Spaces:

oldmonk69
/

gemini-ui-redesign

Running

File size: 21,305 Bytes

<!DOCTYPE html>
<html class="dark" lang="en">
<head>
  <meta charset="utf-8" />
  <meta content="width=device-width, initial-scale=1.0" name="viewport" />
  <title>NeuralNomadAI - HuggingFace Demo (index.html)</title>

  <!-- Google Fonts -->
  <link href="https://fonts.googleapis.com" rel="preconnect" />
  <link crossorigin href="https://fonts.gstatic.com" rel="preconnect" />
  <link href="https://fonts.googleapis.com/css2?family=Space+Grotesk:wght@400;500;700&display=swap" rel="stylesheet" />

  <!-- Tailwind CDN (keeps the original layout & utilities) -->
  <script src="https://cdn.tailwindcss.com?plugins=forms,container-queries"></script>

  <!-- Local CSS for small customizations -->
  <link rel="stylesheet" href="style.css" />
</head>

<body class="bg-gray-50 dark:bg-gray-900 font-display text-gray-900 dark:text-gray-100">
  <div class="flex min-h-screen w-full flex-col">
    <!-- Header -->
    <header class="sticky top-0 z-50 border-b border-gray-200/50 dark:border-gray-700/50 bg-white/80 dark:bg-gray-900/80 backdrop-blur-sm">
      <div class="mx-auto flex max-w-7xl items-center justify-between px-4 py-3">
        <a href="#home" data-page="home" class="flex items-center gap-4 cursor-pointer">
          <svg class="h-8 w-8 text-sky-600" fill="none" viewBox="0 0 48 48" xmlns="http://www.w3.org/2000/svg"><path d="M39.5563 34.1455V13.8546C39.5563 15.708 36.8773 17.3437 32.7927 18.3189C30.2914 18.916 27.263 19.2655 24 19.2655C20.737 19.2655 17.7086 18.916 15.2073 18.3189C11.1227 17.3437 8.44365 15.708 8.44365 13.8546V34.1455C8.44365 35.9988 11.1227 37.6346 15.2073 38.6098C17.7086 39.2069 20.737 39.5564 24 39.5564C27.263 39.5564 30.2914 39.2069 32.7927 38.6098C36.8773 37.6346 39.5563 35.9988 39.5563 34.1455Z" fill="currentColor"></path></svg>
          <h1 class="text-lg font-bold">NeuralNomadAI</h1>
        </a>

        <nav class="hidden md:flex items-center gap-6">
          <a href="#home" data-page="home" class="nav-link text-sm font-medium">Home</a>
          <a href="#services" data-page="services" class="nav-link text-sm font-medium">Services</a>
          <a href="#portfolio" data-page="portfolio" class="nav-link text-sm font-medium">Portfolio</a>
          <a href="#contact" data-page="contact" class="nav-link text-sm font-medium">Contact</a>
        </nav>

        <div class="flex items-center gap-4">
          <a href="#contact" data-page="contact" class="rounded-lg bg-sky-600 px-4 py-2 text-sm font-semibold text-white">Get Started</a>
        </div>
      </div>
    </header>

    <main class="flex-grow">
      <!-- HOME (brief) -->
      <section id="home" class="page-content active">
        <div class="mx-auto max-w-7xl px-4 py-12">
          <div class="text-center">
            <h2 class="text-3xl font-bold">AI Prompt & Image Playground — Hugging Face</h2>
            <p class="mt-3 text-gray-600 dark:text-gray-300">This demo uses the Hugging Face Inference API. Replace the client token with a backend proxy token before production.</p>
          </div>

          <!-- Prompt Ideator -->
          <div class="mt-10 max-w-xl mx-auto">
            <h3 class="text-xl font-semibold mb-3">✨ AI Prompt Ideator</h3>
            <div class="flex gap-3">
              <input id="prompt-input" type="text" placeholder="e.g., a cat astronaut on Mars" aria-label="Prompt input"
                class="w-full rounded-lg border border-gray-300 p-3 bg-white dark:bg-gray-800" />
              <button id="generate-prompt-btn" type="button" class="rounded-lg bg-sky-600 px-4 py-2 text-white font-semibold">
                <span class="btn-text">Generate Prompt</span>
                <span class="btn-spinner hidden ml-2">⏳</span>
              </button>
            </div>
            <div id="prompt-result" class="mt-4 result-box hidden" role="status" aria-live="polite"></div>
          </div>

          <!-- Image Playground -->
          <div class="mt-12 max-w-xl mx-auto">
            <h3 class="text-xl font-semibold mb-3">✨ AI Image Playground</h3>
            <div class="flex gap-3">
              <input id="image-prompt-input" type="text" placeholder="e.g., futuristic city at sunset" aria-label="Image prompt input"
                class="w-full rounded-lg border border-gray-300 p-3 bg-white dark:bg-gray-800" />
              <button id="generate-image-btn" type="button" class="rounded-lg bg-sky-600 px-4 py-2 text-white font-semibold">
                <span class="btn-text">Generate Image</span>
                <span class="btn-spinner hidden ml-2">⏳</span>
              </button>
            </div>

            <div id="image-loading" class="mt-6 hidden text-sm text-gray-600 dark:text-gray-300">
              <div class="flex items-center gap-3">
                <svg class="animate-spin h-6 w-6 text-sky-600" xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 24 24">
                  <circle class="opacity-25" cx="12" cy="12" r="10" stroke="currentColor" stroke-width="4"></circle>
                  <path class="opacity-75" fill="currentColor" d="M4 12a8 8 0 018-8V0C5.373 0 0 5.373 0 12h4z"></path>
                </svg>
                <span>Generating image — this may take 10–30s depending on the model.</span>
              </div>
            </div>

            <div id="image-result" class="mt-6 result-box" role="status" aria-live="polite"></div>
          </div>

        </div>
      </section>

      <!-- SERVICES (placeholder) -->
      <section id="services" class="page-content">
        <div class="mx-auto max-w-7xl px-4 py-12">
          <h2 class="text-2xl font-bold">Services</h2>
          <p class="mt-3 text-gray-600 dark:text-gray-300">Sample content...</p>
        </div>
      </section>

      <!-- PORTFOLIO (placeholder) -->
      <section id="portfolio" class="page-content">
        <div class="mx-auto max-w-7xl px-4 py-12">
          <h2 class="text-2xl font-bold">Portfolio</h2>
          <p class="mt-3 text-gray-600 dark:text-gray-300">Sample content...</p>
        </div>
      </section>

      <!-- CONTACT (placeholder) -->
      <section id="contact" class="page-content">
        <div class="mx-auto max-w-7xl px-4 py-12">
          <h2 class="text-2xl font-bold">Contact</h2>
          <p class="mt-3 text-gray-600 dark:text-gray-300">Sample content...</p>
        </div>
      </section>
    </main>

    <footer class="border-t border-gray-200/50 dark:border-gray-700/50 bg-white/80 dark:bg-gray-900/80">
      <div class="mx-auto max-w-7xl px-4 py-8 text-center text-sm text-gray-600 dark:text-gray-300">
        © 2024 NeuralNomadAI — Hugging Face demo. Remember to proxy your API key on the server.
      </div>
    </footer>
  </div>

  <!-- JS: huggingface API integration + UI handling -->
  <script>
    // -------------------------
    // IMPORTANT (security):
    // Do NOT put real HF tokens in client-side code for production.
    // For quick local testing you can paste a short-lived token here,
    // but the recommended pattern is:
    //   - Create a small backend endpoint (e.g. /api/hf/generate)
    //   - The backend attaches your HF token to the request and forwards it to HF
    // -------------------------

    // If you still want to test client-side (ONLY FOR LOCAL TESTING), put a token here:
    // const HF_TOKEN = 'HF_API_TOKEN_HERE';
    // Otherwise set HF_TOKEN to empty and call a backend endpoint instead.
    const HF_TOKEN = ''; // <-- REPLACE WITH '' FOR PRODUCTION. Use backend proxy.

    // Backend proxy settings (recommended)
    // If you implement a backend, set PROXY_BASE to your server endpoint:
    // e.g. const PROXY_BASE = '/api/hf'; and implement endpoints /api/hf/text and /api/hf/image
    const PROXY_BASE = ''; // e.g. '/api/hf'  (leave empty to call HF directly)

    // Model choices (you can change these)
    // Text generation: use a text generation model (make sure to check the model's capabilities)
    const HF_TEXT_MODEL = 'gpt2'; // small example; swap with a better model for real usage
    // Image generation: choose a model that returns image binary, e.g., "stabilityai/stable-diffusion-2"
    const HF_IMAGE_MODEL = 'stabilityai/stable-diffusion-2';

    // Utilities
    const escapeHtml = (unsafe) => {
      if (unsafe === null || unsafe === undefined) return '';
      return String(unsafe)
        .replace(/&/g, '&amp;')
        .replace(/</g, '&lt;')
        .replace(/>/g, '&gt;')
        .replace(/"/g, '&quot;')
        .replace(/'/g, '&#039;');
    };

    const formatSafeHtml = (text) => {
      let out = escapeHtml(text);
      out = out.replace(/\*\*(.+?)\*\*/g, (m, p1) => `<strong>${p1}</strong>`);
      out = out.replace(/\r\n|\r|\n/g, '<br>');
      return out;
    };

    // DOM
    document.addEventListener('DOMContentLoaded', () => {
      const promptInput = document.getElementById('prompt-input');
      const generatePromptBtn = document.getElementById('generate-prompt-btn');
      const promptResult = document.getElementById('prompt-result');

      const imagePromptInput = document.getElementById('image-prompt-input');
      const generateImageBtn = document.getElementById('generate-image-btn');
      const imageLoading = document.getElementById('image-loading');
      const imageResult = document.getElementById('image-result');

      // Small helpers
      const setButtonLoading = (button, loading = true, loadingText = 'Generating...') => {
        if (!button) return;
        const btnTextNode = button.querySelector('.btn-text');
        const spinner = button.querySelector('.btn-spinner');
        button.disabled = loading;
        button.classList.toggle('opacity-60', loading);
        button.setAttribute('aria-busy', String(loading));
        if (btnTextNode && loading) btnTextNode.textContent = loadingText;
        if (btnTextNode && !loading) {
          if (button.id === 'generate-prompt-btn') btnTextNode.textContent = 'Generate Prompt';
          if (button.id === 'generate-image-btn') btnTextNode.textContent = 'Generate Image';
        }
        if (spinner) spinner.classList.toggle('hidden', !loading);
      };

      const showError = (el, message) => {
        if (!el) return;
        el.classList.remove('hidden');
        el.innerHTML = `<div class="p-3 rounded bg-red-50 text-red-700 dark:bg-red-900/30">${escapeHtml(message)}</div>`;
      };

      const showHtml = (el, html) => {
        if (!el) return;
        el.classList.remove('hidden');
        el.innerHTML = html;
      };

      // Generic function to call Hugging Face Inference API for text
      // If PROXY_BASE is set, the code will call PROXY_BASE + '/text' instead.
      // The backend should forward the request to HF with the Authorization header.
      async function hfTextGenerate(prompt) {
        if (!prompt) throw new Error('Empty prompt');
        const payload = { inputs: prompt, options: { wait_for_model: true } };

        if (PROXY_BASE) {
          // call your backend proxy (recommended)
          const resp = await fetch(`${PROXY_BASE}/text`, {
            method: 'POST',
            headers: { 'Content-Type': 'application/json' },
            body: JSON.stringify({ model: HF_TEXT_MODEL, payload }),
          });
          if (!resp.ok) throw new Error(`Proxy error: ${resp.status} ${await resp.text()}`);
          return resp.json();
        }

        // Direct client-side call to HF Inference
        if (!HF_TOKEN) throw new Error('No HF token set (use a backend proxy in production)');

        const url = `https://api-inference.huggingface.co/models/${encodeURIComponent(HF_TEXT_MODEL)}`;
        const resp = await fetch(url, {
          method: 'POST',
          headers: {
            Authorization: `Bearer ${HF_TOKEN}`,
            'Content-Type': 'application/json',
          },
          body: JSON.stringify(payload),
        });

        if (!resp.ok) {
          const txt = await resp.text().catch(() => '');
          throw new Error(`Hugging Face API error: ${resp.status} ${txt}`);
        }

        return resp.json();
      }

      // Generic function to call Hugging Face Inference API for images
      // Many HF image models return binary image data (image/png). We handle binary responses and JSON fallbacks.
      async function hfImageGenerate(prompt) {
        if (!prompt) throw new Error('Empty prompt');

        if (PROXY_BASE) {
          const resp = await fetch(`${PROXY_BASE}/image`, {
            method: 'POST',
            headers: { 'Content-Type': 'application/json' },
            body: JSON.stringify({ model: HF_IMAGE_MODEL, prompt }),
          });
          if (!resp.ok) throw new Error(`Proxy error: ${resp.status} ${await resp.text()}`);
          // Proxy might forward binary image or JSON with base64; try to detect
          const contentType = resp.headers.get('content-type') || '';
          if (contentType.startsWith('image/')) {
            const blob = await resp.blob();
            return { blob };
          } else {
            return resp.json();
          }
        }

        // Direct client-side call to HF Inference (NOT RECOMMENDED)
        if (!HF_TOKEN) throw new Error('No HF token set (use a backend proxy in production)');

        const url = `https://api-inference.huggingface.co/models/${encodeURIComponent(HF_IMAGE_MODEL)}`;
        const resp = await fetch(url, {
          method: 'POST',
          headers: {
            Authorization: `Bearer ${HF_TOKEN}`,
            'Content-Type': 'application/json',
          },
          // For some image models HF expects {"inputs":"<prompt>"} or a specialized payload
          body: JSON.stringify({ inputs: prompt }),
        });

        if (!resp.ok) {
          const txt = await resp.text().catch(() => '');
          throw new Error(`Hugging Face image error: ${resp.status} ${txt}`);
        }

        const contentType = resp.headers.get('content-type') || '';
        if (contentType.startsWith('image/')) {
          const blob = await resp.blob();
          return { blob };
        } else {
          // JSON response (could contain base64 or URLs); parse it
          const json = await resp.json().catch(() => null);
          return json;
        }
      }

      // Handlers
      const handlePromptGeneration = async () => {
        const input = (promptInput.value || '').trim();
        if (!input) {
          showError(promptResult, 'Please enter a short idea to expand into a prompt.');
          return;
        }
        promptResult.classList.add('hidden');
        setButtonLoading(generatePromptBtn, true, 'Generating...');
        try {
          const systemPrompt = `You are a senior prompt engineer. Expand the user's short idea into a detailed, artistic prompt suitable for image generation. The idea: "${input}"`;
          // Use HF text generation
          const resp = await hfTextGenerate(systemPrompt);
          // Response shapes vary: HF text models often return an array like [{generated_text: "..."}]
          let text = null;
          if (Array.isArray(resp) && resp.length > 0 && resp[0].generated_text) {
            text = resp[0].generated_text;
          } else if (resp.generated_text) {
            text = resp.generated_text;
          } else if (typeof resp === 'string') {
            text = resp;
          } else if (Array.isArray(resp) && typeof resp[0] === 'string') {
            text = resp[0];
          } else if (resp && resp[0] && resp[0].text) {
            text = resp[0].text;
          }

          if (!text) {
            showError(promptResult, 'No text returned by the model (unexpected response shape). Check console for raw output.');
            console.debug('HF text response (raw):', resp);
          } else {
            const safe = formatSafeHtml(text);
            showHtml(promptResult, safe);
          }
        } catch (err) {
          console.error('Prompt generation error:', err);
          showError(promptResult, `Error: ${err.message || err}`);
        } finally {
          setButtonLoading(generatePromptBtn, false);
        }
      };

      const handleImageGeneration = async () => {
        const prompt = (imagePromptInput.value || '').trim();
        if (!prompt) {
          showError(imageResult, 'Please enter an image prompt to generate an image.');
          return;
        }
        imageResult.innerHTML = '';
        imageLoading.classList.remove('hidden');
        setButtonLoading(generateImageBtn, true, 'Generating...');
        try {
          const resp = await hfImageGenerate(prompt);

          // If backend or HF returned binary blob
          if (resp && resp.blob) {
            const url = URL.createObjectURL(resp.blob);
            const img = document.createElement('img');
            img.src = url;
            img.alt = `AI generated: ${prompt}`;
            img.loading = 'lazy';
            img.decoding = 'async';
            img.className = 'rounded-lg shadow-lg mx-auto';
            imageResult.appendChild(img);
            return;
          }

          // HF may return JSON containing base64 or an array of output URLs.
          // Common shapes:
          //  - [{generated_text: "..."}] (unlikely for images)
          //  - { error: "..." }
          //  - [{image_base64: "..."}, ...] or { images: ["data:image/png;..."] }
          // - [{ "generated_image": "<base64>" }]
          // We'll try a few possibilities defensively:

          // If it's an array with objects containing 'image' or 'base64' fields
          if (Array.isArray(resp)) {
            // try to find a base64 string
            const obj = resp.find(r => (r.image || r.base64 || r.image_base64 || r.generated_image));
            const val = obj && (obj.image || obj.base64 || obj.image_base64 || obj.generated_image);
            if (val && typeof val === 'string') {
              // if it already begins with data:, use directly; else assume base64 png
              const src = val.startsWith('data:') ? val : `data:image/png;base64,${val}`;
              const img = document.createElement('img');
              img.src = src;
              img.alt = `AI generated: ${prompt}`;
              img.loading = 'lazy';
              img.decoding = 'async';
              img.className = 'rounded-lg shadow-lg mx-auto';
              imageResult.appendChild(img);
              return;
            }
          }

          // If resp.images is an array of data URLs or URLs
          if (resp && (resp.images || resp.output)) {
            const arr = resp.images || resp.output;
            if (Array.isArray(arr) && arr.length > 0) {
              const imageUrl = arr[0];
              const img = document.createElement('img');
              img.src = imageUrl;
              img.alt = `AI generated: ${prompt}`;
              img.loading = 'lazy';
              img.decoding = 'async';
              img.className = 'rounded-lg shadow-lg mx-auto';
              imageResult.appendChild(img);
              return;
            }
          }

          // If resp contains a base64 string in top-level 'generated_image' or 'b64_json'
          const b64 = resp && (resp.generated_image || resp.b64_json || resp.image_base64 || resp.base64);
          if (b64 && typeof b64 === 'string') {
            const src = b64.startsWith('data:') ? b64 : `data:image/png;base64,${b64}`;
            const img = document.createElement('img');
            img.src = src;
            img.alt = `AI generated: ${prompt}`;
            img.loading = 'lazy';
            img.decoding = 'async';
            img.className = 'rounded-lg shadow-lg mx-auto';
            imageResult.appendChild(img);
            return;
          }

          // As last resort, show the raw JSON for debugging
          showHtml(imageResult, `<pre class="p-3 bg-white/80 dark:bg-gray-800/70 rounded">${escapeHtml(JSON.stringify(resp, null, 2))}</pre>`);
          console.debug('HF image response (raw):', resp);
        } catch (err) {
          console.error('Image generation error:', err);
          showError(imageResult, `Error: ${err.message || err}`);
        } finally {
          imageLoading.classList.add('hidden');
          setButtonLoading(generateImageBtn, false);
        }
      };

      // Attach listeners
      generatePromptBtn.addEventListener('click', handlePromptGeneration);
      generateImageBtn.addEventListener('click', handleImageGeneration);

      // Navigation (simple)
      const navLinks = document.querySelectorAll('[data-page]');
      const pages = document.querySelectorAll('.page-content');
      const showPage = (id) => {
        pages.forEach(p => p.classList.remove('active'));
        const el = document.getElementById(id);
        if (el) el.classList.add('active');
      };
      navLinks.forEach(link => {
        link.addEventListener('click', (e) => {
          e.preventDefault();
          const page = link.getAttribute('data-page');
          if (page) {
            history.pushState({ page }, '', `#${page}`);
            showPage(page);
          }
        });
      });
      window.addEventListener('popstate', (ev) => {
        const page = (ev.state && ev.state.page) || (location.hash && location.hash.substring(1)) || 'home';
        showPage(page);
      });
    });
  </script>
</body>
</html>