab("🔮  Multimodal"):
+            gr.HTML('<p class="section-label">Image + Text → Answer · BLIP-2</p>')
+            with gr.Row():
+                with gr.Column(scale=1):
+                    mm_image = gr.Image(
+                        label="Upload an Image",
+                        type="pil",
+                        show_label=True,
+                    )
+                    mm_question = gr.Textbox(
+                        label="Your question about the image",
+                        placeholder="What is happening in this image? What objects do you see?",
+                        lines=3,
+                    )
+                    mm_btn = gr.Button("🔮 ANALYZE", elem_classes="btn-primary")
+                with gr.Column(scale=1):
+                    mm_answer = gr.Textbox(label="AI Answer", lines=6, show_label=True)
+                    mm_img_out = gr.Image(label="Processed Image", show_label=True)
+
+            mm_btn.click(gen_multimodal, [mm_image, mm_question], [mm_answer, mm_img_out])
+
+    # Footer
+    gr.HTML("""
+    <div style="text-align:center; padding: 1.5rem; border-top: 1px solid #1e1e2e; margin-top: 1.5rem; color: #475569; font-size: 0.75rem; font-family: 'Space Mono', monospace; letter-spacing: 0.05em;">
+        NEXUS AI STUDIO · Built with ❤ on Hugging Face Spaces · Models: Qwen2.5 · FLUX.1-schnell · MusicGen · ModelScope · BLIP-2
+    </div>
+    """)
+
+demo.launch()