|
|
<body> |
|
|
<div class="container"> |
|
|
<h1>HTML Compressor for LLM</h1> |
|
|
<p>Compress HTML content for optimal LLM processing while preserving essential structure.</p> |
|
|
|
|
|
<form id="compressorForm"> |
|
|
<textarea |
|
|
name="html" |
|
|
id="htmlInput" |
|
|
placeholder="Paste your HTML here or upload a file..." |
|
|
></textarea> |
|
|
|
|
|
<div class="options-container"> |
|
|
<h3>Compression Options</h3> |
|
|
<div class="option-grid"> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="cleanHead" name="cleanHead" checked> |
|
|
<label for="cleanHead">Clean head section</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="removeScripts" name="removeScripts" checked> |
|
|
<label for="removeScripts">Remove scripts</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="removeStyles" name="removeStyles" checked> |
|
|
<label for="removeStyles">Remove styles</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="handleRepeatingElements" name="handleRepeatingElements" checked> |
|
|
<label for="handleRepeatingElements">Handle repeating elements</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="truncateText" name="truncateText" checked> |
|
|
<label for="truncateText">Truncate text</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<label for="truncateLength">Max text length:</label> |
|
|
<input type="number" id="truncateLength" name="truncateLength" value="100" min="10" max="1000"> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="minifyHtml" name="minifyHtml" checked> |
|
|
<label for="minifyHtml">Minify HTML</label> |
|
|
</div> |
|
|
<div class="option-item"> |
|
|
<input type="checkbox" id="removeMedia" name="removeMedia" checked> |
|
|
<label for="removeMedia">Remove media</label> |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="extraction-container"> |
|
|
<h3>Data Extraction</h3> |
|
|
<textarea |
|
|
id="extractionQuery" |
|
|
placeholder="Enter your extraction query (e.g., 'extract product title and price')" |
|
|
></textarea> |
|
|
<div class="button-group"> |
|
|
<button type="button" id="generateScript">Generate Extraction Script</button> |
|
|
<button type="button" id="executeExtraction" disabled>Execute Extraction</button> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="button-group"> |
|
|
<input type="file" accept=".html,.htm" id="fileInput"> |
|
|
<button type="submit">Process HTML</button> |
|
|
</div> |
|
|
</form> |
|
|
|
|
|
<div id="operationStatus" class="operation-status" style="display: none;"> |
|
|
<h3>Operation Status</h3> |
|
|
<div class="status-grid"></div> |
|
|
</div> |
|
|
|
|
|
<div id="stats" class="stats-grid" style="display: none;"></div> |
|
|
|
|
|
<div class="results-container" style="display: none;"> |
|
|
<div class="results-tabs"> |
|
|
<div class="tab active" data-view="html">Compressed HTML</div> |
|
|
<div class="tab" data-view="json">JSON Structure</div> |
|
|
<div class="tab" data-view="extraction">Extraction Results</div> |
|
|
</div> |
|
|
|
|
|
<div class="result-panel" id="htmlView"> |
|
|
<div class="result-header"> |
|
|
<h3>HTML Output</h3> |
|
|
<div class="button-group"> |
|
|
<button onclick="copyResult('html')">Copy</button> |
|
|
<button onclick="downloadResult('html')">Download</button> |
|
|
</div> |
|
|
</div> |
|
|
<div class="result-content"> |
|
|
<pre><code class="language-html" id="htmlOutput"></code></pre> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="result-panel" id="jsonView" style="display: none;"> |
|
|
<div class="result-header"> |
|
|
<h3>JSON Structure</h3> |
|
|
<div class="button-group"> |
|
|
<button onclick="copyResult('json')">Copy</button> |
|
|
<button onclick="downloadResult('json')">Download</button> |
|
|
</div> |
|
|
</div> |
|
|
<div class="result-content"> |
|
|
<pre><code class="language-json" id="jsonOutput"></code></pre> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="result-panel" id="extractionView" style="display: none;"> |
|
|
<div class="result-header"> |
|
|
<h3>Extraction Results</h3> |
|
|
<div class="button-group"> |
|
|
<button onclick="copyResult('extraction')">Copy</button> |
|
|
<button onclick="downloadResult('extraction')">Download</button> |
|
|
</div> |
|
|
</div> |
|
|
<div class="result-content"> |
|
|
<pre><code class="language-json" id="extractionOutput"></code></pre> |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
|
|
|
<div class="copy-feedback">Copied to clipboard!</div> |
|
|
</div> |
|
|
|
|
|
<script src="https://cdnjs.cloudflare.com/ajax/libs/prism/1.24.1/prism.min.js"></script> |
|
|
<script src="https://cdnjs.cloudflare.com/ajax/libs/prism/1.24.1/components/prism-markup.min.js"></script> |
|
|
<script src="https://cdnjs.cloudflare.com/ajax/libs/prism/1.24.1/components/prism-json.min.js"></script> |
|
|
<script> |
|
|
const API_URL = 'https://elevatics-ai-web-scraper-chat.hf.space/api/v1/generate-cheerio-script'; |
|
|
const API_KEY = 'ae54a922-ed3a-4634-be4a-4e4dd470800a'; |
|
|
|
|
|
let currentCheerioScript = null; |
|
|
|
|
|
const form = document.getElementById('compressorForm'); |
|
|
const fileInput = document.getElementById('fileInput'); |
|
|
const htmlInput = document.getElementById('htmlInput'); |
|
|
const resultsContainer = document.querySelector('.results-container'); |
|
|
const statsContainer = document.getElementById('stats'); |
|
|
const copyFeedback = document.querySelector('.copy-feedback'); |
|
|
|
|
|
|
|
|
document.querySelectorAll('.tab').forEach(tab => { |
|
|
tab.addEventListener('click', () => { |
|
|
|
|
|
document.querySelectorAll('.tab').forEach(t => t.classList.remove('active')); |
|
|
tab.classList.add('active'); |
|
|
|
|
|
|
|
|
const view = tab.dataset.view; |
|
|
document.getElementById('htmlView').style.display = view === 'html' ? 'block' : 'none'; |
|
|
document.getElementById('jsonView').style.display = view === 'json' ? 'block' : 'none'; |
|
|
document.getElementById('extractionView').style.display = view === 'extraction' ? 'block' : 'none'; |
|
|
}); |
|
|
}); |
|
|
|
|
|
|
|
|
fileInput.addEventListener('change', (e) => { |
|
|
const file = e.target.files[0]; |
|
|
if (file) { |
|
|
const reader = new FileReader(); |
|
|
reader.onload = (e) => htmlInput.value = e.target.result; |
|
|
reader.readAsText(file); |
|
|
} |
|
|
}); |
|
|
|
|
|
|
|
|
async function generateCheerioScript() { |
|
|
const htmlContent = document.getElementById('htmlOutput').textContent; |
|
|
const userInput = document.getElementById('extractionQuery').value; |
|
|
|
|
|
if (!htmlContent || !userInput) { |
|
|
alert('Please process HTML and enter an extraction query first'); |
|
|
return; |
|
|
} |
|
|
|
|
|
try { |
|
|
const response = await fetch(API_URL, { |
|
|
method: 'POST', |
|
|
headers: { |
|
|
'accept': 'application/json', |
|
|
'x-api-key': API_KEY, |
|
|
'Content-Type': 'application/json', |
|
|
}, |
|
|
body: JSON.stringify({ |
|
|
html: htmlContent, |
|
|
user_input: userInput |
|
|
}) |
|
|
}); |
|
|
|
|
|
const data = await response.json(); |
|
|
|
|
|
if (data.status === 'success') { |
|
|
currentCheerioScript = data.cheerio_script; |
|
|
document.getElementById('executeExtraction').disabled = false; |
|
|
|
|
|
|
|
|
document.getElementById('jsonOutput').textContent = JSON.stringify({ |
|
|
cheerio_script: currentCheerioScript |
|
|
}, null, 2); |
|
|
Prism.highlightAll(); |
|
|
} else { |
|
|
alert('Failed to generate extraction script'); |
|
|
} |
|
|
} catch (error) { |
|
|
alert('Error generating script: ' + error.message); |
|
|
} |
|
|
} |
|
|
|
|
|
|
|
|
async function executeExtraction() { |
|
|
if (!currentCheerioScript) { |
|
|
alert('Please generate an extraction script first'); |
|
|
return; |
|
|
} |
|
|
|
|
|
const htmlContent = document.getElementById('htmlOutput').textContent; |
|
|
|
|
|
try { |
|
|
const response = await fetch('/extract', { |
|
|
method: 'POST', |
|
|
headers: { |
|
|
'Content-Type': 'application/json' |
|
|
}, |
|
|
body: JSON.stringify({ |
|
|
html: htmlContent, |
|
|
script: currentCheerioScript |
|
|
}) |
|
|
}); |
|
|
|
|
|
const data = await response.json(); |
|
|
|
|
|
document.getElementById('extractionOutput').textContent = |
|
|
JSON.stringify(data, null, 2); |
|
|
Prism.highlightAll(); |
|
|
|
|
|
|
|
|
document.querySelector('[data-view="extraction"]').click(); |
|
|
} catch (error) { |
|
|
alert('Error executing extraction: ' + error.message); |
|
|
} |
|
|
} |
|
|
|
|
|
|
|
|
form.addEventListener('submit', async (e) => { |
|
|
e.preventDefault(); |
|
|
|
|
|
const formData = new FormData(form); |
|
|
|
|
|
|
|
|
document.querySelectorAll('input[type="checkbox"]').forEach(checkbox => { |
|
|
formData.set(checkbox.name, checkbox.checked); |
|
|
}); |
|
|
|
|
|
try { |
|
|
const response = await fetch('/process', { |
|
|
method: 'POST', |
|
|
body: formData, |
|
|
}); |
|
|
|
|
|
const data = await response.json(); |
|
|
|
|
|
if (data.error) { |
|
|
alert(data.error); |
|
|
return; |
|
|
} |
|
|
|
|
|
|
|
|
const statusContainer = document.querySelector('#operationStatus'); |
|
|
const statusGrid = statusContainer.querySelector('.status-grid'); |
|
|
statusContainer.style.display = 'block'; |
|
|
|
|
|
statusGrid.innerHTML = Object.entries(data.operationStatus) |
|
|
.map(([operation, status]) => ` |
|
|
<div class="status-item"> |
|
|
<div class="status-icon ${status.success ? 'status-success' : 'status-error'}"> |
|
|
${status.success ? '✓' : '✗'} |
|
|
</div> |
|
|
<div> |
|
|
<div>${formatLabel(operation)}</div> |
|
|
${status.error ? `<div class="status-message">Error: ${status.error}</div>` : ''} |
|
|
</div> |
|
|
</div> |
|
|
`).join(''); |
|
|
|
|
|
|
|
|
statsContainer.style.display = 'grid'; |
|
|
statsContainer.innerHTML = Object.entries(data.stats) |
|
|
.map(([key, value]) => ` |
|
|
<div class="stat-item"> |
|
|
<div class="stat-label">${formatLabel(key)}</div> |
|
|
<div class="stat-value">${value}</div> |
|
|
</div> |
|
|
`).join(''); |
|
|
|
|
|
|
|
|
resultsContainer.style.display = 'block'; |
|
|
|
|
|
|
|
|
document.getElementById('htmlOutput').textContent = data.result.html; |
|
|
document.getElementById('jsonOutput').textContent = data.result.json; |
|
|
|
|
|
|
|
|
Prism.highlightAll(); |
|
|
} catch (err) { |
|
|
alert('Error processing HTML: ' + err.message); |
|
|
} |
|
|
}); |
|
|
|
|
|
|
|
|
document.getElementById('generateScript').addEventListener('click', generateCheerioScript); |
|
|
document.getElementById('executeExtraction').addEventListener('click', executeExtraction); |
|
|
|
|
|
|
|
|
function formatLabel(key) { |
|
|
return key |
|
|
.replace(/([A-Z])/g, ' $1') |
|
|
.replace(/([a-z])([A-Z])/g, '$1 $2') |
|
|
.toLowerCase() |
|
|
.replace(/^./, str => str.toUpperCase()) |
|
|
.replace('Html', 'HTML'); |
|
|
} |
|
|
|
|
|
async function copyResult(type) { |
|
|
const content = document.getElementById(`${type}Output`).textContent; |
|
|
try { |
|
|
await navigator.clipboard.writeText(content); |
|
|
showCopyFeedback(); |
|
|
} catch (err) { |
|
|
alert('Failed to copy to clipboard'); |
|
|
} |
|
|
} |
|
|
|
|
|
function downloadResult(type) { |
|
|
const content = document.getElementById(`${type}Output`).textContent; |
|
|
const blob = new Blob([content], { type: 'text/plain' }); |
|
|
const url = URL.createObjectURL(blob); |
|
|
const a = document.createElement('a'); |
|
|
a.href = url; |
|
|
a.download = `compressed.${type}`; |
|
|
document.body.appendChild(a); |
|
|
a.click(); |
|
|
document.body.removeChild(a); |
|
|
URL.revokeObjectURL(url); |
|
|
} |
|
|
|
|
|
function showCopyFeedback() { |
|
|
copyFeedback.style.display = 'block'; |
|
|
setTimeout(() => { |
|
|
copyFeedback.style.display = 'none'; |
|
|
}, 2000); |
|
|
} |
|
|
</script> |
|
|
</body> |
