Files
invoice-master-poc-v2/packages/inference/inference/web/app.py
Yaojia Wang 33ada0350d WIP
2026-01-30 00:44:21 +01:00

937 lines
28 KiB
Python

"""
FastAPI Application Factory
Creates and configures the FastAPI application.
"""
from __future__ import annotations
import logging
from contextlib import asynccontextmanager
from pathlib import Path
from typing import TYPE_CHECKING
from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from fastapi.staticfiles import StaticFiles
from fastapi.responses import HTMLResponse
from .config import AppConfig, default_config
from inference.web.services import InferenceService
# Public API imports
from inference.web.api.v1.public import (
create_inference_router,
create_async_router,
set_async_service,
create_labeling_router,
)
# Async processing imports
from inference.data.async_request_db import AsyncRequestDB
from inference.web.workers.async_queue import AsyncTaskQueue
from inference.web.services.async_processing import AsyncProcessingService
from inference.web.dependencies import init_dependencies
from inference.web.core.rate_limiter import RateLimiter
# Admin API imports
from inference.web.api.v1.admin import (
create_annotation_router,
create_augmentation_router,
create_auth_router,
create_documents_router,
create_locks_router,
create_training_router,
)
from inference.web.core.scheduler import start_scheduler, stop_scheduler
from inference.web.core.autolabel_scheduler import start_autolabel_scheduler, stop_autolabel_scheduler
# Batch upload imports
from inference.web.api.v1.batch.routes import router as batch_upload_router
from inference.web.workers.batch_queue import init_batch_queue, shutdown_batch_queue
from inference.web.services.batch_upload import BatchUploadService
from inference.data.admin_db import AdminDB
if TYPE_CHECKING:
from collections.abc import AsyncGenerator
logger = logging.getLogger(__name__)
def create_app(config: AppConfig | None = None) -> FastAPI:
"""
Create and configure FastAPI application.
Args:
config: Application configuration. Uses default if not provided.
Returns:
Configured FastAPI application
"""
config = config or default_config
# Create model path resolver that reads from database
def get_active_model_path():
"""Resolve active model path from database."""
try:
db = AdminDB()
active_model = db.get_active_model_version()
if active_model and active_model.model_path:
return active_model.model_path
except Exception as e:
logger.warning(f"Failed to get active model from database: {e}")
return None
# Create inference service with database model resolver
inference_service = InferenceService(
model_config=config.model,
storage_config=config.storage,
model_path_resolver=get_active_model_path,
)
# Create async processing components
async_db = AsyncRequestDB()
rate_limiter = RateLimiter(async_db)
task_queue = AsyncTaskQueue(
max_size=config.async_processing.queue_max_size,
worker_count=config.async_processing.worker_count,
)
async_service = AsyncProcessingService(
inference_service=inference_service,
db=async_db,
queue=task_queue,
rate_limiter=rate_limiter,
async_config=config.async_processing,
storage_config=config.storage,
)
# Initialize dependencies for FastAPI
init_dependencies(async_db, rate_limiter)
set_async_service(async_service)
@asynccontextmanager
async def lifespan(app: FastAPI) -> AsyncGenerator[None, None]:
"""Application lifespan manager."""
logger.info("Starting Invoice Inference API...")
# Initialize database tables
try:
async_db.create_tables()
logger.info("Async database tables ready")
except Exception as e:
logger.error(f"Failed to initialize async database: {e}")
# Initialize inference service on startup
try:
inference_service.initialize()
logger.info("Inference service ready")
except Exception as e:
logger.error(f"Failed to initialize inference service: {e}")
# Continue anyway - service will retry on first request
# Start async processing service
try:
async_service.start()
logger.info("Async processing service started")
except Exception as e:
logger.error(f"Failed to start async processing: {e}")
# Start batch upload queue
try:
admin_db = AdminDB()
batch_service = BatchUploadService(admin_db)
init_batch_queue(batch_service)
logger.info("Batch upload queue started")
except Exception as e:
logger.error(f"Failed to start batch upload queue: {e}")
# Start training scheduler
try:
start_scheduler()
logger.info("Training scheduler started")
except Exception as e:
logger.error(f"Failed to start training scheduler: {e}")
# Start auto-label scheduler
try:
start_autolabel_scheduler()
logger.info("AutoLabel scheduler started")
except Exception as e:
logger.error(f"Failed to start autolabel scheduler: {e}")
yield
logger.info("Shutting down Invoice Inference API...")
# Stop auto-label scheduler
try:
stop_autolabel_scheduler()
logger.info("AutoLabel scheduler stopped")
except Exception as e:
logger.error(f"Error stopping autolabel scheduler: {e}")
# Stop training scheduler
try:
stop_scheduler()
logger.info("Training scheduler stopped")
except Exception as e:
logger.error(f"Error stopping training scheduler: {e}")
# Stop batch upload queue
try:
shutdown_batch_queue()
logger.info("Batch upload queue stopped")
except Exception as e:
logger.error(f"Error stopping batch upload queue: {e}")
# Stop async processing service
try:
async_service.stop(timeout=30.0)
logger.info("Async processing service stopped")
except Exception as e:
logger.error(f"Error stopping async service: {e}")
# Close database connection
try:
async_db.close()
logger.info("Database connection closed")
except Exception as e:
logger.error(f"Error closing database: {e}")
# Create FastAPI app
# Store inference service for access by routes (e.g., model reload)
# This will be set after app creation
app = FastAPI(
title="Invoice Field Extraction API",
description="""
REST API for extracting fields from Swedish invoices.
## Features
- YOLO-based field detection
- OCR text extraction
- Field normalization and validation
- Visualization of detections
## Supported Fields
- InvoiceNumber
- InvoiceDate
- InvoiceDueDate
- OCR (reference number)
- Bankgiro
- Plusgiro
- Amount
- supplier_org_number (Swedish organization number)
- customer_number
- payment_line (machine-readable payment code)
""",
version="1.0.0",
lifespan=lifespan,
)
# Add CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Mount static files for results
config.storage.result_dir.mkdir(parents=True, exist_ok=True)
app.mount(
"/static/results",
StaticFiles(directory=str(config.storage.result_dir)),
name="results",
)
# Include public API routes
inference_router = create_inference_router(inference_service, config.storage)
app.include_router(inference_router)
async_router = create_async_router(config.storage.allowed_extensions)
app.include_router(async_router, prefix="/api/v1")
labeling_router = create_labeling_router(inference_service, config.storage)
app.include_router(labeling_router)
# Include admin API routes
auth_router = create_auth_router()
app.include_router(auth_router, prefix="/api/v1")
documents_router = create_documents_router(config.storage)
app.include_router(documents_router, prefix="/api/v1")
locks_router = create_locks_router()
app.include_router(locks_router, prefix="/api/v1")
annotation_router = create_annotation_router()
app.include_router(annotation_router, prefix="/api/v1")
training_router = create_training_router()
app.include_router(training_router, prefix="/api/v1")
augmentation_router = create_augmentation_router()
app.include_router(augmentation_router, prefix="/api/v1/admin")
# Include batch upload routes
app.include_router(batch_upload_router)
# Store inference service in app state for access by routes
app.state.inference_service = inference_service
# Root endpoint - serve HTML UI
@app.get("/", response_class=HTMLResponse)
async def root() -> str:
"""Serve the web UI."""
return get_html_ui()
return app
def get_html_ui() -> str:
"""Generate HTML UI for the web application."""
return """
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Invoice Field Extraction</title>
<style>
* {
box-sizing: border-box;
margin: 0;
padding: 0;
}
body {
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, sans-serif;
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
min-height: 100vh;
padding: 20px;
}
.container {
max-width: 1200px;
margin: 0 auto;
}
header {
text-align: center;
color: white;
margin-bottom: 30px;
}
header h1 {
font-size: 2.5rem;
margin-bottom: 10px;
}
header p {
opacity: 0.9;
font-size: 1.1rem;
}
.main-content {
display: flex;
flex-direction: column;
gap: 20px;
}
.card {
background: white;
border-radius: 16px;
padding: 24px;
box-shadow: 0 10px 40px rgba(0,0,0,0.2);
}
.card h2 {
color: #333;
margin-bottom: 20px;
font-size: 1.3rem;
display: flex;
align-items: center;
gap: 10px;
}
.upload-card {
display: flex;
align-items: center;
gap: 20px;
flex-wrap: wrap;
}
.upload-card h2 {
margin-bottom: 0;
white-space: nowrap;
}
.upload-area {
border: 2px dashed #ddd;
border-radius: 10px;
padding: 15px 25px;
text-align: center;
cursor: pointer;
transition: all 0.3s;
background: #fafafa;
flex: 1;
min-width: 200px;
}
.upload-area:hover, .upload-area.dragover {
border-color: #667eea;
background: #f0f4ff;
}
.upload-area.has-file {
border-color: #10b981;
background: #ecfdf5;
}
.upload-icon {
font-size: 24px;
display: inline;
margin-right: 8px;
}
.upload-area p {
color: #666;
margin: 0;
display: inline;
}
.upload-area small {
color: #999;
display: block;
margin-top: 5px;
}
#file-input {
display: none;
}
.file-name {
margin-top: 15px;
padding: 10px 15px;
background: #e0f2fe;
border-radius: 8px;
color: #0369a1;
font-weight: 500;
}
.btn {
display: inline-block;
padding: 12px 24px;
border: none;
border-radius: 10px;
font-size: 0.9rem;
font-weight: 600;
cursor: pointer;
transition: all 0.3s;
text-transform: uppercase;
letter-spacing: 0.5px;
}
.btn-primary {
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
color: white;
}
.btn-primary:hover:not(:disabled) {
transform: translateY(-2px);
box-shadow: 0 5px 20px rgba(102, 126, 234, 0.4);
}
.btn-primary:disabled {
opacity: 0.6;
cursor: not-allowed;
}
.loading {
display: none;
align-items: center;
gap: 10px;
}
.loading.active {
display: flex;
}
.spinner {
width: 24px;
height: 24px;
border: 3px solid #f3f3f3;
border-top: 3px solid #667eea;
border-radius: 50%;
animation: spin 1s linear infinite;
}
@keyframes spin {
0% { transform: rotate(0deg); }
100% { transform: rotate(360deg); }
}
.results {
display: none;
}
.results.active {
display: block;
}
.result-header {
display: flex;
justify-content: space-between;
align-items: center;
margin-bottom: 20px;
padding-bottom: 15px;
border-bottom: 2px solid #eee;
}
.result-status {
padding: 6px 12px;
border-radius: 20px;
font-size: 0.85rem;
font-weight: 600;
}
.result-status.success {
background: #dcfce7;
color: #166534;
}
.result-status.partial {
background: #fef3c7;
color: #92400e;
}
.result-status.error {
background: #fee2e2;
color: #991b1b;
}
.fields-grid {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
gap: 12px;
}
.field-item {
padding: 12px;
background: #f8fafc;
border-radius: 10px;
border-left: 4px solid #667eea;
}
.field-item label {
display: block;
font-size: 0.75rem;
color: #64748b;
text-transform: uppercase;
letter-spacing: 0.5px;
margin-bottom: 4px;
}
.field-item .value {
font-size: 1.1rem;
font-weight: 600;
color: #1e293b;
}
.field-item .confidence {
font-size: 0.75rem;
color: #10b981;
margin-top: 2px;
}
.visualization {
margin-top: 20px;
}
.visualization img {
width: 100%;
border-radius: 12px;
box-shadow: 0 4px 20px rgba(0,0,0,0.1);
}
.processing-time {
text-align: center;
color: #64748b;
font-size: 0.9rem;
margin-top: 15px;
}
.cross-validation {
background: #f8fafc;
border: 1px solid #e2e8f0;
border-radius: 10px;
padding: 15px;
margin-top: 20px;
}
.cross-validation h3 {
margin: 0 0 10px 0;
color: #334155;
font-size: 1rem;
}
.cv-status {
font-weight: 600;
padding: 8px 12px;
border-radius: 6px;
margin-bottom: 10px;
display: inline-block;
}
.cv-status.valid {
background: #dcfce7;
color: #166534;
}
.cv-status.invalid {
background: #fef3c7;
color: #92400e;
}
.cv-details {
display: flex;
flex-wrap: wrap;
gap: 8px;
margin-top: 10px;
}
.cv-item {
background: white;
border: 1px solid #e2e8f0;
border-radius: 6px;
padding: 6px 12px;
font-size: 0.85rem;
display: flex;
align-items: center;
gap: 6px;
}
.cv-item.match {
border-color: #86efac;
background: #f0fdf4;
}
.cv-item.mismatch {
border-color: #fca5a5;
background: #fef2f2;
}
.cv-icon {
font-weight: bold;
}
.cv-item.match .cv-icon {
color: #16a34a;
}
.cv-item.mismatch .cv-icon {
color: #dc2626;
}
.cv-summary {
margin-top: 10px;
font-size: 0.8rem;
color: #64748b;
}
.error-message {
background: #fee2e2;
color: #991b1b;
padding: 15px;
border-radius: 10px;
margin-top: 15px;
}
footer {
text-align: center;
color: white;
opacity: 0.8;
margin-top: 30px;
font-size: 0.9rem;
}
</style>
</head>
<body>
<div class="container">
<header>
<h1>📄 Invoice Field Extraction</h1>
<p>Upload a Swedish invoice (PDF or image) to extract fields automatically</p>
</header>
<div class="main-content">
<!-- Upload Section - Compact -->
<div class="card upload-card">
<h2>📤 Upload</h2>
<div class="upload-area" id="upload-area">
<span class="upload-icon">📁</span>
<p>Drag & drop or <strong>click to browse</strong></p>
<small>PDF, PNG, JPG (max 50MB)</small>
<input type="file" id="file-input" accept=".pdf,.png,.jpg,.jpeg">
</div>
<div class="file-name" id="file-name" style="display: none;"></div>
<button class="btn btn-primary" id="submit-btn" disabled>
🚀 Extract
</button>
<div class="loading" id="loading">
<div class="spinner"></div>
<p>Processing...</p>
</div>
</div>
<!-- Results Section - Full Width -->
<div class="card">
<h2>📊 Extraction Results</h2>
<div id="placeholder" style="text-align: center; padding: 30px; color: #999;">
<div style="font-size: 48px; margin-bottom: 10px;">🔍</div>
<p>Upload a document to see extraction results</p>
</div>
<div class="results" id="results">
<div class="result-header">
<span>Document: <strong id="doc-id"></strong></span>
<span class="result-status" id="result-status"></span>
</div>
<div class="fields-grid" id="fields-grid"></div>
<div class="processing-time" id="processing-time"></div>
<div class="cross-validation" id="cross-validation" style="display: none;"></div>
<div class="error-message" id="error-message" style="display: none;"></div>
<div class="visualization" id="visualization" style="display: none;">
<h3 style="margin-bottom: 10px; color: #333;">🎯 Detection Visualization</h3>
<img id="viz-image" src="" alt="Detection visualization">
</div>
</div>
</div>
</div>
<footer>
<p>Powered by ColaCoder</p>
</footer>
</div>
<script>
const uploadArea = document.getElementById('upload-area');
const fileInput = document.getElementById('file-input');
const fileName = document.getElementById('file-name');
const submitBtn = document.getElementById('submit-btn');
const loading = document.getElementById('loading');
const placeholder = document.getElementById('placeholder');
const results = document.getElementById('results');
let selectedFile = null;
// Drag and drop handlers
uploadArea.addEventListener('click', () => fileInput.click());
uploadArea.addEventListener('dragover', (e) => {
e.preventDefault();
uploadArea.classList.add('dragover');
});
uploadArea.addEventListener('dragleave', () => {
uploadArea.classList.remove('dragover');
});
uploadArea.addEventListener('drop', (e) => {
e.preventDefault();
uploadArea.classList.remove('dragover');
const files = e.dataTransfer.files;
if (files.length > 0) {
handleFile(files[0]);
}
});
fileInput.addEventListener('change', (e) => {
if (e.target.files.length > 0) {
handleFile(e.target.files[0]);
}
});
function handleFile(file) {
const validTypes = ['.pdf', '.png', '.jpg', '.jpeg'];
const ext = '.' + file.name.split('.').pop().toLowerCase();
if (!validTypes.includes(ext)) {
alert('Please upload a PDF, PNG, or JPG file.');
return;
}
selectedFile = file;
fileName.textContent = `📎 ${file.name}`;
fileName.style.display = 'block';
uploadArea.classList.add('has-file');
submitBtn.disabled = false;
}
submitBtn.addEventListener('click', async () => {
if (!selectedFile) return;
// Show loading
submitBtn.disabled = true;
loading.classList.add('active');
placeholder.style.display = 'none';
results.classList.remove('active');
try {
const formData = new FormData();
formData.append('file', selectedFile);
const response = await fetch('/api/v1/infer', {
method: 'POST',
body: formData,
});
const data = await response.json();
if (!response.ok) {
throw new Error(data.detail || 'Processing failed');
}
displayResults(data);
} catch (error) {
console.error('Error:', error);
document.getElementById('error-message').textContent = error.message;
document.getElementById('error-message').style.display = 'block';
results.classList.add('active');
} finally {
loading.classList.remove('active');
submitBtn.disabled = false;
}
});
function displayResults(data) {
const result = data.result;
// Document ID
document.getElementById('doc-id').textContent = result.document_id;
// Status
const statusEl = document.getElementById('result-status');
statusEl.textContent = result.success ? 'Success' : 'Partial';
statusEl.className = 'result-status ' + (result.success ? 'success' : 'partial');
// Fields
const fieldsGrid = document.getElementById('fields-grid');
fieldsGrid.innerHTML = '';
const fieldOrder = [
'InvoiceNumber', 'InvoiceDate', 'InvoiceDueDate', 'OCR',
'Amount', 'Bankgiro', 'Plusgiro',
'supplier_org_number', 'customer_number', 'payment_line'
];
fieldOrder.forEach(field => {
const value = result.fields[field];
const confidence = result.confidence[field];
if (value !== null && value !== undefined) {
const fieldDiv = document.createElement('div');
fieldDiv.className = 'field-item';
fieldDiv.innerHTML = `
<label>${formatFieldName(field)}</label>
<div class="value">${value}</div>
${confidence ? `<div class="confidence">✓ ${(confidence * 100).toFixed(1)}% confident</div>` : ''}
`;
fieldsGrid.appendChild(fieldDiv);
}
});
// Processing time
document.getElementById('processing-time').textContent =
`⏱️ Processed in ${result.processing_time_ms.toFixed(0)}ms`;
// Cross-validation results
const cvDiv = document.getElementById('cross-validation');
if (result.cross_validation) {
const cv = result.cross_validation;
let cvHtml = '<h3>🔍 Cross-Validation (Payment Line)</h3>';
cvHtml += `<div class="cv-status ${cv.is_valid ? 'valid' : 'invalid'}">`;
cvHtml += cv.is_valid ? '✅ Valid' : '⚠️ Mismatch Detected';
cvHtml += '</div>';
cvHtml += '<div class="cv-details">';
if (cv.payment_line_ocr) {
const matchIcon = cv.ocr_match === true ? '' : (cv.ocr_match === false ? '' : '');
cvHtml += `<div class="cv-item ${cv.ocr_match === true ? 'match' : (cv.ocr_match === false ? 'mismatch' : '')}">`;
cvHtml += `<span class="cv-icon">${matchIcon}</span> OCR: ${cv.payment_line_ocr}</div>`;
}
if (cv.payment_line_amount) {
const matchIcon = cv.amount_match === true ? '' : (cv.amount_match === false ? '' : '');
cvHtml += `<div class="cv-item ${cv.amount_match === true ? 'match' : (cv.amount_match === false ? 'mismatch' : '')}">`;
cvHtml += `<span class="cv-icon">${matchIcon}</span> Amount: ${cv.payment_line_amount}</div>`;
}
if (cv.payment_line_account) {
const accountType = cv.payment_line_account_type === 'bankgiro' ? 'Bankgiro' : 'Plusgiro';
const matchField = cv.payment_line_account_type === 'bankgiro' ? cv.bankgiro_match : cv.plusgiro_match;
const matchIcon = matchField === true ? '' : (matchField === false ? '' : '');
cvHtml += `<div class="cv-item ${matchField === true ? 'match' : (matchField === false ? 'mismatch' : '')}">`;
cvHtml += `<span class="cv-icon">${matchIcon}</span> ${accountType}: ${cv.payment_line_account}</div>`;
}
cvHtml += '</div>';
if (cv.details && cv.details.length > 0) {
cvHtml += '<div class="cv-summary">' + cv.details[cv.details.length - 1] + '</div>';
}
cvDiv.innerHTML = cvHtml;
cvDiv.style.display = 'block';
} else {
cvDiv.style.display = 'none';
}
// Visualization
if (result.visualization_url) {
const vizDiv = document.getElementById('visualization');
const vizImg = document.getElementById('viz-image');
vizImg.src = result.visualization_url;
vizDiv.style.display = 'block';
}
// Errors
if (result.errors && result.errors.length > 0) {
document.getElementById('error-message').textContent = result.errors.join(', ');
document.getElementById('error-message').style.display = 'block';
} else {
document.getElementById('error-message').style.display = 'none';
}
results.classList.add('active');
}
function formatFieldName(name) {
const nameMap = {
'InvoiceNumber': 'Invoice Number',
'InvoiceDate': 'Invoice Date',
'InvoiceDueDate': 'Due Date',
'OCR': 'OCR Reference',
'Amount': 'Amount',
'Bankgiro': 'Bankgiro',
'Plusgiro': 'Plusgiro',
'supplier_org_number': 'Supplier Org Number',
'customer_number': 'Customer Number',
'payment_line': 'Payment Line'
};
return nameMap[name] || name.replace(/([A-Z])/g, ' $1').replace(/_/g, ' ').trim();
}
</script>
</body>
</html>
"""