Add pdf visualization on first part

This commit is contained in:
pjmalandrino 2025-05-22 14:59:24 +02:00
parent c8fbe70232
commit 92de48359e
4 changed files with 259 additions and 0 deletions

21
app.py
View file

@ -39,6 +39,27 @@ def index():
def serve_static(filename):
return send_file(os.path.join('static', filename))
@app.route('/<path:filename>')
def serve_pdf(filename):
"""Serve PDF files from the current directory"""
try:
# Security check: only allow PDF files and prevent directory traversal
if not filename.endswith('.pdf') or '..' in filename or '/' in filename:
logger.warning(f"Blocked access to non-PDF or invalid file: {filename}")
return jsonify({"error": "Access denied"}), 403
# Check if file exists in current directory
file_path = Path(filename)
if not file_path.exists() or not file_path.is_file():
logger.error(f"PDF file not found: {filename}")
return jsonify({"error": f"PDF file not found: {filename}"}), 404
logger.info(f"Serving PDF file: {filename}")
return send_file(filename, mimetype='application/pdf')
except Exception as e:
logger.error(f"Error serving PDF {filename}: {str(e)}")
return jsonify({"error": f"Error serving PDF: {str(e)}"}), 500
@app.route('/pdf-files')
def pdf_files():
try:

View file

@ -62,6 +62,25 @@
<h3>Document Analysis</h3>
<p>Extract comprehensive document structure from your PDF using advanced AI-powered analysis. This process identifies headers, paragraphs, images, tables, and other document elements with precise coordinate information.</p>
<!-- PDF Preview Section -->
<div id="pdf-preview-container" class="pdf-preview-container hidden">
<h4>PDF Preview</h4>
<div class="pdf-preview-controls">
<button id="prev-page-btn" onclick="changePage(-1)" disabled>← Previous</button>
<button id="next-page-btn" onclick="changePage(1)" disabled>Next →</button>
<button id="refresh-preview-btn" onclick="refreshPreview()">Refresh</button>
<div class="pdf-preview-info">
<span id="page-info">Page 1 of 1</span> |
<span id="zoom-info">100%</span>
</div>
</div>
<div id="pdf-preview-loading" class="pdf-preview-loading hidden">
<div class="loader"></div>
Loading PDF preview...
</div>
<canvas id="pdf-preview-canvas" class="hidden"></canvas>
</div>
<button id="analyzer-btn" onclick="runScript('analyzer')">
Start Analysis
</button>
@ -146,6 +165,8 @@
</div>
</div>
<!-- PDF.js Library -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/pdf.js/3.11.174/pdf.min.js"></script>
<script src="static/app.js"></script>
</body>
</html>

View file

@ -2,6 +2,15 @@
const activeTasks = {};
let pollingInterval = null;
// PDF preview variables
let currentPdf = null;
let currentPageNum = 1;
let totalPages = 0;
let renderScale = 1.0;
// Initialize PDF.js
pdfjsLib.GlobalWorkerOptions.workerSrc = 'https://cdnjs.cloudflare.com/ajax/libs/pdf.js/3.11.174/pdf.worker.min.js';
// Tab switching functionality
function switchTab(tabName) {
// Hide all tab contents
@ -39,6 +48,143 @@ function updateProgress(completedSteps) {
}
}
// PDF Preview Functions
async function loadPdfPreview(pdfFile) {
if (!pdfFile) {
hidePdfPreview();
return;
}
const previewContainer = document.getElementById('pdf-preview-container');
const loadingDiv = document.getElementById('pdf-preview-loading');
const canvas = document.getElementById('pdf-preview-canvas');
// Show preview container and loading
previewContainer.classList.remove('hidden');
loadingDiv.classList.remove('hidden');
canvas.classList.add('hidden');
try {
// Load the PDF
const loadingTask = pdfjsLib.getDocument(`/${pdfFile}`);
currentPdf = await loadingTask.promise;
totalPages = currentPdf.numPages;
// Update page info
updatePageInfo();
updatePageControls();
// Render the current page
await renderPage();
// Hide loading, show canvas
loadingDiv.classList.add('hidden');
canvas.classList.remove('hidden');
} catch (error) {
console.error('Error loading PDF preview:', error);
loadingDiv.innerHTML = '<span class="error">Error loading PDF preview: ' + error.message + '</span>';
}
}
async function renderPage() {
if (!currentPdf) return;
const canvas = document.getElementById('pdf-preview-canvas');
const ctx = canvas.getContext('2d');
try {
// Get the page
const page = await currentPdf.getPage(currentPageNum);
// Calculate scale to fit container width (max 600px)
const viewport = page.getViewport({ scale: 1.0 });
const maxWidth = 600;
renderScale = Math.min(maxWidth / viewport.width, 1.5);
const scaledViewport = page.getViewport({ scale: renderScale });
// Set canvas dimensions
canvas.height = scaledViewport.height;
canvas.width = scaledViewport.width;
// Render the page
const renderContext = {
canvasContext: ctx,
viewport: scaledViewport
};
await page.render(renderContext).promise;
// Update zoom info
document.getElementById('zoom-info').textContent = Math.round(renderScale * 100) + '%';
} catch (error) {
console.error('Error rendering page:', error);
}
}
function updatePageInfo() {
document.getElementById('page-info').textContent = `Page ${currentPageNum} of ${totalPages}`;
}
function updatePageControls() {
document.getElementById('prev-page-btn').disabled = currentPageNum <= 1;
document.getElementById('next-page-btn').disabled = currentPageNum >= totalPages;
}
async function changePage(delta) {
const newPage = currentPageNum + delta;
if (newPage >= 1 && newPage <= totalPages) {
currentPageNum = newPage;
// Update the page number input
document.getElementById('page_num').value = currentPageNum;
updatePageInfo();
updatePageControls();
await renderPage();
}
}
async function refreshPreview() {
const pdfFile = document.getElementById('pdf_file').value;
if (pdfFile) {
await loadPdfPreview(pdfFile);
}
}
function hidePdfPreview() {
document.getElementById('pdf-preview-container').classList.add('hidden');
currentPdf = null;
totalPages = 0;
currentPageNum = 1;
}
// Event listeners for PDF selection and page changes
document.addEventListener('DOMContentLoaded', function() {
// PDF file selection change
document.getElementById('pdf_file').addEventListener('change', async function() {
const selectedFile = this.value;
if (selectedFile) {
await loadPdfPreview(selectedFile);
} else {
hidePdfPreview();
}
});
// Page number input change
document.getElementById('page_num').addEventListener('change', async function() {
const pageNum = parseInt(this.value);
if (pageNum >= 1 && pageNum <= totalPages && pageNum !== currentPageNum) {
currentPageNum = pageNum;
updatePageInfo();
updatePageControls();
await renderPage();
}
});
});
// Load PDF files on page load
window.addEventListener('DOMContentLoaded', function() {
const pdfStatus = document.getElementById('pdf-load-status');

View file

@ -68,6 +68,13 @@ h3 {
margin: 0 0 var(--spacing-sm) 0;
}
h4 {
color: var(--primary-dark);
font-weight: 500;
font-size: 1rem;
margin: 0 0 var(--spacing-sm) 0;
}
p {
color: var(--text-medium);
margin: 0 0 var(--spacing-lg) 0;
@ -272,6 +279,60 @@ label:has(input[type="checkbox"]) {
display: block;
}
/* PDF Preview Section */
.pdf-preview-container {
margin-top: var(--spacing-lg);
padding: var(--spacing-lg);
background: var(--surface-light);
border-radius: var(--radius-md);
border: 1px solid var(--border-light);
}
.pdf-preview-controls {
display: flex;
gap: var(--spacing-md);
margin-bottom: var(--spacing-md);
align-items: center;
}
.pdf-preview-controls button {
padding: var(--spacing-xs) var(--spacing-md);
font-size: 0.8rem;
background: var(--text-medium);
}
.pdf-preview-controls button:hover {
background: var(--text-dark);
}
.pdf-preview-controls button:disabled {
background: var(--border-light);
color: var(--text-light);
}
.pdf-preview-info {
font-size: 0.8rem;
color: var(--text-medium);
margin-left: auto;
}
#pdf-preview-canvas {
max-width: 100%;
border: 1px solid var(--border-light);
border-radius: var(--radius-sm);
box-shadow: var(--shadow-sm);
background: white;
display: block;
margin: 0 auto;
}
.pdf-preview-loading {
text-align: center;
padding: var(--spacing-xl);
color: var(--text-medium);
font-size: 0.9rem;
}
/* Buttons */
button {
padding: var(--spacing-md) var(--spacing-lg);
@ -531,4 +592,14 @@ button:disabled {
.tab-content {
padding: var(--spacing-lg);
}
.pdf-preview-controls {
flex-direction: column;
align-items: stretch;
}
.pdf-preview-info {
margin-left: 0;
text-align: center;
}
}