File size: 92,067 Bytes
dc4e6da | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 | """
FastAPI application for DocGenie document generation.
FULLY INTEGRATED PIPELINE (All 19 Stages):
✅ Stage 1-2: Core Pipeline (Stages 01-06)
1. Seed Selection: Download and encode seed images
2. LLM Prompting: Call Claude API (batched client support)
3. Response Processing: Extract and validate HTML/GT
4. PDF Rendering: Generate PDFs with geometry extraction
5. BBox Extraction: Extract bounding boxes from PDFs
6. Validation: Verify geometries and bboxes
✅ Stage 3: Feature Synthesis (Stages 07-13)
7. Extract handwriting definitions from HTML
8. Extract visual element definitions from HTML
9. Generate handwriting images (WordStylist diffusion model)
10. Create visual elements (stamps, barcodes, logos)
11. Render second-pass PDF with features
12. Insert handwriting images into PDF
13. Insert visual elements into PDF
✅ Stage 4: Image Finalization & OCR (Stages 14-15)
14. Render final PDF to high-quality image (pdf2image)
15. Perform OCR on final image (Microsoft Document Intelligence)
✅ Stage 5: Dataset Packaging (Stages 16-19)
16. Normalize bounding boxes to [0,1] scale
17. Verify and prepare ground truth annotations
18. Generate document analysis and statistics
19. Create debug visualization overlays
See API_PIPELINE_STATUS.md for detailed integration status.
"""
import os
import sys
import pathlib
import tempfile
import uuid
import json
import zipfile
import asyncio
import shutil
import warnings
from typing import List, Optional
from contextlib import asynccontextmanager
# Suppress resource_tracker warnings in development mode (with uvicorn --reload)
# These warnings are harmless - they occur because the reloader creates child processes
# that share semaphores. The lifespan handler below ensures proper cleanup.
warnings.filterwarnings("ignore", category=UserWarning, module="resource_tracker")
# Load environment variables from .env file if it exists
from dotenv import load_dotenv
load_dotenv()
# Add parent directory to path for docgenie imports
sys.path.insert(0, str(pathlib.Path(__file__).parent.parent))
from fastapi import FastAPI, HTTPException, status, BackgroundTasks
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse, StreamingResponse
import uvicorn
import io
from docgenie import ENV
from .schemas import (
GenerateDocumentRequest,
GenerateDocumentResponse,
DocumentResult,
BoundingBox,
HealthResponse,
DatasetExportInfo
)
from .utils import (
download_image_to_base64,
build_prompt,
call_claude_api_direct,
extract_html_documents_from_response,
extract_ground_truth,
extract_css_from_html,
render_html_to_pdf,
extract_bboxes_from_rendered_pdf,
pdf_to_base64,
validate_html_structure,
validate_pdf,
validate_bboxes,
process_stage3_complete,
process_stage4_ocr,
process_stage5_complete,
retry_on_network_error
)
from .config import settings
# Lifespan context manager for proper startup/shutdown
@asynccontextmanager
async def lifespan(app: FastAPI):
"""Handle application lifecycle - startup and shutdown."""
# Startup
print("🚀 DocGenie API starting up...")
yield
# Shutdown - give pending tasks time to complete
print("🛑 DocGenie API shutting down gracefully...")
await asyncio.sleep(0.5) # Allow pending async operations to complete
print("✓ Shutdown complete")
# Initialize FastAPI app with lifespan
app = FastAPI(
title="DocGenie API",
description="API for generating synthetic documents using LLMs",
version="1.0.0",
docs_url="/docs",
lifespan=lifespan
)
# Add CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=settings.get_cors_origins(), # Configure in .env
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
@app.get("/", response_model=HealthResponse)
async def root():
"""Root endpoint - health check."""
return HealthResponse(status="healthy", version="1.0.0")
@app.get("/health", response_model=HealthResponse)
async def health_check():
"""Health check endpoint."""
return HealthResponse(status="healthy", version="1.0.0")
@app.post("/generate", response_model=GenerateDocumentResponse)
async def generate_documents(request: GenerateDocumentRequest):
"""
Generate synthetic documents from seed images.
Pipeline:
1. Download seed images from URLs
2. Convert images to base64
3. Build prompt with user parameters
4. Call Claude API
5. Extract HTML documents from response
6. Extract ground truth and CSS
7. Render HTML to PDF
8. Extract bounding boxes
9. Return results
"""
try:
# Step 1 & 2: Download and convert seed images to base64
print(f"Downloading {len(request.seed_images)} seed images...")
seed_images_base64 = []
# Parse request_id and handle assets
user_id_from_input, request_id = parse_request_id(request.request_id)
user_id = user_id_from_input
# Sanitize Google Drive tokens (ignore Swagger UI defaults)
if request.google_drive_token == "string":
request.google_drive_token = None
if request.google_drive_refresh_token == "string":
request.google_drive_refresh_token = None
assets_temp_dir = None
# Download assets if possible
try:
from .supabase_client import supabase_client
# Try to get user_id from database if not in request_id
effective_user_id = user_id
if not effective_user_id:
effective_user_id = supabase_client.get_user_id_from_request(request_id)
if effective_user_id and request_id:
assets_path = f"{effective_user_id}/{request_id}/assets"
files = supabase_client.list_files("doc_storage", assets_path)
asset_files = [f for f in files if f.get('id') is not None]
if asset_files:
assets_temp_dir = pathlib.Path(tempfile.mkdtemp())
print(f"Found {len(asset_files)} assets in storage, downloading...")
for file_info in asset_files:
file_name = file_info['name']
try:
file_content = supabase_client.download_file("doc_storage", f"{assets_path}/{file_name}")
with open(assets_temp_dir / file_name, 'wb') as f:
f.write(file_content)
except Exception as e:
print(f" ⚠ Failed to download asset {file_name}: {e}")
except Exception as e:
print(f" ⚠ Asset check failed: {e}")
for url in request.seed_images:
try:
img_b64 = await download_image_to_base64(str(url))
seed_images_base64.append(img_b64)
except Exception as e:
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
detail=f"Failed to download image from {url}: {str(e)}"
)
print(f"Successfully downloaded {len(seed_images_base64)} images")
# Step 3: Build prompt
prompt_template_path = ENV.PROMPT_TEMPLATES_DIR / "ClaudeRefined12" / "seed-based-json.txt"
if not prompt_template_path.exists():
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Prompt template not found at {prompt_template_path}"
)
prompt = build_prompt(
language=request.prompt_params.language,
doc_type=request.prompt_params.doc_type,
gt_type=request.prompt_params.gt_type,
gt_format=request.prompt_params.gt_format,
num_solutions=request.prompt_params.num_solutions,
num_seed_images=len(seed_images_base64),
prompt_template_path=prompt_template_path,
enable_visual_elements=request.prompt_params.enable_visual_elements,
visual_element_types=request.prompt_params.visual_element_types
)
print("Prompt built successfully")
# Step 4: Call Claude API (using settings)
if not settings.ANTHROPIC_API_KEY:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="ANTHROPIC_API_KEY environment variable not set"
)
print(f"Calling Claude API with model {settings.CLAUDE_MODEL}...")
llm_data = await call_claude_api_direct(
prompt=prompt,
seed_images_base64=seed_images_base64,
api_key=settings.ANTHROPIC_API_KEY,
model=settings.CLAUDE_MODEL
)
llm_response = llm_data["response"]
usage_data = llm_data["usage"]
# Calculate cost for the entire request (direct call = no batch discount)
from .utils import calculate_message_cost
total_request_cost = calculate_message_cost(
model=usage_data["model"],
input_tokens=usage_data["input_tokens"],
output_tokens=usage_data["output_tokens"],
cache_creation_input_tokens=usage_data["cache_creation_tokens"],
cache_read_input_tokens=usage_data["cache_read_tokens"]
)
print(f"Received LLM response ({len(llm_response)} chars, Cost: ${total_request_cost:.4f})")
# Step 5: Extract HTML documents
html_documents = extract_html_documents_from_response(llm_response)
if not html_documents:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="No valid HTML documents found in LLM response"
)
print(f"Extracted {len(html_documents)} HTML documents")
# Process each document
results = []
# Create temporary directory for PDFs
with tempfile.TemporaryDirectory() as tmp_dir:
tmp_path = pathlib.Path(tmp_dir)
for idx, html in enumerate(html_documents):
try:
doc_id = f"{uuid.uuid4()}_{idx}"
print(f"Processing document {idx + 1}/{len(html_documents)} (ID: {doc_id})")
# Initialize original_pdf_path (will be set after rendering)
original_pdf_path = None
# Validate HTML structure (pipeline_03 validation)
is_valid, error_msg = validate_html_structure(html)
if not is_valid:
print(f" ⚠ HTML validation failed: {error_msg}")
continue
# Step 6: Extract ground truth and CSS (pipeline_03)
gt, html_clean = extract_ground_truth(html)
css, _ = extract_css_from_html(html_clean)
# DEBUG: Check if LLM generated handwriting classes
print(f"\n 🔍 DEBUG - Handwriting Detection:")
print(f" - Contains 'handwritten' class: {'handwritten' in html_clean}")
# Check for author classes (format: author1, author2, etc. - NO DASH)
import re
author_pattern = re.compile(r'\bauthor\d+\b')
author_matches = author_pattern.findall(html_clean)
if 'handwritten' in html_clean:
# Count occurrences
hw_count = html_clean.count('handwritten')
print(f" - 'handwritten' occurrences: {hw_count}")
print(f" - Author classes found: {len(author_matches)}")
if author_matches:
unique_authors = set(author_matches)
print(f" - Unique author IDs: {sorted(unique_authors)}")
else:
print(f" - ⚠️ NO author classes found (expected format: author1, author2, etc.)")
# Show first match context
idx = html_clean.find('handwritten')
context_start = max(0, idx - 50)
context_end = min(len(html_clean), idx + 150)
print(f" - First match context: ...{html_clean[context_start:context_end]}...")
else:
print(f" - ⚠️ NO handwriting classes found in LLM output!")
# Show sample of HTML to see structure
print(f" - HTML sample (first 500 chars): {html_clean[:500]}")
print(f" 🔍 DEBUG - Visual Elements Detection:")
print(f" - Contains 'data-placeholder': {'data-placeholder' in html_clean}")
if 'data-placeholder' in html_clean:
ve_count = html_clean.count('data-placeholder')
print(f" - 'data-placeholder' occurrences: {ve_count}")
print()
# Step 7: Render to PDF (pipeline_04) and extract geometries
pdf_path = tmp_path / f"{doc_id}.pdf"
pdf_path, width_mm, height_mm, geometries = await render_html_to_pdf(
html=html_clean,
output_pdf_path=pdf_path
)
print(f" ✓ Rendered PDF: {width_mm:.1f}mm x {height_mm:.1f}mm")
# Validate PDF (pipeline_06 style validation)
is_valid, error_msg = validate_pdf(pdf_path)
if not is_valid:
print(f" ⚠ PDF validation failed: {error_msg}")
continue
# Step 8: Extract bounding boxes (pipeline_05)
bboxes_raw = extract_bboxes_from_rendered_pdf(pdf_path)
# Validate bboxes (pipeline_06 style validation)
is_valid, error_msg = validate_bboxes(bboxes_raw, min_bbox_count=1)
if not is_valid:
print(f" ⚠ BBox validation failed: {error_msg}")
# Continue anyway with empty bboxes for API response
bboxes = [BoundingBox(**bbox) for bbox in bboxes_raw]
print(f" ✓ Extracted {len(bboxes)} bounding boxes")
# Step 9: Convert PDF to base64
pdf_b64 = pdf_to_base64(pdf_path)
# Step 10: Process Stage 3 (Handwriting & Visual Elements) if enabled
final_image_b64 = None
handwriting_regions = []
visual_elements = []
handwriting_images = {}
visual_element_images = {}
ocr_results = None
modified_pdf_path = None
# Track original PDF path before modification
original_pdf_path = pdf_path
if request.prompt_params.enable_handwriting or request.prompt_params.enable_visual_elements:
print(f" 🎨 Processing Stages 07-13 (Handwriting & Visual Elements)...")
try:
final_image_b64, handwriting_regions, visual_elements, handwriting_images, visual_element_images, pdf_with_handwriting_path, pdf_final_path = await process_stage3_complete(
pdf_path=pdf_path,
geometries=geometries,
ground_truth=gt,
bboxes_raw=bboxes_raw,
page_width_mm=width_mm,
page_height_mm=height_mm,
enable_handwriting=request.prompt_params.enable_handwriting,
handwriting_ratio=request.prompt_params.handwriting_ratio,
handwriting_apply_ink_filter=request.prompt_params.handwriting_apply_ink_filter,
handwriting_num_inference_steps=request.prompt_params.handwriting_num_inference_steps,
handwriting_writer_ids=request.prompt_params.handwriting_writer_ids,
enable_visual_elements=request.prompt_params.enable_visual_elements,
visual_element_types=request.prompt_params.visual_element_types,
seed=request.prompt_params.seed,
assets_dir=assets_temp_dir,
barcode_number=request.prompt_params.barcode_number
)
# Use final PDF if modifications were made
if pdf_final_path and pdf_final_path.exists():
pdf_path = pdf_final_path
pdf_b64 = pdf_to_base64(pdf_path)
elif pdf_with_handwriting_path and pdf_with_handwriting_path.exists():
pdf_path = pdf_with_handwriting_path
pdf_b64 = pdf_to_base64(pdf_path)
print(f" ✓ Stages 07-13 complete: {len(handwriting_regions)} handwriting regions, {len(visual_elements)} visual elements")
print(f" - Individual tokens: {len(handwriting_images)} handwriting, {len(visual_element_images)} visual elements")
except Exception as e:
print(f" ⚠ Stages 07-13 processing failed: {str(e)}")
# Continue with original PDF if Stage 3 fails
# Step 11: Process Stages 14-15 (Image Finalization & OCR) if needed
if request.prompt_params.enable_ocr or (final_image_b64 is None and (request.prompt_params.enable_handwriting or request.prompt_params.enable_visual_elements)):
print(f" 📄 Processing Stages 14-15 (Image Finalization & OCR)...")
try:
stage4_image, ocr_results = await process_stage4_ocr(
pdf_path=pdf_path,
enable_ocr=request.prompt_params.enable_ocr,
dpi=settings.OCR_DPI
)
# Use Stage 4 image if Stage 3 didn't generate one
if final_image_b64 is None and stage4_image:
final_image_b64 = stage4_image
if ocr_results:
print(f" ✓ Stages 14-15 complete: Image rendered, OCR: {len(ocr_results.get('words', []))} words")
else:
print(f" ✓ Stage 14 complete: Image rendered")
except Exception as e:
print(f" ⚠ Stages 14-15 processing failed: {str(e)}")
# Continue without Stage 4
# Step 12: Process Stages 16-18 (Dataset Packaging) if needed
stage5_results = {}
if any([
request.prompt_params.enable_bbox_normalization,
request.prompt_params.enable_gt_verification,
request.prompt_params.enable_analysis,
request.prompt_params.enable_debug_visualization
]):
print(f" 📦 Processing Stages 16-18 (Dataset Packaging)...")
try:
stage5_results = await process_stage5_complete(
document_id=doc_id,
pdf_path=str(pdf_path),
image_base64=final_image_b64,
ocr_results=ocr_results,
ground_truth=gt,
bboxes_raw=bbox_pdf_word,
has_handwriting=request.prompt_params.enable_handwriting,
has_visual_elements=request.prompt_params.enable_visual_elements,
layout_elements=visual_elements,
handwriting_regions=handwriting_regions,
page_width_mm=width_mm,
page_height_mm=height_mm,
enable_bbox_normalization=request.prompt_params.enable_bbox_normalization,
enable_gt_verification=request.prompt_params.enable_gt_verification,
enable_analysis=request.prompt_params.enable_analysis,
enable_debug_visualization=request.prompt_params.enable_debug_visualization
)
print(f" ✓ Stages 16-18 complete")
except Exception as e:
print(f" ⚠ Stages 16-18 processing failed: {str(e)}")
# Continue without Stage 5
# Step 13: Export to dataset format if requested
dataset_export_info = None
if request.prompt_params.enable_dataset_export:
print(f" 📦 Exporting dataset format ({request.prompt_params.dataset_export_format})...")
try:
from .utils import export_to_msgpack
# Only msgpack format is currently supported
if request.prompt_params.dataset_export_format.lower() == "msgpack":
# Prepare data for export
export_words = []
export_word_bboxes = []
export_segment_bboxes = []
# Get normalized bboxes if available (Stage 5), otherwise use raw OCR
if stage5_results.get('normalized_bboxes_word'):
# Use Stage 5 normalized bboxes
for bbox_entry in stage5_results['normalized_bboxes_word']:
export_words.append(bbox_entry.get('text', ''))
bbox = bbox_entry.get('bbox', [0, 0, 1, 1])
export_word_bboxes.append(bbox)
if stage5_results.get('normalized_bboxes_segment'):
for bbox_entry in stage5_results['normalized_bboxes_segment']:
bbox = bbox_entry.get('bbox', [0, 0, 1, 1])
export_segment_bboxes.append(bbox)
elif ocr_results:
# Fallback: normalize OCR bboxes manually
from pdf2image import convert_from_path
images = convert_from_path(pdf_path, dpi=settings.OCR_DPI)
img_width, img_height = images[0].size if images else (1000, 1000)
for word in ocr_results.get('words', []):
export_words.append(word.get('text', ''))
bbox = word.get('bbox', {'x0': 0, 'y0': 0, 'x1': 1, 'y1': 1})
# Normalize to [0,1]
norm_bbox = [
bbox['x0'] / img_width,
bbox['y0'] / img_height,
bbox['x1'] / img_width,
bbox['y1'] / img_height
]
export_word_bboxes.append(norm_bbox)
export_segment_bboxes.append(norm_bbox) # Use words as segments
else:
print(f" ⚠ No OCR data available for msgpack export")
if export_words and export_word_bboxes:
# Create msgpack file in temp directory
msgpack_path = pathlib.Path(tempfile.gettempdir()) / f"{doc_id}_dataset.msgpack"
await export_to_msgpack(
document_id=doc_id,
image_path=None,
image_base64=final_image_b64,
words=export_words,
word_bboxes=export_word_bboxes,
segment_bboxes=export_segment_bboxes if export_segment_bboxes else export_word_bboxes,
ground_truth=gt,
output_path=msgpack_path,
image_width=None,
image_height=None
)
# Read msgpack file as base64 for response
if msgpack_path.exists():
with open(msgpack_path, 'rb') as f:
msgpack_bytes = f.read()
msgpack_b64 = base64.b64encode(msgpack_bytes).decode('utf-8')
dataset_export_info = DatasetExportInfo(
format="msgpack",
num_samples=1,
output_path=str(msgpack_path),
msgpack_base64=msgpack_b64 if len(msgpack_bytes) < 10_000_000 else None, # Only include if < 10MB
metadata={
"document_id": doc_id,
"num_words": len(export_words),
"has_ground_truth": gt is not None,
"has_ocr": ocr_results is not None
}
)
print(f" ✓ Dataset exported to msgpack: {msgpack_path}")
else:
print(f" ⚠ Export format '{request.prompt_params.dataset_export_format}' not supported. Only 'msgpack' is available.")
except Exception as e:
print(f" ⚠ Dataset export failed: {str(e)}")
import traceback
traceback.print_exc()
# Prepare individual tokens based on output_detail level
handwriting_token_images_response = None
visual_element_images_response = None
token_mapping_response = None
output_detail = request.prompt_params.output_detail
if output_detail in ["dataset", "complete"]:
# Include individual token images for dataset/complete levels
from .utils import create_token_mapping_json
if handwriting_images or visual_element_images:
handwriting_token_images_response = handwriting_images
visual_element_images_response = visual_element_images
token_mapping_response = create_token_mapping_json(
handwriting_regions,
handwriting_images,
visual_elements,
visual_element_images
)
print(f" 📦 Output detail '{output_detail}': Including {len(handwriting_images)} handwriting tokens, {len(visual_element_images)} visual elements")
# Calculate per-document cost share
num_docs = len(html_documents)
doc_cost_info = None
if num_docs > 0:
doc_cost_info = CostInfo(
input_tokens=usage_data["input_tokens"] // num_docs,
output_tokens=usage_data["output_tokens"] // num_docs,
cache_creation_tokens=usage_data["cache_creation_tokens"] // num_docs,
cache_read_tokens=usage_data["cache_read_tokens"] // num_docs,
cost_usd=total_request_cost / num_docs,
batch_discount_applied=False
)
# Create result
result = DocumentResult(
document_id=doc_id,
html=html_clean,
css=css,
ground_truth=gt,
pdf_base64=pdf_b64,
bboxes=bboxes,
page_width_mm=width_mm,
page_height_mm=height_mm,
image_base64=final_image_b64,
handwriting_regions=handwriting_regions,
visual_elements=visual_elements,
handwriting_token_images=handwriting_token_images_response,
visual_element_images=visual_element_images_response,
token_mapping=token_mapping_response,
ocr_results=ocr_results,
# Stage 5 results
normalized_bboxes_word=stage5_results.get('normalized_bboxes_word'),
normalized_bboxes_segment=stage5_results.get('normalized_bboxes_segment'),
gt_verification=stage5_results.get('gt_verification'),
analysis_stats=stage5_results.get('analysis_stats'),
debug_visualization=stage5_results.get('debug_visualization'),
dataset_export=dataset_export_info,
cost_info=doc_cost_info
)
results.append(result)
except Exception as e:
print(f"Error processing document {idx}: {str(e)}")
# Continue with other documents
continue
if not results:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="Failed to process any documents"
)
print(f"Successfully generated {len(results)} documents")
# Add warning message for large responses
output_detail = request.prompt_params.output_detail
message = f"Successfully generated {len(results)} documents"
if output_detail == "complete":
message += " ⚠️ WARNING: 'complete' output detail level may result in 50+ MB response"
elif output_detail == "dataset":
message += " (dataset mode: includes individual tokens)"
return GenerateDocumentResponse(
success=True,
message=message,
documents=results,
total_documents=len(results),
total_cost=CostInfo(
input_tokens=usage_data["input_tokens"],
output_tokens=usage_data["output_tokens"],
cache_creation_tokens=usage_data["cache_creation_tokens"],
cache_read_tokens=usage_data["cache_read_tokens"],
cost_usd=total_request_cost,
batch_discount_applied=False
)
)
except HTTPException:
raise
except Exception as e:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Internal server error: {str(e)}"
)
finally:
# Clean up assets directory if it exists
if 'assets_temp_dir' in locals() and assets_temp_dir and assets_temp_dir.exists():
try:
shutil.rmtree(assets_temp_dir, ignore_errors=True)
print(f"✓ Cleaned up assets directory {assets_temp_dir}")
except:
pass
def parse_request_id(input_str: str) -> tuple:
"""Extract user_id and request_id from input string (format: user_id/request_id or just request_id)."""
if "/" in input_str:
parts = input_str.split("/", 1)
return parts[0], parts[1]
return None, input_str
@app.post("/generate/pdf")
async def generate_document_pdf(
request: GenerateDocumentRequest,
background_tasks: BackgroundTasks
):
"""
Generate documents and return them as downloadable PDF files (FAST DEMO ENDPOINT).
This endpoint generates documents and returns a ZIP file immediately (20-60 seconds).
**Workflow:**
1. Frontend creates document_requests entry in Supabase with status="pending"
2. Frontend sends request_id to this endpoint along with tokens and seed images
3. API fetches existing request, validates, and starts generation
4. API updates status through: processing → generating → completed/failed
5. ZIP file is returned immediately
6. If google_drive_token provided: ZIP is uploaded to GDrive in background
**Request Parameters:**
- request_id: UUID of existing document_requests entry (required)
- seed_images: List of image URLs to use as document backgrounds (required)
- google_drive_token: OAuth token for GDrive upload (optional, enables backup)
- google_drive_refresh_token: Refresh token for GDrive (optional)
- prompt_params: Document generation parameters
**Use Cases:**
- Quick demos and testing (with direct Claude API)
- Production with progress tracking and GDrive backup
**For batch processing:** Use `/generate/async` (50% cheaper, 5-30 minutes)
"""
# Get request_id from database
user_id_from_input, request_id = parse_request_id(request.request_id)
user_id = user_id_from_input
supabase_enabled = False
gdrive_enabled = False
try:
# Import supabase_client
from .supabase_client import supabase_client
# Get existing request from database
existing_request = supabase_client.get_request(request_id)
if not existing_request:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"Request {request_id} not found in database"
)
# Use user_id from input if available, otherwise from database
if not user_id:
user_id = existing_request["user_id"]
supabase_enabled = True
print(f"[Request {request_id}] Processing request for user {user_id}")
print(f"[Request {request_id}] Current status: {existing_request['status']}")
# Validate Google Drive token if provided
if request.google_drive_token:
gdrive_enabled = True
# Download assets from Supabase storage if they exist
assets_temp_dir = None
if supabase_enabled:
try:
assets_path = f"{user_id}/{request_id}/assets"
files = supabase_client.list_files("doc_storage", assets_path)
# Filter out directories
asset_files = [f for f in files if f.get('id') is not None]
if asset_files:
assets_temp_dir = pathlib.Path(tempfile.mkdtemp())
print(f"[Request {request_id}] Found {len(asset_files)} assets in storage, downloading...")
for file_info in asset_files:
file_name = file_info['name']
try:
file_content = supabase_client.download_file("doc_storage", f"{assets_path}/{file_name}")
with open(assets_temp_dir / file_name, 'wb') as f:
f.write(file_content)
print(f" ✓ Downloaded {file_name}")
except Exception as download_err:
print(f" ⚠ Failed to download {file_name}: {download_err}")
else:
print(f"[Request {request_id}] No assets found in {assets_path}")
except Exception as e:
print(f"[Request {request_id}] ⚠ Asset check/download failed: {e}")
print(f"[Request {request_id}] GDrive integration enabled")
# Log analytics
try:
supabase_client.log_analytics_event(
user_id=user_id,
event_type="document_generation_started_sync",
entity_id=request_id
)
except Exception as e:
print(f"[Request {request_id}] Warning: Analytics logging failed: {e}")
except HTTPException:
raise
except Exception as e:
print(f"Error: Failed to fetch request from database: {e}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Failed to fetch request: {str(e)}"
)
# Update status: Downloading seed images
if supabase_enabled:
try:
supabase_client.update_request_status(request_id, "downloading")
print(f"[Request {request_id}] Status: downloading (fetching seed images)")
except Exception as e:
print(f"Warning: Status update failed: {e}")
try:
# Step 1 & 2: Download and convert seed images to base64
print(f"Downloading {len(request.seed_images)} seed images...")
seed_images_base64 = []
for url in request.seed_images:
try:
img_b64 = await download_image_to_base64(str(url))
seed_images_base64.append(img_b64)
except Exception as e:
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
detail=f"Failed to download image from {url}: {str(e)}"
)
print(f"Successfully downloaded {len(seed_images_base64)} images")
# Step 3: Build prompt
prompt_template_path = ENV.PROMPT_TEMPLATES_DIR / "ClaudeRefined12" / "seed-based-json.txt"
if not prompt_template_path.exists():
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Prompt template not found at {prompt_template_path}"
)
prompt = build_prompt(
language=request.prompt_params.language,
doc_type=request.prompt_params.doc_type,
gt_type=request.prompt_params.gt_type,
gt_format=request.prompt_params.gt_format,
num_solutions=request.prompt_params.num_solutions,
num_seed_images=len(seed_images_base64),
prompt_template_path=prompt_template_path,
enable_visual_elements=request.prompt_params.enable_visual_elements,
visual_element_types=request.prompt_params.visual_element_types
)
print("Prompt built successfully")
# Extract output_detail early to use in ZIP packaging later
output_detail = request.prompt_params.output_detail
# Create temporary directory and exporter BEFORE LLM call (so we can track costs)
with tempfile.TemporaryDirectory() as tmp_dir:
tmp_path = pathlib.Path(tmp_dir)
# Initialize DatasetExporter for organized structure
from .dataset_exporter import DatasetExporter
exporter = DatasetExporter(tmp_path, dataset_name="docgenie_documents")
# Update status: Generating (calling LLM)
if supabase_enabled:
try:
supabase_client.update_request_status(request_id, "generating")
print(f"[Request {request_id}] Status: generating (calling LLM)")
except Exception as e:
print(f"Warning: Status update failed: {e}")
# Step 4: Call Claude API (using settings)
print(f"Calling Claude API with model {settings.CLAUDE_MODEL}...")
llm_data = await call_claude_api_direct(
prompt=prompt,
seed_images_base64=seed_images_base64,
api_key=settings.ANTHROPIC_API_KEY,
model=settings.CLAUDE_MODEL
)
llm_response = llm_data["response"]
usage_data = llm_data["usage"]
# Calculate cost and add to exporter (Research Parity)
from .utils import calculate_message_cost
total_request_cost = calculate_message_cost(
model=usage_data["model"],
input_tokens=usage_data["input_tokens"],
output_tokens=usage_data["output_tokens"],
cache_creation_input_tokens=usage_data["cache_creation_tokens"],
cache_read_input_tokens=usage_data["cache_read_tokens"]
)
exporter.add_cost(
cost_usd=total_request_cost,
input_tokens=usage_data["input_tokens"],
output_tokens=usage_data["output_tokens"],
cache_creation_tokens=usage_data["cache_creation_tokens"],
cache_read_tokens=usage_data["cache_read_tokens"]
)
print(f"Received LLM response ({len(llm_response)} chars, Cost: ${total_request_cost:.4f})")
# Step 5: Extract HTML documents
html_documents = extract_html_documents_from_response(llm_response)
if not html_documents:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="No valid HTML documents found in LLM response"
)
print(f"Extracted {len(html_documents)} HTML documents")
pdf_files = []
metadata = []
for idx, html in enumerate(html_documents):
try:
doc_id = f"document_{idx + 1}"
print(f"Processing document {idx + 1}/{len(html_documents)} (ID: {doc_id})")
# Initialize original_pdf_path (will be set after rendering)
original_pdf_path = None
# Extract ground truth
gt, html_clean = extract_ground_truth(html)
# DEBUG: Check if LLM generated handwriting classes
print(f"\n 🔍 DEBUG - Handwriting Detection:")
print(f" - Contains 'handwritten' class: {'handwritten' in html_clean}")
# Check for author classes (format: author1, author2, etc. - NO DASH)
import re
author_pattern = re.compile(r'\bauthor\d+\b')
author_matches = author_pattern.findall(html_clean)
if 'handwritten' in html_clean:
# Count occurrences
hw_count = html_clean.count('handwritten')
print(f" - 'handwritten' occurrences: {hw_count}")
print(f" - Author classes found: {len(author_matches)}")
if author_matches:
unique_authors = set(author_matches)
print(f" - Unique author IDs: {sorted(unique_authors)}")
else:
print(f" - ⚠️ NO author classes found (expected format: author1, author2, etc.)")
# Show first match context
idx = html_clean.find('handwritten')
context_start = max(0, idx - 50)
context_end = min(len(html_clean), idx + 150)
print(f" - First match context: ...{html_clean[context_start:context_end]}...")
else:
print(f" - ⚠️ NO handwriting classes found in LLM output!")
# Show sample of HTML to see structure
print(f" - HTML sample (first 500 chars): {html_clean[:500]}")
print(f" 🔍 DEBUG - Visual Elements Detection:")
print(f" - Contains 'data-placeholder': {'data-placeholder' in html_clean}")
if 'data-placeholder' in html_clean:
ve_count = html_clean.count('data-placeholder')
print(f" - 'data-placeholder' occurrences: {ve_count}")
print()
# Render to PDF and extract geometries
pdf_path = tmp_path / f"{doc_id}.pdf"
pdf_path, width_mm, height_mm, geometries = await render_html_to_pdf(
html=html_clean,
output_pdf_path=pdf_path
)
print(f" - Rendered PDF: {width_mm:.1f}mm x {height_mm:.1f}mm")
# Extract bounding boxes
bboxes_raw = extract_bboxes_from_rendered_pdf(pdf_path)
print(f" - Extracted {len(bboxes_raw)} bounding boxes")
# Extract CSS for Stage 3
css, _ = extract_css_from_html(html_clean)
# Step: Process Stage 3 (Handwriting & Visual Elements) if enabled
final_image_b64 = None
handwriting_regions = []
visual_elements = []
handwriting_images = {}
visual_element_images = {}
ocr_results = None
pdf_with_handwriting_path = None
pdf_final_path = None
# Track original PDF path before modification
original_pdf_path = pdf_path
if request.prompt_params.enable_handwriting or request.prompt_params.enable_visual_elements:
print(f" 🎨 Processing Stages 07-13 (Handwriting & Visual Elements)...")
try:
final_image_b64, handwriting_regions, visual_elements, handwriting_images, visual_element_images, pdf_with_handwriting_path, pdf_final_path = await process_stage3_complete(
pdf_path=pdf_path,
geometries=geometries,
ground_truth=gt,
bboxes_raw=bboxes_raw,
page_width_mm=width_mm,
page_height_mm=height_mm,
enable_handwriting=request.prompt_params.enable_handwriting,
handwriting_ratio=request.prompt_params.handwriting_ratio,
handwriting_apply_ink_filter=request.prompt_params.handwriting_apply_ink_filter,
handwriting_enable_enhancements=request.prompt_params.handwriting_enable_enhancements,
handwriting_num_inference_steps=request.prompt_params.handwriting_num_inference_steps,
handwriting_writer_ids=request.prompt_params.handwriting_writer_ids,
enable_visual_elements=request.prompt_params.enable_visual_elements,
visual_element_types=request.prompt_params.visual_element_types,
seed=request.prompt_params.seed,
assets_dir=assets_temp_dir,
barcode_number=request.prompt_params.barcode_number
)
# Use final PDF if modifications were made
if pdf_final_path and pdf_final_path.exists():
pdf_path = pdf_final_path
elif pdf_with_handwriting_path and pdf_with_handwriting_path.exists():
pdf_path = pdf_with_handwriting_path
print(f" ✓ Stages 07-13 complete: {len(handwriting_regions)} handwriting regions, {len(visual_elements)} visual elements")
print(f" - Individual tokens: {len(handwriting_images)} handwriting, {len(visual_element_images)} visual elements")
except Exception as e:
print(f" ⚠ Stages 07-13 processing failed: {str(e)}")
# Continue with original PDF if Stage 3 fails
# Step: Process Stages 14-15 (Image Finalization & OCR) if needed
if request.prompt_params.enable_ocr:
print(f" 📄 Processing Stages 14-15 (OCR)...")
try:
stage4_image, ocr_results = await process_stage4_ocr(
pdf_path=pdf_path,
enable_ocr=True,
dpi=settings.OCR_DPI
)
if ocr_results:
print(f" ✓ Stages 14-15 complete: OCR: {len(ocr_results.get('words', []))} words")
except Exception as e:
print(f" ⚠ Stages 14-15 processing failed: {str(e)}")
# Continue without Stage 4
# Step: Extract bbox_pdf (word + char) from original PDF (Stage 1 parity)
from .utils import extract_all_bboxes_from_pdf
print(f" 📦 Extracting Stage 1 bboxes from PDF for normalization...")
try:
bboxes_pdf = extract_all_bboxes_from_pdf(original_pdf_path if original_pdf_path else pdf_path)
bbox_pdf_word = bboxes_pdf.get('word', [])
bbox_pdf_char = bboxes_pdf.get('char', [])
except Exception as e:
print(f" ⚠ Stage 1 bbox extraction failed: {e}")
bbox_pdf_word = bboxes_raw
bbox_pdf_char = []
# Step: Process Stages 16-19 (Dataset Packaging) if needed
stage5_results = {}
if any([
request.prompt_params.enable_bbox_normalization,
request.prompt_params.enable_gt_verification,
request.prompt_params.enable_analysis,
request.prompt_params.enable_debug_visualization
]):
print(f" 📦 Processing Stages 16-18 (Dataset Packaging)...")
try:
stage5_results = await process_stage5_complete(
document_id=doc_id,
pdf_path=str(pdf_path),
image_base64=final_image_b64,
ocr_results=ocr_results,
ground_truth=gt,
bboxes_raw=bbox_pdf_word,
has_handwriting=request.prompt_params.enable_handwriting,
has_visual_elements=request.prompt_params.enable_visual_elements,
layout_elements=visual_elements,
handwriting_regions=handwriting_regions,
page_width_mm=width_mm,
page_height_mm=height_mm,
enable_bbox_normalization=request.prompt_params.enable_bbox_normalization,
enable_gt_verification=request.prompt_params.enable_gt_verification,
enable_analysis=request.prompt_params.enable_analysis,
enable_debug_visualization=request.prompt_params.enable_debug_visualization
)
print(f" ✓ Stages 16-19 complete")
except Exception as e:
print(f" ⚠ Stages 16-19 processing failed: {str(e)}")
import traceback
traceback.print_exc()
# Track PDFs for metadata
if original_pdf_path and pdf_path != original_pdf_path:
pdf_files.append(original_pdf_path)
pdf_files.append(pdf_path)
else:
pdf_files.append(pdf_path)
# Extract raw_annotations (layout boxes before normalization)
raw_annotations = None
if geometries:
print(f" 📦 Extracting raw_annotations from geometries...")
try:
raw_annotations = extract_raw_annotations_from_geometries(geometries)
print(f" ✓ Extracted {len(raw_annotations)} layout annotations")
except Exception as e:
print(f" ⚠ raw_annotations extraction failed: {e}")
# Decode final image to bytes
final_image_bytes = None
if final_image_b64:
import base64
final_image_bytes = base64.b64decode(final_image_b64)
# Decode debug visualization
debug_viz_bytes = None
if stage5_results.get('debug_visualization'):
debug_viz_dict = stage5_results['debug_visualization']
if debug_viz_dict and 'bbox_overlay_base64' in debug_viz_dict:
debug_viz_b64 = debug_viz_dict['bbox_overlay_base64']
debug_viz_bytes = base64.b64decode(debug_viz_b64)
# Prepare token mapping if tokens exist
token_mapping_data = None
if output_detail in ["dataset", "complete"]:
if handwriting_images or visual_element_images:
from .utils import create_token_mapping_json
token_mapping_data = create_token_mapping_json(
handwriting_regions,
handwriting_images,
visual_elements,
visual_element_images
)
print(f" 📦 Output detail '{output_detail}': Prepared {len(handwriting_images)} handwriting tokens, {len(visual_element_images)} visual elements")
# Extract bbox_final_word and bbox_final_segment (from OCR or PDF)
bbox_final_word = None
bbox_final_segment = None
if ocr_results and ocr_results.get('words'):
# Use OCR results as final bboxes
bbox_final_word = ocr_results.get('words', [])
bbox_final_segment = ocr_results.get('lines', [])
else:
# Fallback to PDF bboxes if no OCR
bbox_final_word = bbox_pdf_word
bbox_final_segment = [] # No line-level data without OCR
# Read PDF bytes for exporter (capture all stages)
pdf_initial_bytes = original_pdf_path.read_bytes()
pdf_with_handwriting_bytes = pdf_with_handwriting_path.read_bytes() if pdf_with_handwriting_path and pdf_with_handwriting_path.exists() else None
pdf_final_bytes = pdf_final_path.read_bytes() if pdf_final_path and pdf_final_path.exists() else None
# For visual elements only (no handwriting), pdf_final_path actually points to the VE-only PDF
pdf_with_visual_elements_bytes = None
if pdf_final_bytes and not pdf_with_handwriting_bytes:
# Only visual elements were added, not handwriting
pdf_with_visual_elements_bytes = pdf_final_bytes
pdf_final_bytes = None # No "final" with both modifications
# Add document to exporter
print(f" 📦 Adding document to dataset exporter...")
# Pick the best available bboxes for normalized export
norm_word = stage5_results.get('normalized_bboxes_word') or stage5_results.get('normalized_bboxes_word_raw')
norm_segment = stage5_results.get('normalized_bboxes_segment')
exporter.add_document(
document_id=doc_id,
html=html_clean,
css=css,
pdf_initial=pdf_initial_bytes,
pdf_with_handwriting=pdf_with_handwriting_bytes,
pdf_with_visual_elements=pdf_with_visual_elements_bytes,
pdf_final=pdf_final_bytes,
final_image=final_image_bytes,
ground_truth=gt,
raw_annotations=raw_annotations,
bboxes_pdf_word=bbox_pdf_word,
bboxes_pdf_char=bbox_pdf_char,
bboxes_final_word=bbox_final_word,
bboxes_final_segment=bbox_final_segment,
bboxes_normalized_word=norm_word,
bboxes_normalized_segment=norm_segment,
gt_verification=stage5_results.get('gt_verification'),
token_mapping=token_mapping_data,
handwriting_regions=handwriting_regions,
handwriting_images=handwriting_images,
visual_elements=visual_elements,
visual_element_images=visual_element_images,
layout_elements=stage5_results.get('normalized_layout_elements') or visual_elements,
geometries=geometries,
ocr_results=ocr_results,
analysis_stats=stage5_results.get('analysis_stats'),
debug_visualization=debug_viz_bytes
)
print(f" ✓ Document {doc_id} added to dataset")
# Store metadata
metadata.append({
"document_id": doc_id,
"filename": f"{doc_id}.pdf",
"bboxes": bboxes_raw,
"ground_truth": gt,
"geometries": geometries,
"page_width_mm": width_mm,
"page_height_mm": height_mm,
"handwriting_regions": handwriting_regions,
"visual_elements": visual_elements,
"has_stage3_image": final_image_b64 is not None,
"ocr_results": ocr_results,
# Stage 5 results
"normalized_bboxes_word": norm_word,
"normalized_bboxes_segment": norm_segment,
"gt_verification": stage5_results.get('gt_verification'),
"analysis_stats": stage5_results.get('analysis_stats'),
"debug_visualization_available": stage5_results.get('debug_visualization') is not None
})
except Exception as e:
print(f"Error processing document {idx}: {str(e)}")
# Continue with other documents
continue
if not pdf_files:
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail="Failed to process any documents"
)
print(f"Successfully generated {len(pdf_files)} documents")
# Finalize dataset export (writes metadata.json and README.md)
print(f"📦 Finalizing dataset export...")
exporter.finalize(
request_id=request_id if request_id else "unnamed",
user_id=user_id,
prompt_params=request.prompt_params.dict(),
api_mode="sync"
)
print(f"✓ Dataset structure finalized at {exporter.base_path}")
# Update status: Zipping
if supabase_enabled:
try:
supabase_client.update_request_status(request_id, "zipping")
print(f"[Request {request_id}] Status: zipping (creating ZIP archive)")
except Exception as e:
print(f"Warning: Status update failed: {e}")
# Create ZIP from organized dataset
print(f"📦 Creating ZIP archive from dataset...")
zip_buffer = io.BytesIO()
with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
# Add all files from exporter.base_path
for file_path in exporter.base_path.rglob('*'):
if file_path.is_file():
arcname = file_path.relative_to(exporter.base_path.parent)
zip_file.write(file_path, arcname)
zip_buffer.seek(0)
zip_size_mb = len(zip_buffer.getvalue()) / (1024 * 1024)
print(f"✓ ZIP created: {zip_size_mb:.2f} MB")
# Update status: Completed
if supabase_enabled and request_id:
try:
from .supabase_client import supabase_client
supabase_client.update_request_status(request_id, "completed")
print(f"[Request {request_id}] Status: completed")
except Exception as e:
print(f"[Request {request_id}] ⚠ Supabase update failed: {e}")
# Save ZIP to temporary file for background upload
temp_zip_path = pathlib.Path(tempfile.gettempdir()) / f"docgenie_{request_id}.zip"
temp_zip_path.write_bytes(zip_buffer.getvalue())
# Schedule background task: Upload to Google Drive
has_gdrive_token = request.google_drive_token and request.google_drive_token != "string"
if gdrive_enabled and request_id and has_gdrive_token:
# Update status: Uploading
try:
supabase_client.update_request_status(request_id, "uploading")
print(f"[Request {request_id}] Status: uploading (uploading to Google Drive)")
except Exception as e:
print(f"Warning: Status update failed: {e}")
print(f"[Request {request_id}] Scheduling GDrive upload in background...")
background_tasks.add_task(
upload_zip_to_gdrive_background,
request_id=request_id,
zip_path=temp_zip_path,
access_token=request.google_drive_token,
refresh_token=request.google_drive_refresh_token,
num_documents=len(pdf_files)
)
# Save files for Supabase background upload
if supabase_enabled:
import shutil
supabase_temp_dir = pathlib.Path(tempfile.gettempdir()) / f"docgenie_supabase_{request_id}"
if supabase_temp_dir.exists():
shutil.rmtree(supabase_temp_dir, ignore_errors=True)
# Copy exporter base_path to persistent temp dir
shutil.copytree(exporter.base_path, supabase_temp_dir)
print(f"[Request {request_id}] Scheduling Supabase document upload in background...")
background_tasks.add_task(
upload_documents_to_supabase_background,
request_id=request_id,
user_id=str(user_id),
temp_dir=str(supabase_temp_dir),
num_documents=len(exporter.documents),
model_version=settings.LLM_MODEL,
zip_path=str(temp_zip_path) if 'temp_zip_path' in locals() else None
)
# Prepare response headers with tracking info
headers = {
"Content-Disposition": f"attachment; filename=docgenie_documents_{uuid.uuid4().hex[:8]}.zip"
}
# Add tracking header if Supabase enabled
if supabase_enabled and request_id:
headers["X-Request-ID"] = request_id
headers["X-Status-URL"] = f"/jobs/{request_id}/status"
print(f"[Request {request_id}] Returning ZIP with tracking headers")
return StreamingResponse(
zip_buffer,
media_type="application/zip",
headers=headers
)
except HTTPException as e:
# Update status to failed if Supabase enabled
if supabase_enabled and request_id:
try:
from .supabase_client import supabase_client
supabase_client.update_request_status(request_id, "failed", error_message=str(e.detail))
print(f"[Request {request_id}] Status: failed - {e.detail}")
except Exception as update_error:
print(f"Warning: Status update failed: {update_error}")
raise
except Exception as e:
# Update status to failed if Supabase enabled
if supabase_enabled and request_id:
try:
from .supabase_client import supabase_client
supabase_client.update_request_status(request_id, "failed", error_message=str(e))
print(f"[Request {request_id}] Status: failed - {str(e)}")
except Exception as sup_err:
print(f"[Request {request_id}] ⚠ Supabase update failed: {sup_err}")
print(f"Unexpected error: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Internal server error: {str(e)}"
)
# ==================== Background Task Functions ====================
def upload_documents_to_supabase_background(
request_id: str,
user_id: str,
temp_dir: str,
num_documents: int,
model_version: str,
zip_path: Optional[str] = None
):
"""
Background task to upload individual documents to Supabase Storage.
"""
import shutil
import pathlib
import traceback
try:
print(f"[Background Task {request_id}] Starting Supabase individual document upload...")
from .supabase_client import supabase_client
# Clean up any old generated documents for this request (parity with async worker)
try:
supabase_client.delete_generated_documents(request_id)
print(f"[Background Task {request_id}] ✓ Cleaned up old document records")
except Exception as cleanup_err:
print(f"[Background Task {request_id}] ⚠ Cleanup of old records failed: {cleanup_err}")
base_path = pathlib.Path(temp_dir)
# Upload zip if provided
zip_url = None
if zip_path and pathlib.Path(zip_path).exists():
zip_file = pathlib.Path(zip_path)
zip_storage_path = f"{user_id}/{request_id}/generated/docgenie_{request_id}.zip"
retry_on_network_error(lambda: supabase_client.upload_to_storage("doc_storage", zip_storage_path, zip_file.read_bytes(), "application/zip"))
zip_url = supabase_client.get_public_url("doc_storage", zip_storage_path)
print(f"[Background Task {request_id}] ✓ Uploaded ZIP to Supabase: {zip_url}")
for idx in range(num_documents):
doc_id = f"document_{idx + 1}"
# Paths to upload
doc_storage_path = f"{user_id}/{request_id}/generated/{idx}_doc.pdf"
gt_storage_path = f"{user_id}/{request_id}/generated/{idx}_gt.json"
html_storage_path = f"{user_id}/{request_id}/generated/{idx}_src.html"
bbox_storage_path = f"{user_id}/{request_id}/generated/{idx}_bbox.json"
# Local paths
local_pdf = base_path / "pdf" / "pdf_final" / f"{doc_id}.pdf"
if not local_pdf.exists():
local_pdf = base_path / "pdf" / "pdf_initial" / f"{doc_id}.pdf"
local_gt = base_path / "annotations" / "gt" / f"{doc_id}.json"
local_html = base_path / "html" / f"{doc_id}.html"
local_bbox = base_path / "bbox" / "bbox_final" / "word" / f"{doc_id}.json"
try:
# Step 10: Upload Individual Files and Create Record
# We wrap each upload in a retry, and use a nested try-except for the whole group
try:
# Upload PDF (Critical)
pdf_url = None
if local_pdf.exists():
retry_on_network_error(lambda: supabase_client.upload_to_storage("doc_storage", doc_storage_path, local_pdf.read_bytes(), "application/pdf"))
pdf_url = supabase_client.get_public_url("doc_storage", doc_storage_path)
# Upload Ground Truth (Important)
if local_gt.exists():
retry_on_network_error(lambda: supabase_client.upload_to_storage("doc_storage", gt_storage_path, local_gt.read_bytes(), "application/json"))
# Upload HTML Source (Optional)
if local_html.exists():
retry_on_network_error(lambda: supabase_client.upload_to_storage("doc_storage", html_storage_path, local_html.read_bytes(), "text/html"))
# Upload Bounding Boxes (Optional)
if local_bbox.exists():
retry_on_network_error(lambda: supabase_client.upload_to_storage("doc_storage", bbox_storage_path, local_bbox.read_bytes(), "application/json"))
except Exception as upload_err:
print(f"[Background Task {request_id}] ⚠ Some file uploads failed for document {idx}: {upload_err}")
# Create record in database (Always try this)
try:
retry_on_network_error(lambda: supabase_client.create_generated_document(
request_id=request_id,
file_url=pdf_url,
model_version=model_version,
doc_index=idx,
doc_storage_path=doc_storage_path if local_pdf.exists() else None,
gt_storage_path=gt_storage_path if local_gt.exists() else None,
html_storage_path=html_storage_path if local_html.exists() else None,
bbox_storage_path=bbox_storage_path if local_bbox.exists() else None
))
print(f"[Background Task {request_id}] ✓ Uploaded and tracked document {idx}")
except Exception as db_err:
print(f"[Background Task {request_id}] ❌ Failed to create DB record for document {idx}: {db_err}")
except Exception as doc_err:
print(f"[Background Task {request_id}] ⚠ Unexpected error for document {idx}: {doc_err}")
# Final Step: Update the request record with the ZIP URL and final status
# This happens AFTER the loop finishes all document uploads
if zip_url:
try:
supabase_client.update_request_status(
request_id=request_id,
status="completed",
zip_url=zip_url
)
print(f"[Background Task {request_id}] ✓ Updated request {request_id} with zip_url")
except Exception as status_err:
print(f"[Background Task {request_id}] ❌ Failed to update request status: {status_err}")
except Exception as e:
print(f"[Background Task {request_id}] ⚠ Supabase upload failed: {str(e)}")
traceback.print_exc()
# Update status to error if we failed catastrophically
try:
from .supabase_client import supabase_client
supabase_client.update_request_status(
request_id=request_id,
status="error",
error_message=f"Supabase upload failed: {str(e)}"
)
except:
pass
finally:
try:
# Clean up temporary directory
shutil.rmtree(temp_dir, ignore_errors=True)
print(f"[Background Task {request_id}] ✓ Cleaned up temporary directory {temp_dir}")
except Exception as e:
print(f"[Background Task {request_id}] ⚠ Failed to clean up temp dir: {e}")
def upload_zip_to_gdrive_background(
request_id: str,
zip_path: pathlib.Path,
access_token: str,
refresh_token: Optional[str],
num_documents: int
):
"""
Background task to upload ZIP file to Google Drive.
Args:
request_id: Supabase request ID
zip_path: Path to temporary ZIP file
access_token: Google Drive OAuth access token
refresh_token: Google Drive refresh token (optional)
num_documents: Number of documents in ZIP
"""
try:
print(f"[Background Task {request_id}] Starting GDrive upload...")
from .google_drive import GoogleDriveClient
from .supabase_client import supabase_client
# Upload to Google Drive
client = GoogleDriveClient(
access_token=access_token,
refresh_token=refresh_token
)
gdrive_url = None
gdrive_failed = False
gdrive_skipped = False
# Determine if we should attempt GDrive upload
has_gdrive_token = access_token and access_token != "string"
if not has_gdrive_token:
print(f"[Background Task {request_id}] No GDrive token provided. Skipping.")
gdrive_skipped = True
else:
try:
filename = f"docgenie_{request_id}.zip"
gdrive_url = client.upload_file(
file_path=zip_path,
filename=filename,
folder_name=settings.GOOGLE_DRIVE_FOLDER_NAME,
mime_type="application/zip"
)
print(f"[Background Task {request_id}] ✓ Uploaded to GDrive: {gdrive_url}")
except Exception as drive_err:
print(f"[Background Task {request_id}] ⚠ Google Drive upload failed: {drive_err}")
gdrive_failed = True
# Update status to completed and save zip_url
if gdrive_skipped:
final_status = "completed_no_gdrive"
elif gdrive_failed:
final_status = "completed_gdrive_failed"
else:
final_status = "completed"
supabase_client.update_request_status(
request_id=request_id,
status=final_status,
zip_url=zip_url
)
print(f"[Background Task {request_id}] ✓ Status updated to {final_status} with zip_url")
# Clean up temporary file
zip_path.unlink(missing_ok=True)
print(f"[Background Task {request_id}] ✓ Cleaned up temp file")
except Exception as e:
print(f"[Background Task {request_id}] ✗ GDrive upload failed: {str(e)}")
import traceback
traceback.print_exc()
# Update status to completed_gdrive_failed since token was provided
try:
from .supabase_client import supabase_client
supabase_client.update_request_status(request_id, "completed_gdrive_failed")
print(f"[Background Task {request_id}] Status updated to completed_gdrive_failed")
except Exception as status_err:
print(f"[Background Task {request_id}] Failed to update status: {status_err}")
# Clean up temp file even if upload failed
try:
zip_path.unlink(missing_ok=True)
except Exception:
pass
# ==================== New Async Endpoints (Batched API) ====================
from redis import Redis
from rq import Queue
from rq.job import Job
from .supabase_client import supabase_client
from .worker import process_document_generation_job
# Initialize Redis and RQ
try:
redis_conn = Redis.from_url(settings.REDIS_URL)
job_queue = Queue(settings.RQ_QUEUE_NAME, connection=redis_conn)
print(f"✓ Connected to Redis: [HIDDEN]")
print(f"✓ RQ Queue: {settings.RQ_QUEUE_NAME}")
except Exception as e:
print(f"⚠ Warning: Redis connection failed: {e}")
print(" Async endpoints will not work without Redis")
redis_conn = None
job_queue = None
@app.post("/generate/async")
async def generate_documents_async(request: GenerateDocumentRequest):
"""
Generate synthetic documents asynchronously using batched Claude API.
**Workflow:**
1. Frontend creates document_requests entry in Supabase with status="pending"
2. Frontend sends request_id to this endpoint along with tokens and seed images
3. API fetches existing request, validates, and enqueues background job
4. API returns immediately with job info
5. Background worker processes job and updates status: processing → generating → completed/failed
6. User polls /jobs/{request_id}/status for progress
7. Upon completion, ZIP is automatically uploaded to Google Drive
Uses batched Claude API for 50% cost savings (but takes 5-30 minutes).
Request body:
- request_id: UUID of existing document_requests entry (required)
- seed_images: List[str] (Supabase storage URLs) (required)
- google_drive_token: OAuth token for GDrive upload (optional)
- google_drive_refresh_token: Refresh token for GDrive (optional)
- prompt_params: dict (language, doc_type, num_solutions, etc.)
Returns:
- request_id: UUID to track job
- status: "pending"
- estimated_time_minutes: int
- poll_url: URL to check status
"""
if not job_queue:
raise HTTPException(
status_code=status.HTTP_503_SERVICE_UNAVAILABLE,
detail="Background job queue not available. Redis connection required."
)
# Get request_id from request
user_id_from_input, request_id = parse_request_id(request.request_id)
user_id = user_id_from_input
# Sanitize Google Drive tokens (ignore Swagger UI defaults)
if request.google_drive_token == "string":
request.google_drive_token = None
if request.google_drive_refresh_token == "string":
request.google_drive_refresh_token = None
try:
# Fetch request from Supabase
existing_request = supabase_client.get_request(request_id)
if not existing_request:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"Request {request_id} not found in database"
)
# Use user_id from input if available, otherwise from database
if not user_id:
user_id = existing_request["user_id"]
print(f"[Request {request_id}] Processing async request for user {user_id}")
print(f"[Request {request_id}] Current status: {existing_request['status']}")
# Validate seed images
if not request.seed_images:
raise HTTPException(
status_code=status.HTTP_400_BAD_REQUEST,
detail="At least one seed image is required"
)
# Update status to processing (job is being queued)
supabase_client.update_request_status(request_id, "processing")
print(f"[Request {request_id}] Status: processing (queuing job)")
# Prepare job data
job_data = {
"user_id": user_id,
"google_drive_token": request.google_drive_token,
"google_drive_refresh_token": request.google_drive_refresh_token,
"seed_images": [str(url) for url in request.seed_images],
"prompt_params": request.prompt_params.dict()
}
# Enqueue background job
job = job_queue.enqueue(
process_document_generation_job,
request_id=request_id,
request_data=job_data,
job_timeout='2h', # 2 hours max (batched API can take time)
result_ttl=86400, # Keep result for 24 hours
failure_ttl=86400 # Keep failure info for 24 hours
)
print(f"Enqueued job {job.id} for request {request_id}")
# Estimate time based on num_solutions
num_solutions = request.prompt_params.num_solutions
if num_solutions <= 3:
estimated_time = 10 # ~10 minutes for small batch
elif num_solutions <= 10:
estimated_time = 20 # ~20 minutes for medium batch
else:
estimated_time = 30 + (num_solutions - 10) * 2 # Scale up
# Log analytics
supabase_client.log_analytics_event(
user_id=user_id,
event_type="document_generation_requested",
entity_id=request_id
)
return {
"request_id": request_id,
"status": "pending",
"estimated_time_minutes": estimated_time,
"num_documents": num_solutions,
"poll_url": f"/jobs/{request_id}/status",
"message": f"Job queued successfully. Check status at /jobs/{request_id}/status"
}
except HTTPException as http_exc:
# Update status to failed
try:
supabase_client.update_request_status(request_id, "failed", error_message=str(http_exc.detail))
print(f"[Request {request_id}] Status: failed - {http_exc.detail}")
except Exception as update_error:
print(f"Warning: Status update failed: {update_error}")
raise
except Exception as e:
print(f"Error creating async job: {str(e)}")
import traceback
traceback.print_exc()
# Update status to failed
try:
supabase_client.update_request_status(request_id, "failed", error_message=str(e))
print(f"[Request {request_id}] Status: failed - {str(e)}")
except Exception as update_error:
print(f"Warning: Status update failed: {update_error}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Failed to create job: {str(e)}"
)
finally:
# Clean up assets directory if it exists
if 'assets_temp_dir' in locals() and assets_temp_dir and assets_temp_dir.exists():
try:
shutil.rmtree(assets_temp_dir, ignore_errors=True)
print(f"[Request {request_id}] ✓ Cleaned up assets directory {assets_temp_dir}")
except:
pass
@app.get("/jobs/{request_id}/status")
async def get_job_status(request_id: str):
"""
Get status of a document generation job.
Returns:
- request_id: UUID
- status: pending | processing | generating | completed | failed
- created_at: ISO timestamp
- updated_at: ISO timestamp
- error_message: str (if failed)
- results: dict with download_url (if completed)
"""
try:
# Get request from Supabase
request_data = supabase_client.get_request(request_id)
if not request_data:
raise HTTPException(
status_code=status.HTTP_404_NOT_FOUND,
detail=f"Request {request_id} not found"
)
response = {
"request_id": request_id,
"status": request_data["status"],
"created_at": request_data["created_at"],
"updated_at": request_data["updated_at"],
"num_documents": request_data["metadata"]["prompt_params"]["num_solutions"]
}
# Add error message if failed
if request_data["status"] == "failed":
response["error_message"] = request_data.get("error_message")
# Add result URL if completed
if request_data["status"] == "completed":
# Get generated documents
generated_docs = supabase_client.get_generated_documents(request_id)
if generated_docs:
response["results"] = {
"documents": [
{
"id": doc.get("id"),
"doc_index": doc.get("doc_index"),
"pdf_url": doc.get("file_url"),
"doc_storage_path": doc.get("doc_storage_path"),
"gt_storage_path": doc.get("gt_storage_path"),
"html_storage_path": doc.get("html_storage_path"),
"bbox_storage_path": doc.get("bbox_storage_path")
} for doc in generated_docs if doc.get("doc_index") is not None
],
"zip_filename": f"docgenie_{request_id}.zip"
}
# If there's a zip file (legacy or background GDrive task), add it too
zip_docs = [doc for doc in generated_docs if doc.get("file_type") == "application/zip"]
if zip_docs:
response["results"]["download_url"] = zip_docs[0].get("file_url")
return response
except HTTPException:
raise
except Exception as e:
print(f"Error fetching job status: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Failed to fetch job status: {str(e)}"
)
@app.get("/jobs/user/{user_id}")
async def get_user_jobs(user_id: int, limit: int = 50, offset: int = 0):
"""
Get all jobs for a user.
Query params:
- limit: int (default: 50, max: 100)
- offset: int (default: 0)
Returns:
List of job status objects
"""
try:
# Validate limit
if limit > 100:
limit = 100
# Get user's requests from Supabase
requests = supabase_client.get_user_requests(user_id, limit, offset)
results = []
for request_data in requests:
result = {
"request_id": request_data["id"],
"status": request_data["status"],
"created_at": request_data["created_at"],
"updated_at": request_data["updated_at"],
"num_documents": request_data["metadata"]["prompt_params"]["num_solutions"]
}
if request_data["status"] == "failed":
result["error_message"] = request_data.get("error_message")
if request_data["status"] == "completed":
# Get generated documents
generated_docs = supabase_client.get_generated_documents(request_data["id"])
if generated_docs:
result["download_url"] = generated_docs[0]["file_url"]
results.append(result)
return {
"user_id": user_id,
"jobs": results,
"count": len(results),
"limit": limit,
"offset": offset
}
except Exception as e:
print(f"Error fetching user jobs: {str(e)}")
raise HTTPException(
status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
detail=f"Failed to fetch user jobs: {str(e)}"
)
if __name__ == "__main__":
uvicorn.run(
"main:app",
host=settings.API_HOST,
port=settings.API_PORT,
reload=settings.DEBUG_MODE
)
|