docling-studio/index.html
2025-05-21 21:33:17 +02:00

908 lines
No EOL
31 KiB
HTML

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>DocTags Tools</title>
<style>
:root {
--primary-color: #4CAF50;
--primary-dark: #45a049;
--secondary-color: #2196F3;
--error-color: #d32f2f;
--success-color: #43a047;
--warning-color: #ff9800;
--text-color: #333;
--light-bg: #f5f5f5;
--border-color: #ddd;
}
body {
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
max-width: 1200px;
margin: 0 auto;
padding: 20px;
color: var(--text-color);
background-color: #f9f9f9;
}
header {
display: flex;
align-items: center;
margin-bottom: 30px;
}
header img {
width: 50px;
margin-right: 15px;
}
h1, h2, h3 {
color: #333;
margin-top: 0;
}
.form-group {
margin-bottom: 15px;
}
label {
display: block;
margin-bottom: 5px;
font-weight: bold;
}
select, input[type="number"] {
padding: 10px 12px;
width: 100%;
max-width: 300px;
border: 1px solid var(--border-color);
border-radius: 4px;
font-size: 14px;
}
input[type="checkbox"] {
margin-right: 8px;
}
button {
padding: 10px 16px;
background-color: var(--primary-color);
color: white;
border: none;
border-radius: 4px;
cursor: pointer;
margin-right: 10px;
margin-bottom: 10px;
font-weight: 500;
display: inline-flex;
align-items: center;
transition: background-color 0.2s;
}
button:hover {
background-color: var(--primary-dark);
}
button:disabled {
background-color: #cccccc;
cursor: not-allowed;
}
.workflow-container {
display: flex;
flex-direction: column;
gap: 30px;
margin-top: 30px;
}
.step {
background-color: white;
padding: 25px;
border-radius: 8px;
box-shadow: 0 2px 10px rgba(0,0,0,0.05);
transition: all 0.3s ease;
border-left: 5px solid var(--primary-color);
}
.step h3 {
margin-top: 0;
display: flex;
align-items: center;
}
.step h3 span {
background-color: var(--primary-color);
color: white;
width: 28px;
height: 28px;
border-radius: 50%;
display: inline-flex;
justify-content: center;
align-items: center;
margin-right: 10px;
font-size: 16px;
}
.status {
font-weight: 500;
margin-top: 15px;
padding: 8px 12px;
border-radius: 4px;
display: inline-block;
}
.error {
background-color: rgba(211, 47, 47, 0.1);
color: var(--error-color);
border-left: 3px solid var(--error-color);
}
.success {
background-color: rgba(67, 160, 71, 0.1);
color: var(--success-color);
border-left: 3px solid var(--success-color);
}
.working {
background-color: rgba(25, 118, 210, 0.1);
color: var(--secondary-color);
border-left: 3px solid var(--secondary-color);
}
.loader {
border: 3px solid #f3f3f3;
border-top: 3px solid var(--secondary-color);
border-radius: 50%;
width: 16px;
height: 16px;
animation: spin 2s linear infinite;
display: inline-block;
vertical-align: middle;
margin-right: 8px;
}
@keyframes spin {
0% { transform: rotate(0deg); }
100% { transform: rotate(360deg); }
}
.output-section {
margin-top: 20px;
padding: 15px;
border: 1px solid var(--border-color);
border-radius: 4px;
background-color: var(--light-bg);
white-space: pre-wrap;
max-height: 200px;
overflow-y: auto;
font-family: monospace;
font-size: 14px;
}
.preview-section {
margin-top: 20px;
display: flex;
flex-direction: column;
gap: 15px;
}
.preview-container {
background-color: white;
border: 1px solid var(--border-color);
border-radius: 8px;
padding: 15px;
box-shadow: 0 2px 5px rgba(0,0,0,0.05);
}
.preview-container h4 {
margin-top: 0;
margin-bottom: 15px;
color: var(--secondary-color);
border-bottom: 1px solid var(--border-color);
padding-bottom: 8px;
}
.pdf-preview {
max-height: 400px;
overflow-y: auto;
display: flex;
justify-content: center;
background-color: #f5f5f5;
padding: 15px;
border-radius: 4px;
}
.pdf-preview img {
max-width: 100%;
border: 1px solid #ddd;
box-shadow: 0 2px 5px rgba(0,0,0,0.1);
}
.doctags-preview {
max-height: 300px;
overflow-y: auto;
background-color: #272822;
color: #f8f8f2;
padding: 15px;
border-radius: 4px;
font-family: 'Courier New', monospace;
}
.visualization-preview {
display: flex;
justify-content: center;
max-height: 400px;
overflow-y: auto;
background-color: #f5f5f5;
padding: 15px;
border-radius: 4px;
}
.visualization-preview img {
max-width: 100%;
border: 1px solid #ddd;
box-shadow: 0 2px 5px rgba(0,0,0,0.1);
}
.extracted-images-preview {
display: grid;
grid-template-columns: repeat(auto-fill, minmax(200px, 1fr));
gap: 15px;
max-height: 400px;
overflow-y: auto;
padding: 10px;
background-color: #f5f5f5;
border-radius: 4px;
}
.extracted-image {
background-color: white;
padding: 10px;
border-radius: 4px;
box-shadow: 0 1px 3px rgba(0,0,0,0.1);
transition: transform 0.2s;
}
.extracted-image:hover {
transform: scale(1.02);
}
.extracted-image img {
width: 100%;
border-radius: 4px;
display: block;
}
.extracted-image .caption {
margin-top: 8px;
font-size: 12px;
color: #666;
text-align: center;
overflow: hidden;
text-overflow: ellipsis;
white-space: nowrap;
}
.hidden {
display: none !important;
}
.controls {
display: flex;
gap: 10px;
margin-top: 10px;
}
#environment-check {
margin-top: 40px;
padding: 20px;
background-color: white;
border-radius: 8px;
box-shadow: 0 2px 10px rgba(0,0,0,0.05);
}
.env-success {
color: var(--success-color);
}
.env-error {
color: var(--error-color);
}
.debug-section {
margin-top: 50px;
border-top: 1px solid var(--border-color);
padding-top: 20px;
}
.debug-section button {
background-color: #607d8b;
}
.debug-section button:hover {
background-color: #546e7a;
}
/* Tabs for extracted images */
.tabs {
display: flex;
margin-bottom: 15px;
}
.tab {
padding: 8px 16px;
background-color: #f0f0f0;
border: 1px solid #ddd;
cursor: pointer;
}
.tab.active {
background-color: white;
border-bottom-color: white;
font-weight: bold;
}
/* Responsive adjustments */
@media (max-width: 768px) {
.extracted-images-preview {
grid-template-columns: repeat(auto-fill, minmax(150px, 1fr));
}
}
</style>
</head>
<body>
<header>
<h1>DocTags Analyzer and Visualizer</h1>
</header>
<div id="environment-check" class="hidden">
<h3>Environment Check</h3>
<div id="env-details"></div>
</div>
<!-- Main form controls -->
<div class="form-group">
<label for="pdf_file">PDF File:</label>
<select id="pdf_file" name="pdf_file"></select>
<div id="pdf-load-status"></div>
</div>
<div class="form-group">
<label for="page_num">Page Number:</label>
<input type="number" id="page_num" name="page_num" value="1" min="1">
</div>
<div class="form-group">
<label for="adjust">
<input type="checkbox" id="adjust" name="adjust" checked>
Auto-adjust coordinates
</label>
</div>
<!-- Workflow steps -->
<div class="workflow-container">
<!-- Step 1: Analyze PDF -->
<div class="step" id="step-analyze">
<h3><span>1</span> Analyze PDF</h3>
<p>Extract DocTags structure from the PDF page</p>
<button id="analyzer-btn" onclick="runScript('analyzer')">Run Analyzer</button>
<div id="analyzer-status" class="status"></div>
<!-- PDF and DocTags Preview Section -->
<div id="analyze-preview-section" class="preview-section hidden">
<div class="preview-container">
<h4>PDF Preview</h4>
<div class="pdf-preview" id="pdf-preview">
<!-- PDF preview will be inserted here -->
</div>
</div>
<div class="preview-container">
<h4>Generated DocTags</h4>
<div class="doctags-preview" id="doctags-preview">
<!-- DocTags content will be inserted here -->
</div>
</div>
</div>
</div>
<!-- Step 2: Generate Visualization -->
<div class="step" id="step-visualize">
<h3><span>2</span> Generate Visualization</h3>
<p>Create visual representation of the extracted DocTags</p>
<button id="visualizer-btn" onclick="runScript('visualizer')" disabled>Run Visualizer</button>
<div id="visualizer-status" class="status"></div>
<!-- Visualization Preview Section -->
<div id="visualize-preview-section" class="preview-section hidden">
<div class="preview-container">
<h4>Zones Visualization</h4>
<div class="visualization-preview" id="visualization-preview">
<!-- Visualization image will be inserted here -->
</div>
</div>
</div>
</div>
<!-- Step 3: Extract Pictures -->
<div class="step" id="step-extract">
<h3><span>3</span> Extract Pictures</h3>
<p>Extract images from the document based on DocTags</p>
<button id="extractor-btn" onclick="runScript('extractor')" disabled>Run Picture Extractor</button>
<div id="extractor-status" class="status"></div>
<!-- Extracted Images Preview Section -->
<div id="extract-preview-section" class="preview-section hidden">
<div class="preview-container">
<h4>Extracted Images</h4>
<div class="extracted-images-preview" id="extracted-images-preview">
<!-- Extracted images will be inserted here -->
</div>
</div>
</div>
</div>
</div>
<div id="output" class="output-section hidden"></div>
<!-- Debug Section -->
<div class="debug-section">
<h3>Advanced Options</h3>
<button onclick="checkEnvironment()">Run Environment Check</button>
<button onclick="manuallyRunScript()">Run Manual Command</button>
</div>
<script>
// Keep track of active tasks
const activeTasks = {};
let pollingInterval = null;
let pdfPages = {};
// Load PDF files on page load
window.addEventListener('DOMContentLoaded', function() {
const pdfStatus = document.getElementById('pdf-load-status');
pdfStatus.innerHTML = '<div class="loader"></div> Loading PDF files...';
fetch('/pdf-files')
.then(response => {
if (!response.ok) {
throw new Error('Failed to load PDF files');
}
return response.json();
})
.then(data => {
const select = document.getElementById('pdf_file');
if (data.length === 0) {
pdfStatus.innerHTML = '<span class="error">No PDF files found in the current directory</span>';
return;
}
data.forEach(file => {
const option = document.createElement('option');
option.value = file;
option.textContent = file;
select.appendChild(option);
});
pdfStatus.innerHTML = '<span class="success">Loaded ' + data.length + ' PDF files</span>';
// Load the preview for the first PDF
if (data.length > 0) {
loadPdfPreview(data[0], 1);
}
})
.catch(error => {
console.error('Error loading PDFs:', error);
pdfStatus.innerHTML = '<span class="error">Error: ' + error.message + '</span>';
});
// Add event listeners for PDF and page selection changes
document.getElementById('pdf_file').addEventListener('change', function() {
const pageNum = document.getElementById('page_num').value;
loadPdfPreview(this.value, pageNum);
});
document.getElementById('page_num').addEventListener('change', function() {
const pdfFile = document.getElementById('pdf_file').value;
if (pdfFile) {
loadPdfPreview(pdfFile, this.value);
}
});
// Run an environment check on startup
checkEnvironment();
});
// Function to load PDF preview
function loadPdfPreview(pdfFile, pageNum) {
const previewContainer = document.getElementById('pdf-preview');
previewContainer.innerHTML = '<div class="loader"></div> Loading PDF preview...';
// Generate a temp preview for the PDF page
const formData = new FormData();
formData.append('pdf_file', pdfFile);
formData.append('page_num', pageNum);
fetch('/generate-preview', {
method: 'POST',
body: formData
})
.then(response => response.json())
.then(data => {
if (data.success && data.preview_image) {
const img = document.createElement('img');
img.src = data.preview_image + '?t=' + new Date().getTime(); // Add timestamp to prevent caching
img.alt = `Page ${pageNum} of ${pdfFile}`;
// Clear previous content and add the new image
previewContainer.innerHTML = '';
previewContainer.appendChild(img);
// Store page count if available
if (data.page_count) {
pdfPages[pdfFile] = data.page_count;
document.getElementById('page_num').max = data.page_count;
}
} else {
previewContainer.innerHTML = '<div class="error">Failed to load preview: ' + (data.error || 'Unknown error') + '</div>';
}
})
.catch(error => {
console.error('Error loading PDF preview:', error);
previewContainer.innerHTML = '<div class="error">Error loading preview: ' + error.message + '</div>';
});
}
// Check the environment
function checkEnvironment() {
const envDiv = document.getElementById('environment-check');
const envDetails = document.getElementById('env-details');
envDiv.classList.remove('hidden');
envDetails.innerHTML = '<div class="loader"></div> Checking environment...';
fetch('/check-environment')
.then(response => response.json())
.then(data => {
let html = '<ul>';
// Current working directory
html += '<li>Working directory: <code>' + data.cwd + '</code></li>';
// Python version
html += '<li>Python version: <code>' + data.python_version + '</code></li>';
// Required scripts check
if (data.missing_scripts.length === 0) {
html += '<li class="env-success">✓ All required scripts found</li>';
} else {
html += '<li class="env-error">✗ Missing scripts: <code>' + data.missing_scripts.join(', ') + '</code></li>';
}
// PDF files check
if (data.pdf_files.length > 0) {
html += '<li class="env-success">✓ Found ' + data.pdf_files.length + ' PDF files: <code>' + data.pdf_files.join(', ') + '</code></li>';
} else {
html += '<li class="env-error">✗ No PDF files found in the working directory</li>';
}
// Results directory check
if (data.results_dir_exists) {
html += '<li class="env-success">✓ Results directory exists</li>';
if (data.results_dir_writable) {
html += '<li class="env-success">✓ Results directory is writable</li>';
} else {
html += '<li class="env-error">✗ Results directory is not writable</li>';
}
} else {
html += '<li class="env-error">✗ Results directory does not exist</li>';
}
html += '</ul>';
// List of all files for debugging
html += '<details><summary>All files in directory (' + data.files.length + ' files)</summary><pre>' +
data.files.join('\n') + '</pre></details>';
envDetails.innerHTML = html;
})
.catch(error => {
envDetails.innerHTML = '<div class="env-error">Error checking environment: ' + error.message + '</div>';
});
}
// Manual command execution for debugging
function manuallyRunScript() {
const command = prompt("Enter command to run (e.g., 'python analyzer.py --image document.pdf --page 1')");
if (!command) return;
const outputDiv = document.getElementById('output');
outputDiv.textContent = 'Running command: ' + command + '\nPlease wait...';
outputDiv.classList.remove('hidden');
// Create form data
const formData = new FormData();
formData.append('command', command);
// Send the command directly to backend
fetch('/run-manual-command', {
method: 'POST',
body: formData
})
.then(response => response.json())
.then(data => {
outputDiv.textContent = 'Command: ' + command + '\n\n' +
(data.success ? 'Success!\n\n' : 'Failed!\n\n') +
(data.output || '') +
(data.error ? '\n\nError: ' + data.error : '');
})
.catch(error => {
outputDiv.textContent = 'Error running command: ' + error.message;
});
}
// Start polling for task updates
function startPolling() {
if (pollingInterval) {
return; // Already polling
}
pollingInterval = setInterval(pollTasks, 1000);
}
// Stop polling when no active tasks
function checkAndStopPolling() {
if (Object.keys(activeTasks).length === 0 && pollingInterval) {
clearInterval(pollingInterval);
pollingInterval = null;
}
}
// Function to load DocTags preview
function loadDocTagsPreview(doctagsPath) {
fetch('/get-doctags-content?path=' + encodeURIComponent(doctagsPath))
.then(response => response.text())
.then(content => {
const doctagsPreview = document.getElementById('doctags-preview');
doctagsPreview.textContent = content;
document.getElementById('analyze-preview-section').classList.remove('hidden');
})
.catch(error => {
console.error('Error loading DocTags content:', error);
});
}
// Function to load visualization preview
function loadVisualizationPreview(imagePath) {
const container = document.getElementById('visualization-preview');
const img = document.createElement('img');
img.src = imagePath + '?t=' + new Date().getTime();
img.alt = 'DocTags Visualization';
container.innerHTML = '';
container.appendChild(img);
document.getElementById('visualize-preview-section').classList.remove('hidden');
}
// Function to load extracted images preview
function loadExtractedImagesPreview() {
const pdfFile = document.getElementById('pdf_file').value;
const pageNum = document.getElementById('page_num').value;
// Request the extracted images list
fetch(`/get-extracted-images?pdf=${encodeURIComponent(pdfFile)}&page=${pageNum}`)
.then(response => response.json())
.then(data => {
const container = document.getElementById('extracted-images-preview');
container.innerHTML = '';
if (data.success && data.images && data.images.length > 0) {
data.images.forEach(image => {
const imageDiv = document.createElement('div');
imageDiv.className = 'extracted-image';
const img = document.createElement('img');
img.src = image.path + '?t=' + new Date().getTime();
img.alt = image.caption || `Extracted image ${image.id}`;
imageDiv.appendChild(img);
if (image.caption) {
const caption = document.createElement('div');
caption.className = 'caption';
caption.textContent = image.caption;
imageDiv.appendChild(caption);
}
container.appendChild(imageDiv);
});
} else {
container.innerHTML = '<div class="error">No extracted images found</div>';
}
document.getElementById('extract-preview-section').classList.remove('hidden');
})
.catch(error => {
console.error('Error loading extracted images:', error);
const container = document.getElementById('extracted-images-preview');
container.innerHTML = '<div class="error">Error loading images: ' + error.message + '</div>';
});
}
// Poll for task updates
function pollTasks() {
for (const taskId in activeTasks) {
const taskInfo = activeTasks[taskId];
fetch(`/task-status/${taskId}`)
.then(response => {
if (!response.ok) {
throw new Error('Failed to check task status');
}
return response.json();
})
.then(data => {
// Update status display
const statusElement = document.getElementById(`${taskInfo.type}-status`);
if (data.done) {
if (data.success) {
// Task completed successfully
statusElement.innerHTML = '<span class="success">✓ Completed</span>';
// Enable button
document.getElementById(`${taskInfo.type}-btn`).disabled = false;
// Display output
const outputDiv = document.getElementById('output');
outputDiv.textContent = data.output;
outputDiv.classList.remove('hidden');
// Update preview based on task type
if (taskInfo.type === 'analyzer') {
// Load DocTags preview
loadDocTagsPreview(data.doctags_file);
// Enable next step button
document.getElementById('visualizer-btn').disabled = false;
}
else if (taskInfo.type === 'visualizer' && data.image_file) {
// Load visualization preview
loadVisualizationPreview('/' + data.image_file);
// Enable next step button
document.getElementById('extractor-btn').disabled = false;
}
else if (taskInfo.type === 'extractor') {
// Load extracted images preview
loadExtractedImagesPreview();
}
// Remove from active tasks
delete activeTasks[taskId];
checkAndStopPolling();
} else {
// Task failed
statusElement.innerHTML = '<span class="error">✗ Failed: ' + (data.error || 'Unknown error') + '</span>';
document.getElementById(`${taskInfo.type}-btn`).disabled = false;
// Display error output
const outputDiv = document.getElementById('output');
outputDiv.textContent = 'Error: ' + (data.error || 'Unknown error');
outputDiv.classList.remove('hidden');
// Remove from active tasks
delete activeTasks[taskId];
checkAndStopPolling();
}
} else {
// Still running
statusElement.innerHTML = '<div class="loader"></div><span class="working">Running...</span>';
}
})
.catch(error => {
console.error('Error checking task status:', error);
});
}
}
function runScript(script) {
const pdfFile = document.getElementById('pdf_file').value;
const pageNum = document.getElementById('page_num').value;
const adjust = document.getElementById('adjust').checked;
if (!pdfFile) {
alert('Please select a PDF file');
return;
}
// Disable button
const button = document.getElementById(`${script}-btn`);
button.disabled = true;
// Create form data
const formData = new FormData();
formData.append('pdf_file', pdfFile);
formData.append('page_num', pageNum);
formData.append('adjust', adjust);
// Show running status
const statusElement = document.getElementById(`${script}-status`);
statusElement.innerHTML = '<div class="loader"></div><span class="working">Starting...</span>';
// Clear previous output
document.getElementById('output').classList.add('hidden');
// Determine endpoint
let endpoint;
switch(script) {
case 'analyzer':
endpoint = '/run-analyzer';
// Disable next step buttons
document.getElementById('visualizer-btn').disabled = true;
document.getElementById('extractor-btn').disabled = true;
break;
case 'visualizer':
endpoint = '/run-visualizer';
// Disable next step button
document.getElementById('extractor-btn').disabled = true;
break;
case 'extractor':
endpoint = '/run-extractor';
break;
}
// Send request
fetch(endpoint, {
method: 'POST',
body: formData
})
.then(response => {
if (!response.ok) {
throw new Error('Failed to start task');
}
return response.json();
})
.then(data => {
if (data.success && data.task_id) {
// Store task information
activeTasks[data.task_id] = {
type: script,
pageNum: pageNum
};
// Start polling for updates
startPolling();
// Update status
statusElement.innerHTML = '<div class="loader"></div><span class="working">Running...</span>';
// Show initial output
const outputDiv = document.getElementById('output');
outputDiv.textContent = data.message || 'Task started, please wait...';
outputDiv.classList.remove('hidden');
} else {
// Failed to start task
statusElement.innerHTML = '<span class="error">✗ Failed to start task</span>';
button.disabled = false;
// Show error
const outputDiv = document.getElementById('output');
outputDiv.textContent = 'Error: ' + (data.error || 'Failed to start task');
outputDiv.classList.remove('hidden');
}
})
.catch(error => {
console.error('Error starting task:', error);
statusElement.innerHTML = '<span class="error">✗ ' + error.message + '</span>';
button.disabled = false;
// Show error
const outputDiv = document.getElementById('output');
outputDiv.textContent = 'Error: ' + error.message;
outputDiv.classList.remove('hidden');
});
}
</script>
</body>
</html>