Improve frontend and adapt API
This commit is contained in:
pjmalandrino
parent
e84a9715b2
commit
ce3a131ec2
3 changed files with 630 additions and 88 deletions
142
app.py
142
app.py
|
|
@ -4,7 +4,12 @@ import os
|
|||
import time
|
||||
import threading
|
||||
import logging
|
||||
import json
|
||||
import re
|
||||
from pathlib import Path
|
||||
import glob
|
||||
from PIL import Image
|
||||
import pdf2image
|
||||
|
||||
# Configure logging
|
||||
logging.basicConfig(level=logging.INFO,
|
||||
|
|
@ -358,6 +363,143 @@ def run_manual_command():
|
|||
logger.error(f"Error running manual command: {str(e)}")
|
||||
return jsonify({'success': False, 'error': str(e)}), 500
|
||||
|
||||
# NEW ROUTES TO SUPPORT ENHANCED UI
|
||||
|
||||
@app.route('/generate-preview', methods=['POST'])
|
||||
def generate_preview():
|
||||
"""Generate a preview image of a PDF page"""
|
||||
try:
|
||||
# Get parameters
|
||||
pdf_file = request.form.get('pdf_file')
|
||||
page_num = int(request.form.get('page_num', 1))
|
||||
|
||||
if not pdf_file or not os.path.exists(pdf_file):
|
||||
return jsonify({'success': False, 'error': 'PDF file not found'}), 400
|
||||
|
||||
# Ensure results directory exists
|
||||
results_dir = ensure_results_folder()
|
||||
preview_path = results_dir / f"preview_{os.path.basename(pdf_file)}_{page_num}.png"
|
||||
|
||||
# Convert PDF page to image
|
||||
try:
|
||||
# Get total page count
|
||||
from pdf2image.pdf2image import pdfinfo_from_path
|
||||
pdf_info = pdfinfo_from_path(pdf_file)
|
||||
page_count = pdf_info["Pages"]
|
||||
|
||||
# Ensure the page number is valid
|
||||
if page_num > page_count:
|
||||
return jsonify({'success': False, 'error': f'Page number {page_num} exceeds total pages {page_count}'}), 400
|
||||
|
||||
# Convert the page
|
||||
images = pdf2image.convert_from_path(
|
||||
pdf_file,
|
||||
dpi=150, # Lower DPI for preview
|
||||
first_page=page_num,
|
||||
last_page=page_num
|
||||
)
|
||||
|
||||
if images:
|
||||
# Resize for preview if too large
|
||||
MAX_HEIGHT = 600
|
||||
img = images[0]
|
||||
ratio = MAX_HEIGHT / img.height if img.height > MAX_HEIGHT else 1
|
||||
if ratio < 1:
|
||||
new_width = int(img.width * ratio)
|
||||
img = img.resize((new_width, MAX_HEIGHT), Image.Resampling.LANCZOS)
|
||||
|
||||
# Save the preview
|
||||
img.save(preview_path)
|
||||
|
||||
return jsonify({
|
||||
'success': True,
|
||||
'preview_image': f'/results/{preview_path.name}',
|
||||
'page_count': page_count
|
||||
})
|
||||
else:
|
||||
return jsonify({'success': False, 'error': 'Failed to convert PDF page'}), 500
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error converting PDF: {str(e)}")
|
||||
return jsonify({'success': False, 'error': str(e)}), 500
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error generating preview: {str(e)}")
|
||||
return jsonify({'success': False, 'error': str(e)}), 500
|
||||
|
||||
@app.route('/get-doctags-content')
|
||||
def get_doctags_content():
|
||||
"""Get the content of a DocTags file"""
|
||||
try:
|
||||
path = request.args.get('path')
|
||||
if not path or not os.path.exists(path):
|
||||
# Try prepending 'results/' if the path doesn't exist
|
||||
if not path.startswith('results/'):
|
||||
path = f"results/{path}"
|
||||
|
||||
if not os.path.exists(path):
|
||||
return "DocTags file not found", 404
|
||||
|
||||
with open(path, 'r', encoding='utf-8') as f:
|
||||
content = f.read()
|
||||
|
||||
return content
|
||||
except Exception as e:
|
||||
logger.error(f"Error reading DocTags file: {str(e)}")
|
||||
return f"Error: {str(e)}", 500
|
||||
|
||||
@app.route('/get-extracted-images')
|
||||
def get_extracted_images():
|
||||
"""Get a list of extracted images for a specific PDF page"""
|
||||
try:
|
||||
pdf_file = request.args.get('pdf')
|
||||
page_num = request.args.get('page', 1)
|
||||
|
||||
if not pdf_file:
|
||||
return jsonify({'success': False, 'error': 'PDF file not specified'}), 400
|
||||
|
||||
# Get the base name of the PDF without extension
|
||||
pdf_basename = os.path.splitext(os.path.basename(pdf_file))[0]
|
||||
|
||||
# Look for extracted images in the results/pictures directory
|
||||
pictures_dir = Path("results") / "pictures"
|
||||
|
||||
if not pictures_dir.exists():
|
||||
return jsonify({'success': True, 'images': []})
|
||||
|
||||
# Pattern to match: picture_*.*
|
||||
image_pattern = pictures_dir / "picture_*.png"
|
||||
image_files = glob.glob(str(image_pattern))
|
||||
|
||||
images = []
|
||||
for image_file in image_files:
|
||||
img_path = Path(image_file)
|
||||
img_id = "unknown"
|
||||
img_caption = ""
|
||||
|
||||
# Try to extract ID from filename (picture_1_caption.png)
|
||||
match = re.search(r'picture_(\d+)', img_path.name)
|
||||
if match:
|
||||
img_id = match.group(1)
|
||||
|
||||
# Look for a matching caption file
|
||||
caption_file = img_path.with_suffix('.txt')
|
||||
if caption_file.exists():
|
||||
with open(caption_file, 'r', encoding='utf-8') as f:
|
||||
img_caption = f.read().strip()
|
||||
|
||||
images.append({
|
||||
'id': img_id,
|
||||
'path': f'/results/pictures/{img_path.name}',
|
||||
'caption': img_caption
|
||||
})
|
||||
|
||||
return jsonify({'success': True, 'images': images})
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"Error getting extracted images: {str(e)}")
|
||||
return jsonify({'success': False, 'error': str(e)}), 500
|
||||
|
||||
if __name__ == '__main__':
|
||||
ensure_results_folder()
|
||||
app.run(debug=True, host='127.0.0.1', port=5000)
|
||||
BIN
dat-isia.pdf
Normal file
BIN
dat-isia.pdf
Normal file
Binary file not shown.
576
index.html
576
index.html
|
|
@ -5,133 +5,363 @@
|
|||
<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: Arial, sans-serif;
|
||||
max-width: 800px;
|
||||
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
|
||||
max-width: 1200px;
|
||||
margin: 0 auto;
|
||||
padding: 20px;
|
||||
color: var(--text-color);
|
||||
background-color: #f9f9f9;
|
||||
}
|
||||
h1 {
|
||||
|
||||
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 {
|
||||
padding: 8px;
|
||||
|
||||
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 15px;
|
||||
background-color: #4CAF50;
|
||||
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: #45a049;
|
||||
background-color: var(--primary-dark);
|
||||
}
|
||||
|
||||
button:disabled {
|
||||
background-color: #cccccc;
|
||||
cursor: not-allowed;
|
||||
}
|
||||
.output {
|
||||
margin-top: 20px;
|
||||
padding: 15px;
|
||||
border: 1px solid #ddd;
|
||||
border-radius: 4px;
|
||||
background-color: #f9f9f9;
|
||||
white-space: pre-wrap;
|
||||
max-height: 300px;
|
||||
overflow-y: auto;
|
||||
}
|
||||
.hidden {
|
||||
display: none;
|
||||
}
|
||||
img {
|
||||
max-width: 100%;
|
||||
margin-top: 10px;
|
||||
border: 1px solid #ddd;
|
||||
}
|
||||
.container {
|
||||
|
||||
.workflow-container {
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
gap: 20px;
|
||||
gap: 30px;
|
||||
margin-top: 30px;
|
||||
}
|
||||
|
||||
.step {
|
||||
background-color: #f5f5f5;
|
||||
padding: 15px;
|
||||
border-radius: 5px;
|
||||
border-left: 5px solid #4CAF50;
|
||||
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;
|
||||
}
|
||||
.status {
|
||||
font-weight: bold;
|
||||
margin-top: 5px;
|
||||
}
|
||||
.error {
|
||||
color: #d32f2f;
|
||||
}
|
||||
.success {
|
||||
color: #43a047;
|
||||
}
|
||||
.working {
|
||||
color: #1976d2;
|
||||
}
|
||||
.loader {
|
||||
border: 4px solid #f3f3f3;
|
||||
border-top: 4px solid #3498db;
|
||||
|
||||
.step h3 span {
|
||||
background-color: var(--primary-color);
|
||||
color: white;
|
||||
width: 28px;
|
||||
height: 28px;
|
||||
border-radius: 50%;
|
||||
width: 20px;
|
||||
height: 20px;
|
||||
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: 10px;
|
||||
margin-right: 8px;
|
||||
}
|
||||
|
||||
@keyframes spin {
|
||||
0% { transform: rotate(0deg); }
|
||||
100% { transform: rotate(360deg); }
|
||||
}
|
||||
#environment-check {
|
||||
|
||||
.output-section {
|
||||
margin-top: 20px;
|
||||
padding: 10px;
|
||||
background-color: #fff3cd;
|
||||
border: 1px solid #ffeeba;
|
||||
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: green;
|
||||
color: var(--success-color);
|
||||
}
|
||||
|
||||
.env-error {
|
||||
color: red;
|
||||
color: var(--error-color);
|
||||
}
|
||||
|
||||
.debug-section {
|
||||
margin-top: 30px;
|
||||
border-top: 1px solid #ccc;
|
||||
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>
|
||||
<h1>DocTags Analyzer and Visualizer</h1>
|
||||
<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>
|
||||
|
|
@ -150,37 +380,75 @@
|
|||
</label>
|
||||
</div>
|
||||
|
||||
<div class="container">
|
||||
<div class="step">
|
||||
<h3>Step 1: Analyze PDF</h3>
|
||||
<!-- 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>
|
||||
|
||||
<div class="step">
|
||||
<h3>Step 2: Generate Visualization</h3>
|
||||
<!-- 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')">Run Visualizer</button>
|
||||
<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>
|
||||
|
||||
<div class="step">
|
||||
<h3>Step 3: Extract Pictures</h3>
|
||||
<!-- 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')">Run Picture Extractor</button>
|
||||
<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 hidden"></div>
|
||||
|
||||
<div id="image-container" class="hidden">
|
||||
<h2>Visualization Result</h2>
|
||||
<img id="result-image" src="" alt="Visualization">
|
||||
</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>
|
||||
|
|
@ -189,6 +457,7 @@
|
|||
// Keep track of active tasks
|
||||
const activeTasks = {};
|
||||
let pollingInterval = null;
|
||||
let pdfPages = {};
|
||||
|
||||
// Load PDF files on page load
|
||||
window.addEventListener('DOMContentLoaded', function() {
|
||||
|
|
@ -216,16 +485,74 @@
|
|||
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');
|
||||
|
|
@ -331,6 +658,77 @@
|
|||
}
|
||||
}
|
||||
|
||||
// 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) {
|
||||
|
|
@ -360,18 +758,25 @@
|
|||
outputDiv.textContent = data.output;
|
||||
outputDiv.classList.remove('hidden');
|
||||
|
||||
// Show image if available for visualizer
|
||||
if (taskInfo.type === 'visualizer' && data.image_file) {
|
||||
document.getElementById('result-image').src = '/' + data.image_file + '?t=' + new Date().getTime();
|
||||
document.getElementById('image-container').classList.remove('hidden');
|
||||
}
|
||||
|
||||
// Enable next step button if applicable
|
||||
// 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') {
|
||||
}
|
||||
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];
|
||||
|
|
@ -428,11 +833,6 @@
|
|||
// Clear previous output
|
||||
document.getElementById('output').classList.add('hidden');
|
||||
|
||||
// Hide previous image if not running visualizer
|
||||
if (script !== 'visualizer') {
|
||||
document.getElementById('image-container').classList.add('hidden');
|
||||
}
|
||||
|
||||
// Determine endpoint
|
||||
let endpoint;
|
||||
switch(script) {
|
||||
|
|
|
|||
Loading…
Reference in a new issue