Spaces:
Sleeping
Sleeping
File size: 77,001 Bytes
6c982a7 eda7f22 6c982a7 eda7f22 6c982a7 eda7f22 6c982a7 5aa7215 358773d 5aa7215 6c982a7 eda7f22 6c982a7 eda7f22 6c982a7 eda7f22 6c982a7 eda7f22 358773d 5aa7215 358773d 5aa7215 c99ab26 5aa7215 6c982a7 eda7f22 6c982a7 eda7f22 358773d eda7f22 5aa7215 358773d eda7f22 5aa7215 358773d eda7f22 5aa7215 358773d 6c982a7 911cc77 6c982a7 5aa7215 358773d 6c982a7 5aa7215 358773d eda7f22 358773d 5aa7215 3900fea 5aa7215 3900fea 911cc77 3900fea 911cc77 5aa7215 358773d 3900fea 5aa7215 358773d 911cc77 5aa7215 911cc77 5aa7215 911cc77 5aa7215 911cc77 5aa7215 358773d 5aa7215 911cc77 5aa7215 358773d 5aa7215 358773d 5aa7215 911cc77 5aa7215 911cc77 5aa7215 358773d 911cc77 358773d 5aa7215 911cc77 5aa7215 911cc77 5aa7215 911cc77 5aa7215 911cc77 eda7f22 911cc77 5aa7215 911cc77 5aa7215 911cc77 358773d 5aa7215 911cc77 5aa7215 358773d 5aa7215 eda7f22 6c982a7 5aa7215 3900fea 358773d 6c982a7 5aa7215 358773d 6c982a7 3900fea 5aa7215 3900fea 358773d 6c982a7 3900fea 358773d 3900fea 6c982a7 5aa7215 6c982a7 5aa7215 6c982a7 5aa7215 358773d 5aa7215 358773d 5aa7215 358773d 5aa7215 358773d 5aa7215 6c982a7 5aa7215 358773d 5aa7215 6c982a7 5aa7215 3900fea 358773d 6c982a7 5aa7215 358773d 6c982a7 358773d 5aa7215 6c982a7 eda7f22 5aa7215 358773d eda7f22 5aa7215 358773d eda7f22 5aa7215 358773d eda7f22 358773d eda7f22 5aa7215 358773d eda7f22 5aa7215 358773d 5aa7215 358773d 5aa7215 358773d 5aa7215 358773d 5aa7215 eda7f22 065c4b9 eda7f22 358773d eda7f22 5aa7215 358773d eda7f22 358773d 5aa7215 6c982a7 |
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 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 |
from fastapi import FastAPI, UploadFile, File, Form, HTTPException
from fastapi.responses import JSONResponse
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from typing import Optional, List, Dict
from PIL import Image
import io
import numpy as np
import os
from datetime import datetime
from pymongo import MongoClient
from huggingface_hub import InferenceClient
from embedding_service import JinaClipEmbeddingService
from qdrant_service import QdrantVectorService
from advanced_rag import AdvancedRAG
from cag_service import CAGService
from pdf_parser import PDFIndexer
from multimodal_pdf_parser import MultimodalPDFIndexer
# Initialize FastAPI app
app = FastAPI(
title="Event Social Media Embeddings & ChatbotRAG API",
description="API để embeddings, search và ChatbotRAG với Jina CLIP v2 + Qdrant + MongoDB + LLM",
version="2.0.0"
)
# CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Initialize services
print("Initializing services...")
embedding_service = JinaClipEmbeddingService(model_path="jinaai/jina-clip-v2")
collection_name = os.getenv("COLLECTION_NAME", "event_social_media")
qdrant_service = QdrantVectorService(
collection_name=collection_name,
vector_size=embedding_service.get_embedding_dimension()
)
print(f"✓ Qdrant collection: {collection_name}")
# MongoDB connection
mongodb_uri = os.getenv("MONGODB_URI", "mongodb+srv://truongtn7122003:7KaI9OT5KTUxWjVI@truongtn7122003.xogin4q.mongodb.net/")
mongo_client = MongoClient(mongodb_uri)
db = mongo_client[os.getenv("MONGODB_DB_NAME", "chatbot_rag")]
documents_collection = db["documents"]
chat_history_collection = db["chat_history"]
print("✓ MongoDB connected")
# Hugging Face token
hf_token = os.getenv("HUGGINGFACE_TOKEN")
if hf_token:
print("✓ Hugging Face token configured")
# Initialize Advanced RAG (Best Case 2025)
advanced_rag = AdvancedRAG(
embedding_service=embedding_service,
qdrant_service=qdrant_service
)
print("✓ Advanced RAG pipeline initialized (with Cross-Encoder)")
# Initialize CAG Service (Semantic Cache)
try:
cag_service = CAGService(
embedding_service=embedding_service,
cache_collection="semantic_cache",
vector_size=embedding_service.get_embedding_dimension(),
similarity_threshold=0.9,
ttl_hours=24
)
print("✓ CAG Service initialized (Semantic Caching enabled)")
except Exception as e:
print(f"Warning: CAG Service initialization failed: {e}")
print("Continuing without semantic caching...")
cag_service = None
# Initialize PDF Indexer
pdf_indexer = PDFIndexer(
embedding_service=embedding_service,
qdrant_service=qdrant_service,
documents_collection=documents_collection
)
print("✓ PDF Indexer initialized")
# Initialize Multimodal PDF Indexer (for PDFs with images)
multimodal_pdf_indexer = MultimodalPDFIndexer(
embedding_service=embedding_service,
qdrant_service=qdrant_service,
documents_collection=documents_collection
)
print("✓ Multimodal PDF Indexer initialized")
print("✓ Services initialized successfully")
# Pydantic models for embeddings
class SearchRequest(BaseModel):
text: Optional[str] = None
limit: int = 10
score_threshold: Optional[float] = None
text_weight: float = 0.5
image_weight: float = 0.5
class SearchResponse(BaseModel):
id: str
confidence: float
metadata: dict
class IndexResponse(BaseModel):
success: bool
id: str
message: str
# Pydantic models for ChatbotRAG
class ChatRequest(BaseModel):
message: str
use_rag: bool = True
top_k: int = 3
system_message: Optional[str] = """Bạn là trợ lý AI chuyên biệt cho hệ thống quản lý sự kiện và mạng xã hội.
Vai trò của bạn là trả lời các câu hỏi CHÍNH XÁC dựa trên dữ liệu được cung cấp từ hệ thống.
Quy tắc tuyệt đối:
- CHỈ trả lời câu hỏi liên quan đến: events, social media posts, PDFs đã upload, và dữ liệu trong knowledge base
- KHÔNG trả lời câu hỏi ngoài phạm vi (tin tức, thời tiết, toán học, lập trình, tư vấn cá nhân, v.v.)
- Nếu câu hỏi nằm ngoài phạm vi: BẮT BUỘC trả lời "Chúng tôi không thể trả lời câu hỏi này vì nó nằm ngoài vùng application xử lí."
- Luôn ưu tiên thông tin từ context được cung cấp"""
max_tokens: int = 512
temperature: float = 0.7
top_p: float = 0.95
hf_token: Optional[str] = None
# Advanced RAG options
use_advanced_rag: bool = True
use_query_expansion: bool = True
use_reranking: bool = True
use_compression: bool = True
score_threshold: float = 0.5
# Advanced RAG options
use_advanced_rag: bool = True
use_query_expansion: bool = True
use_reranking: bool = True
use_compression: bool = True
score_threshold: float = 0.5
class ChatResponse(BaseModel):
response: str
context_used: List[Dict]
timestamp: str
rag_stats: Optional[Dict] = None # Stats from advanced RAG pipeline
rag_stats: Optional[Dict] = None # Stats from advanced RAG pipeline
class AddDocumentRequest(BaseModel):
text: str
metadata: Optional[Dict] = None
class AddDocumentResponse(BaseModel):
success: bool
doc_id: str
message: str
class UploadPDFResponse(BaseModel):
success: bool
document_id: str
filename: str
chunks_indexed: int
message: str
class UploadPDFResponse(BaseModel):
success: bool
document_id: str
filename: str
chunks_indexed: int
message: str
@app.get("/")
async def root():
"""Health check endpoint with comprehensive API documentation"""
return {
"status": "running",
"service": "ChatbotRAG API - Advanced RAG with Multimodal Support",
"version": "3.0.0",
"service": "ChatbotRAG API - Advanced RAG with Multimodal Support",
"version": "3.0.0",
"vector_db": "Qdrant",
"document_db": "MongoDB",
"features": {
"multiple_inputs": "Index up to 10 texts + 10 images per request",
"advanced_rag": "Query expansion, reranking, contextual compression",
"pdf_support": "Upload PDFs and chat about their content",
"multimodal_pdf": "PDFs with text and image URLs - perfect for user guides",
"chat_history": "Track conversation history",
"hybrid_search": "Text + image search with Jina CLIP v2"
},
"document_db": "MongoDB",
"features": {
"multiple_inputs": "Index up to 10 texts + 10 images per request",
"advanced_rag": "Query expansion, reranking, contextual compression",
"pdf_support": "Upload PDFs and chat about their content",
"multimodal_pdf": "PDFs with text and image URLs - perfect for user guides",
"chat_history": "Track conversation history",
"hybrid_search": "Text + image search with Jina CLIP v2"
},
"endpoints": {
"indexing": {
"POST /index": {
"description": "Index multiple texts and images (NEW: up to 10 each)",
"content_type": "multipart/form-data",
"body": {
"id": "string (required) - Document ID (primary)",
"texts": "List[string] (optional) - Up to 10 texts",
"images": "List[UploadFile] (optional) - Up to 10 images",
"id_use": "string (optional) - ID của SocialMedia hoặc EventCode",
"id_user": "string (optional) - ID của User"
},
"example": "curl -X POST '/index' -F 'id=doc1' -F 'id_use=social_123' -F 'id_user=user_789' -F 'texts=Text 1' -F 'images=@img1.jpg'",
"indexing": {
"POST /index": {
"description": "Index multiple texts and images (NEW: up to 10 each)",
"content_type": "multipart/form-data",
"body": {
"id": "string (required) - Document ID (primary)",
"texts": "List[string] (optional) - Up to 10 texts",
"images": "List[UploadFile] (optional) - Up to 10 images",
"id_use": "string (optional) - ID của SocialMedia hoặc EventCode",
"id_user": "string (optional) - ID của User"
},
"example": "curl -X POST '/index' -F 'id=doc1' -F 'id_use=social_123' -F 'id_user=user_789' -F 'texts=Text 1' -F 'images=@img1.jpg'",
"response": {
"success": True,
"id": "doc1",
"message": "Indexed successfully with 2 texts and 1 images"
"success": True,
"id": "doc1",
"message": "Indexed successfully with 2 texts and 1 images"
},
"use_cases": {
"social_media_post": {
"id": "post_uuid_123",
"id_use": "social_media_456",
"id_user": "user_789",
"description": "Link post to social media account and user"
},
"event_post": {
"id": "post_uuid_789",
"id_use": "event_code_ABC123",
"id_user": "user_101",
"description": "Link post to event and user"
}
}
"use_cases": {
"social_media_post": {
"id": "post_uuid_123",
"id_use": "social_media_456",
"id_user": "user_789",
"description": "Link post to social media account and user"
},
"event_post": {
"id": "post_uuid_789",
"id_use": "event_code_ABC123",
"id_user": "user_101",
"description": "Link post to event and user"
}
}
},
"POST /documents": {
"description": "Add text document to knowledge base",
"content_type": "application/json",
"body": {
"text": "string (required) - Document content",
"metadata": "object (optional) - Additional metadata"
},
"example": {
"text": "How to create event: Click 'Create Event' button...",
"metadata": {"category": "tutorial", "source": "user_guide"}
}
},
"POST /upload-pdf": {
"description": "Upload PDF file (text only)",
"content_type": "multipart/form-data",
"body": {
"file": "UploadFile (required) - PDF file",
"title": "string (optional) - Document title",
"category": "string (optional) - Category",
"description": "string (optional) - Description"
},
"example": "curl -X POST '/upload-pdf' -F 'file=@guide.pdf' -F 'title=User Guide'"
},
"POST /upload-pdf-multimodal": {
"description": "Upload PDF with text and image URLs (RECOMMENDED for user guides)",
"content_type": "multipart/form-data",
"features": [
"Extracts text from PDF",
"Detects image URLs (http://, https://)",
"Supports markdown: ",
"Supports HTML: <img src='url'>",
"Links images to text chunks",
"Returns images with context in chat"
],
"body": {
"file": "UploadFile (required) - PDF file with image URLs",
"title": "string (optional) - Document title",
"category": "string (optional) - e.g. 'user_guide', 'tutorial'",
"description": "string (optional)"
},
"example": "curl -X POST '/upload-pdf-multimodal' -F 'file=@guide_with_images.pdf' -F 'category=user_guide'",
"description": "Add text document to knowledge base",
"content_type": "application/json",
"body": {
"text": "string (required) - Document content",
"metadata": "object (optional) - Additional metadata"
},
"example": {
"text": "How to create event: Click 'Create Event' button...",
"metadata": {"category": "tutorial", "source": "user_guide"}
}
},
"POST /upload-pdf": {
"description": "Upload PDF file (text only)",
"content_type": "multipart/form-data",
"body": {
"file": "UploadFile (required) - PDF file",
"title": "string (optional) - Document title",
"category": "string (optional) - Category",
"description": "string (optional) - Description"
},
"example": "curl -X POST '/upload-pdf' -F 'file=@guide.pdf' -F 'title=User Guide'"
},
"POST /upload-pdf-multimodal": {
"description": "Upload PDF with text and image URLs (RECOMMENDED for user guides)",
"content_type": "multipart/form-data",
"features": [
"Extracts text from PDF",
"Detects image URLs (http://, https://)",
"Supports markdown: ",
"Supports HTML: <img src='url'>",
"Links images to text chunks",
"Returns images with context in chat"
],
"body": {
"file": "UploadFile (required) - PDF file with image URLs",
"title": "string (optional) - Document title",
"category": "string (optional) - e.g. 'user_guide', 'tutorial'",
"description": "string (optional)"
},
"example": "curl -X POST '/upload-pdf-multimodal' -F 'file=@guide_with_images.pdf' -F 'category=user_guide'",
"response": {
"success": True,
"document_id": "pdf_multimodal_20251029_150000",
"chunks_indexed": 25,
"message": "PDF indexed with 25 chunks and 15 images"
"success": True,
"document_id": "pdf_multimodal_20251029_150000",
"chunks_indexed": 25,
"message": "PDF indexed with 25 chunks and 15 images"
},
"use_case": "Perfect for user guides with screenshots, tutorials with diagrams"
}
},
"search": {
"POST /search": {
"description": "Hybrid search with text and/or image",
"body": {
"text": "string (optional) - Query text",
"image": "UploadFile (optional) - Query image",
"limit": "int (default: 10)",
"score_threshold": "float (optional, 0-1)",
"text_weight": "float (default: 0.5)",
"image_weight": "float (default: 0.5)"
}
},
"POST /search/text": {
"description": "Text-only search",
"body": {"text": "string", "limit": "int", "score_threshold": "float"}
},
"POST /search/image": {
"description": "Image-only search",
"body": {"image": "UploadFile", "limit": "int", "score_threshold": "float"}
"use_case": "Perfect for user guides with screenshots, tutorials with diagrams"
}
},
"search": {
"POST /search": {
"description": "Hybrid search with text and/or image",
"body": {
"text": "string (optional) - Query text",
"image": "UploadFile (optional) - Query image",
"limit": "int (default: 10)",
"score_threshold": "float (optional, 0-1)",
"text_weight": "float (default: 0.5)",
"image_weight": "float (default: 0.5)"
}
},
"POST /search/text": {
"description": "Text-only search",
"body": {"text": "string", "limit": "int", "score_threshold": "float"}
},
"POST /search/image": {
"description": "Image-only search",
"body": {"image": "UploadFile", "limit": "int", "score_threshold": "float"}
},
"POST /rag/search": {
"description": "Search in RAG knowledge base",
"body": {"query": "string", "top_k": "int (default: 5)", "score_threshold": "float (default: 0.5)"}
}
},
"chat": {
"POST /chat": {
"description": "Chat với Advanced RAG (Query expansion + Reranking + Compression)",
"content_type": "application/json",
"body": {
"message": "string (required) - User question",
"use_rag": "bool (default: true) - Enable RAG retrieval",
"use_advanced_rag": "bool (default: true) - Use advanced RAG pipeline (RECOMMENDED)",
"use_query_expansion": "bool (default: true) - Expand query with variations",
"use_reranking": "bool (default: true) - Rerank results for accuracy",
"use_compression": "bool (default: true) - Compress context to relevant parts",
"top_k": "int (default: 3) - Number of documents to retrieve",
"score_threshold": "float (default: 0.5) - Min relevance score (0-1)",
"max_tokens": "int (default: 512) - Max response tokens",
"temperature": "float (default: 0.7) - Creativity (0-1)",
"hf_token": "string (optional) - Hugging Face token"
},
"response": {
"response": "string - AI answer",
"context_used": "array - Retrieved documents with metadata",
"timestamp": "string",
"rag_stats": "object - RAG pipeline statistics (query variants, retrieval counts)"
},
"example_advanced": {
"message": "Làm sao để upload PDF có hình ảnh?",
"use_advanced_rag": True,
"use_reranking": True,
"top_k": 5,
"score_threshold": 0.5
"description": "Search in RAG knowledge base",
"body": {"query": "string", "top_k": "int (default: 5)", "score_threshold": "float (default: 0.5)"}
}
},
"chat": {
"POST /chat": {
"description": "Chat với Advanced RAG (Query expansion + Reranking + Compression)",
"content_type": "application/json",
"body": {
"message": "string (required) - User question",
"use_rag": "bool (default: true) - Enable RAG retrieval",
"use_advanced_rag": "bool (default: true) - Use advanced RAG pipeline (RECOMMENDED)",
"use_query_expansion": "bool (default: true) - Expand query with variations",
"use_reranking": "bool (default: true) - Rerank results for accuracy",
"use_compression": "bool (default: true) - Compress context to relevant parts",
"top_k": "int (default: 3) - Number of documents to retrieve",
"score_threshold": "float (default: 0.5) - Min relevance score (0-1)",
"max_tokens": "int (default: 512) - Max response tokens",
"temperature": "float (default: 0.7) - Creativity (0-1)",
"hf_token": "string (optional) - Hugging Face token"
},
"response": {
"response": "string - AI answer",
"context_used": "array - Retrieved documents with metadata",
"timestamp": "string",
"rag_stats": "object - RAG pipeline statistics (query variants, retrieval counts)"
},
"example_advanced": {
"message": "Làm sao để upload PDF có hình ảnh?",
"use_advanced_rag": True,
"use_reranking": True,
"top_k": 5,
"score_threshold": 0.5
},
"example_response_with_images": {
"response": "Để upload PDF có hình ảnh, sử dụng endpoint /upload-pdf-multimodal...",
"context_used": [
{
"id": "pdf_multimodal_...._p2_c1",
"confidence": 0.89,
"metadata": {
"text": "Bước 1: Chuẩn bị PDF với image URLs...",
"has_images": True,
"image_urls": [
"https://example.com/screenshot1.png",
"https://example.com/diagram.jpg"
],
"num_images": 2,
"page": 2
}
}
],
"rag_stats": {
"original_query": "Làm sao để upload PDF có hình ảnh?",
"expanded_queries": ["upload PDF hình ảnh", "PDF có ảnh"],
"initial_results": 10,
"after_rerank": 5,
"after_compression": 5
}
},
"notes": [
"Advanced RAG significantly improves answer quality",
"When multimodal PDF is used, images are returned in metadata",
"Requires HUGGINGFACE_TOKEN for actual LLM generation"
]
"example_response_with_images": {
"response": "Để upload PDF có hình ảnh, sử dụng endpoint /upload-pdf-multimodal...",
"context_used": [
{
"id": "pdf_multimodal_...._p2_c1",
"confidence": 0.89,
"metadata": {
"text": "Bước 1: Chuẩn bị PDF với image URLs...",
"has_images": True,
"image_urls": [
"https://example.com/screenshot1.png",
"https://example.com/diagram.jpg"
],
"num_images": 2,
"page": 2
}
}
],
"rag_stats": {
"original_query": "Làm sao để upload PDF có hình ảnh?",
"expanded_queries": ["upload PDF hình ảnh", "PDF có ảnh"],
"initial_results": 10,
"after_rerank": 5,
"after_compression": 5
}
},
"notes": [
"Advanced RAG significantly improves answer quality",
"When multimodal PDF is used, images are returned in metadata",
"Requires HUGGINGFACE_TOKEN for actual LLM generation"
]
},
"GET /history": {
"description": "Get chat history",
"query_params": {"limit": "int (default: 10)", "skip": "int (default: 0)"},
"response": {"history": "array", "total": "int"}
}
},
"management": {
"GET /documents/pdf": {
"description": "List all PDF documents",
"response": {"documents": "array", "total": "int"}
},
"DELETE /documents/pdf/{document_id}": {
"description": "Delete PDF and all its chunks",
"response": {"success": "bool", "message": "string"}
},
"GET /document/{doc_id}": {
"description": "Get document by ID",
"response": {"success": "bool", "data": "object"}
},
"DELETE /delete/{doc_id}": {
"description": "Delete document by ID",
"response": {"success": "bool", "message": "string"}
},
"GET /stats": {
"description": "Get Qdrant collection statistics",
"response": {"vectors_count": "int", "segments": "int", "indexed_vectors_count": "int"}
}
}
},
"quick_start": {
"1_upload_multimodal_pdf": "curl -X POST '/upload-pdf-multimodal' -F 'file=@user_guide.pdf' -F 'title=Guide'",
"2_verify_upload": "curl '/documents/pdf'",
"3_chat_with_rag": "curl -X POST '/chat' -H 'Content-Type: application/json' -d '{\"message\": \"How to...?\", \"use_advanced_rag\": true}'",
"4_see_images_in_context": "response['context_used'][0]['metadata']['image_urls']"
},
"use_cases": {
"user_guide_with_screenshots": {
"endpoint": "/upload-pdf-multimodal",
"description": "PDFs with text instructions + image URLs for visual guidance",
"benefits": ["Images linked to text chunks", "Chatbot returns relevant screenshots", "Perfect for step-by-step guides"]
},
"simple_text_docs": {
"endpoint": "/upload-pdf",
"description": "Simple PDFs with text only (FAQ, policies, etc.)"
},
"social_media_posts": {
"endpoint": "/index",
"description": "Index multiple posts with texts (up to 10) and images (up to 10)"
},
"complex_queries": {
"endpoint": "/chat",
"description": "Use advanced RAG for better accuracy on complex questions",
"settings": {"use_advanced_rag": True, "use_reranking": True, "use_compression": True}
}
},
"best_practices": {
"pdf_format": [
"Include image URLs in text (http://, https://)",
"Use markdown format:  or HTML: <img src='url'>",
"Clear structure with headings and sections",
"Link images close to their related text"
],
"chat_settings": {
"for_accuracy": {"temperature": 0.3, "use_advanced_rag": True, "use_reranking": True},
"for_creativity": {"temperature": 0.8, "use_advanced_rag": False},
"for_factual_answers": {"temperature": 0.3, "use_compression": True, "score_threshold": 0.6}
},
"retrieval_tuning": {
"not_finding_info": "Lower score_threshold to 0.3-0.4, increase top_k to 7-10",
"too_much_context": "Increase score_threshold to 0.6-0.7, decrease top_k to 3-5",
"slow_responses": "Disable compression, use basic RAG, decrease top_k"
}
"description": "Get chat history",
"query_params": {"limit": "int (default: 10)", "skip": "int (default: 0)"},
"response": {"history": "array", "total": "int"}
}
},
"management": {
"GET /documents/pdf": {
"description": "List all PDF documents",
"response": {"documents": "array", "total": "int"}
},
"DELETE /documents/pdf/{document_id}": {
"description": "Delete PDF and all its chunks",
"response": {"success": "bool", "message": "string"}
},
"GET /document/{doc_id}": {
"description": "Get document by ID",
"response": {"success": "bool", "data": "object"}
},
"DELETE /delete/{doc_id}": {
"description": "Delete document by ID",
"response": {"success": "bool", "message": "string"}
},
"GET /stats": {
"description": "Get Qdrant collection statistics",
"response": {"vectors_count": "int", "segments": "int", "indexed_vectors_count": "int"}
}
}
},
"quick_start": {
"1_upload_multimodal_pdf": "curl -X POST '/upload-pdf-multimodal' -F 'file=@user_guide.pdf' -F 'title=Guide'",
"2_verify_upload": "curl '/documents/pdf'",
"3_chat_with_rag": "curl -X POST '/chat' -H 'Content-Type: application/json' -d '{\"message\": \"How to...?\", \"use_advanced_rag\": true}'",
"4_see_images_in_context": "response['context_used'][0]['metadata']['image_urls']"
},
"use_cases": {
"user_guide_with_screenshots": {
"endpoint": "/upload-pdf-multimodal",
"description": "PDFs with text instructions + image URLs for visual guidance",
"benefits": ["Images linked to text chunks", "Chatbot returns relevant screenshots", "Perfect for step-by-step guides"]
},
"simple_text_docs": {
"endpoint": "/upload-pdf",
"description": "Simple PDFs with text only (FAQ, policies, etc.)"
},
"social_media_posts": {
"endpoint": "/index",
"description": "Index multiple posts with texts (up to 10) and images (up to 10)"
},
"complex_queries": {
"endpoint": "/chat",
"description": "Use advanced RAG for better accuracy on complex questions",
"settings": {"use_advanced_rag": True, "use_reranking": True, "use_compression": True}
}
},
"best_practices": {
"pdf_format": [
"Include image URLs in text (http://, https://)",
"Use markdown format:  or HTML: <img src='url'>",
"Clear structure with headings and sections",
"Link images close to their related text"
],
"chat_settings": {
"for_accuracy": {"temperature": 0.3, "use_advanced_rag": True, "use_reranking": True},
"for_creativity": {"temperature": 0.8, "use_advanced_rag": False},
"for_factual_answers": {"temperature": 0.3, "use_compression": True, "score_threshold": 0.6}
},
"retrieval_tuning": {
"not_finding_info": "Lower score_threshold to 0.3-0.4, increase top_k to 7-10",
"too_much_context": "Increase score_threshold to 0.6-0.7, decrease top_k to 3-5",
"slow_responses": "Disable compression, use basic RAG, decrease top_k"
}
},
"links": {
"docs": "http://localhost:8000/docs",
"redoc": "http://localhost:8000/redoc",
"openapi": "http://localhost:8000/openapi.json",
"guides": {
"multimodal_pdf": "See MULTIMODAL_PDF_GUIDE.md",
"advanced_rag": "See ADVANCED_RAG_GUIDE.md",
"pdf_general": "See PDF_RAG_GUIDE.md",
"quick_start": "See QUICK_START_PDF.md"
}
},
"system_info": {
"embedding_model": "Jina CLIP v2 (multimodal)",
"vector_db": "Qdrant with HNSW index",
"document_db": "MongoDB",
"rag_pipeline": "Advanced RAG with query expansion, reranking, compression",
"pdf_parser": "pypdfium2 with URL extraction",
"max_inputs": "10 texts + 10 images per /index request"
"openapi": "http://localhost:8000/openapi.json",
"guides": {
"multimodal_pdf": "See MULTIMODAL_PDF_GUIDE.md",
"advanced_rag": "See ADVANCED_RAG_GUIDE.md",
"pdf_general": "See PDF_RAG_GUIDE.md",
"quick_start": "See QUICK_START_PDF.md"
}
},
"system_info": {
"embedding_model": "Jina CLIP v2 (multimodal)",
"vector_db": "Qdrant with HNSW index",
"document_db": "MongoDB",
"rag_pipeline": "Advanced RAG with query expansion, reranking, compression",
"pdf_parser": "pypdfium2 with URL extraction",
"max_inputs": "10 texts + 10 images per /index request"
}
}
@app.post("/index", response_model=IndexResponse)
async def index_data(
id: str = Form(...),
texts: Optional[List[str]] = Form(None),
images: Optional[List[UploadFile]] = File(None),
id_use: Optional[str] = Form(None),
id_user: Optional[str] = Form(None)
texts: Optional[List[str]] = Form(None),
images: Optional[List[UploadFile]] = File(None),
id_use: Optional[str] = Form(None),
id_user: Optional[str] = Form(None)
):
"""
Index data vào vector database (hỗ trợ nhiều texts và images)
Index data vào vector database (hỗ trợ nhiều texts và images)
Body:
- id: Document ID (primary ID)
- texts: List of text contents (tiếng Việt supported) - Tối đa 10 texts
- images: List of image files (optional) - Tối đa 10 images
- id_use: ID của SocialMedia hoặc EventCode (optional)
- id_user: ID của User (optional)
- id: Document ID (primary ID)
- texts: List of text contents (tiếng Việt supported) - Tối đa 10 texts
- images: List of image files (optional) - Tối đa 10 images
- id_use: ID của SocialMedia hoặc EventCode (optional)
- id_user: ID của User (optional)
Returns:
- success: True/False
- id: Document ID
- message: Status message
Example:
```bash
curl -X POST '/index' \
-F 'id=doc123' \
-F 'id_use=social_media_456' \
-F 'id_user=user_789' \
-F 'texts=Post content 1' \
-F 'texts=Post content 2' \
-F 'images=@image1.jpg'
```
Example:
```bash
curl -X POST '/index' \
-F 'id=doc123' \
-F 'id_use=social_media_456' \
-F 'id_user=user_789' \
-F 'texts=Post content 1' \
-F 'texts=Post content 2' \
-F 'images=@image1.jpg'
```
"""
try:
# Validation
if texts is None and images is None:
raise HTTPException(status_code=400, detail="Phải cung cấp ít nhất texts hoặc images")
if texts and len(texts) > 10:
raise HTTPException(status_code=400, detail="Tối đa 10 texts")
if images and len(images) > 10:
raise HTTPException(status_code=400, detail="Tối đa 10 images")
# Validation
if texts is None and images is None:
raise HTTPException(status_code=400, detail="Phải cung cấp ít nhất texts hoặc images")
if texts and len(texts) > 10:
raise HTTPException(status_code=400, detail="Tối đa 10 texts")
if images and len(images) > 10:
raise HTTPException(status_code=400, detail="Tối đa 10 images")
# Prepare embeddings
text_embeddings = []
image_embeddings = []
text_embeddings = []
image_embeddings = []
# Encode multiple texts (tiếng Việt)
if texts:
for text in texts:
if text and text.strip():
text_emb = embedding_service.encode_text(text)
text_embeddings.append(text_emb)
# Encode multiple texts (tiếng Việt)
if texts:
for text in texts:
if text and text.strip():
text_emb = embedding_service.encode_text(text)
text_embeddings.append(text_emb)
# Encode multiple images
if images:
for image in images:
if image.filename: # Check if image is provided
image_bytes = await image.read()
pil_image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
image_emb = embedding_service.encode_image(pil_image)
image_embeddings.append(image_emb)
# Encode multiple images
if images:
for image in images:
if image.filename: # Check if image is provided
image_bytes = await image.read()
pil_image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
image_emb = embedding_service.encode_image(pil_image)
image_embeddings.append(image_emb)
# Combine embeddings
all_embeddings = []
if text_embeddings:
# Average all text embeddings
avg_text_embedding = np.mean(text_embeddings, axis=0)
all_embeddings.append(avg_text_embedding)
if image_embeddings:
# Average all image embeddings
avg_image_embedding = np.mean(image_embeddings, axis=0)
all_embeddings.append(avg_image_embedding)
if not all_embeddings:
raise HTTPException(status_code=400, detail="Không có embedding nào được tạo từ texts hoặc images")
# Final combined embedding
combined_embedding = np.mean(all_embeddings, axis=0)
all_embeddings = []
if text_embeddings:
# Average all text embeddings
avg_text_embedding = np.mean(text_embeddings, axis=0)
all_embeddings.append(avg_text_embedding)
if image_embeddings:
# Average all image embeddings
avg_image_embedding = np.mean(image_embeddings, axis=0)
all_embeddings.append(avg_image_embedding)
if not all_embeddings:
raise HTTPException(status_code=400, detail="Không có embedding nào được tạo từ texts hoặc images")
# Final combined embedding
combined_embedding = np.mean(all_embeddings, axis=0)
# Normalize
combined_embedding = combined_embedding / np.linalg.norm(combined_embedding, axis=1, keepdims=True)
# Index vào Qdrant
metadata = {
"texts": texts if texts else [],
"text_count": len(texts) if texts else 0,
"image_count": len(images) if images else 0,
"image_filenames": [img.filename for img in images] if images else [],
"id_use": id_use if id_use else None, # ID của SocialMedia hoặc EventCode
"id_user": id_user if id_user else None # ID của User
"texts": texts if texts else [],
"text_count": len(texts) if texts else 0,
"image_count": len(images) if images else 0,
"image_filenames": [img.filename for img in images] if images else [],
"id_use": id_use if id_use else None, # ID của SocialMedia hoặc EventCode
"id_user": id_user if id_user else None # ID của User
}
result = qdrant_service.index_data(
doc_id=id,
embedding=combined_embedding,
metadata=metadata
)
return IndexResponse(
success=True,
id=result["original_id"], # Trả về MongoDB ObjectId
message=f"Đã index thành công document {result['original_id']} với {len(texts) if texts else 0} texts và {len(images) if images else 0} images (Qdrant UUID: {result['qdrant_id']})"
message=f"Đã index thành công document {result['original_id']} với {len(texts) if texts else 0} texts và {len(images) if images else 0} images (Qdrant UUID: {result['qdrant_id']})"
)
except HTTPException:
raise
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi index: {str(e)}")
@app.post("/search", response_model=List[SearchResponse])
async def search(
text: Optional[str] = Form(None),
image: Optional[UploadFile] = File(None),
limit: int = Form(10),
score_threshold: Optional[float] = Form(None),
text_weight: float = Form(0.5),
image_weight: float = Form(0.5)
):
"""
Search similar documents bằng text và/hoặc image
Body:
- text: Query text (tiếng Việt supported)
- image: Query image (optional)
- limit: Số lượng kết quả (default: 10)
- score_threshold: Minimum confidence score (0-1)
- text_weight: Weight cho text search (default: 0.5)
- image_weight: Weight cho image search (default: 0.5)
Returns:
- List of results với id, confidence, và metadata
"""
try:
# Prepare query embeddings
text_embedding = None
image_embedding = None
# Encode text query
if text and text.strip():
text_embedding = embedding_service.encode_text(text)
# Encode image query
if image:
image_bytes = await image.read()
pil_image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
image_embedding = embedding_service.encode_image(pil_image)
# Validate input
if text_embedding is None and image_embedding is None:
raise HTTPException(status_code=400, detail="Phải cung cấp ít nhất text hoặc image để search")
# Hybrid search với Qdrant
results = qdrant_service.hybrid_search(
text_embedding=text_embedding,
image_embedding=image_embedding,
text_weight=text_weight,
image_weight=image_weight,
limit=limit,
score_threshold=score_threshold,
ef=256 # High accuracy search
)
# Format response
return [
SearchResponse(
id=result["id"],
confidence=result["confidence"],
metadata=result["metadata"]
)
for result in results
]
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi search: {str(e)}")
@app.post("/search/text", response_model=List[SearchResponse])
async def search_by_text(
text: str = Form(...),
limit: int = Form(10),
score_threshold: Optional[float] = Form(None)
):
"""
Search chỉ bằng text (tiếng Việt)
Body:
- text: Query text (tiếng Việt)
- limit: Số lượng kết quả
- score_threshold: Minimum confidence score
Returns:
- List of results
"""
try:
# Encode text
text_embedding = embedding_service.encode_text(text)
# Search
results = qdrant_service.search(
query_embedding=text_embedding,
limit=limit,
score_threshold=score_threshold,
ef=256
)
return [
SearchResponse(
id=result["id"],
confidence=result["confidence"],
metadata=result["metadata"]
)
for result in results
]
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi search: {str(e)}")
@app.post("/search/image", response_model=List[SearchResponse])
async def search_by_image(
image: UploadFile = File(...),
limit: int = Form(10),
score_threshold: Optional[float] = Form(None)
):
"""
Search chỉ bằng image
Body:
- image: Query image
- limit: Số lượng kết quả
- score_threshold: Minimum confidence score
Returns:
- List of results
"""
try:
# Encode image
image_bytes = await image.read()
pil_image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
image_embedding = embedding_service.encode_image(pil_image)
# Search
results = qdrant_service.search(
query_embedding=image_embedding,
limit=limit,
score_threshold=score_threshold,
ef=256
)
return [
SearchResponse(
id=result["id"],
confidence=result["confidence"],
metadata=result["metadata"]
)
for result in results
]
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi search: {str(e)}")
@app.delete("/delete/{doc_id}")
async def delete_document(doc_id: str):
"""
Delete document by ID (MongoDB ObjectId hoặc UUID)
Args:
- doc_id: Document ID to delete
Returns:
- Success message
"""
try:
qdrant_service.delete_by_id(doc_id)
return {"success": True, "message": f"Đã xóa document {doc_id}"}
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi xóa: {str(e)}")
@app.get("/document/{doc_id}")
async def get_document(doc_id: str):
"""
Get document by ID (MongoDB ObjectId hoặc UUID)
Args:
- doc_id: Document ID (MongoDB ObjectId)
Returns:
- Document data
"""
try:
doc = qdrant_service.get_by_id(doc_id)
if doc:
return {
"success": True,
"data": doc
}
raise HTTPException(status_code=404, detail=f"Không tìm thấy document {doc_id}")
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi get document: {str(e)}")
@app.get("/stats")
async def get_stats():
"""
Lấy thông tin thống kê collection
Returns:
- Collection statistics
"""
try:
info = qdrant_service.get_collection_info()
return info
except Exception as e:
raise HTTPException(status_code=500, detail=f"Lỗi khi lấy stats: {str(e)}")
# ============================================
# ChatbotRAG Endpoints
# ============================================
@app.post("/chat", response_model=ChatResponse)
async def chat(request: ChatRequest):
"""
Chat endpoint với Advanced RAG
Chat endpoint với Advanced RAG
Body:
- message: User message
- use_rag: Enable RAG retrieval (default: true)
- top_k: Number of documents to retrieve (default: 3)
- system_message: System prompt (optional)
- max_tokens: Max tokens for response (default: 512)
- temperature: Temperature for generation (default: 0.7)
- hf_token: Hugging Face token (optional, sẽ dùng env nếu không truyền)
- use_advanced_rag: Use advanced RAG pipeline (default: true)
- use_query_expansion: Enable query expansion (default: true)
- use_reranking: Enable reranking (default: true)
- use_compression: Enable context compression (default: true)
- score_threshold: Minimum relevance score (default: 0.5)
- use_advanced_rag: Use advanced RAG pipeline (default: true)
- use_query_expansion: Enable query expansion (default: true)
- use_reranking: Enable reranking (default: true)
- use_compression: Enable context compression (default: true)
- score_threshold: Minimum relevance score (default: 0.5)
Returns:
- response: Generated response
- context_used: Retrieved context documents
- timestamp: Response timestamp
- rag_stats: Statistics from RAG pipeline
- rag_stats: Statistics from RAG pipeline
"""
try:
# ============================================
# CAG Layer: Check Semantic Cache First
# ============================================
cache_hit = None
if cag_service and request.use_rag:
cache_hit = cag_service.check_cache(request.message)
if cache_hit:
# Cache hit! Return cached response immediately
return ChatResponse(
response=cache_hit["response"],
context_used=cache_hit["context_used"],
timestamp=datetime.utcnow().isoformat(),
rag_stats={
**cache_hit.get("rag_stats", {}),
"cache_hit": True,
"cached_query": cache_hit["cached_query"],
"similarity_score": cache_hit["similarity_score"],
"cached_at": cache_hit["cached_at"]
}
)
# ============================================
# RAG Pipeline (if cache miss)
# ============================================
# Retrieve context if RAG enabled
context_used = []
rag_stats = None
rag_stats = None
if request.use_rag:
if request.use_advanced_rag:
# Initialize LLM client for query expansion
hf_client = None
if request.hf_token or hf_token:
hf_client = InferenceClient(token=request.hf_token or hf_token)
# Use Advanced RAG Pipeline (Best Case 2025)
documents, stats = advanced_rag.hybrid_rag_pipeline(
query=request.message,
top_k=request.top_k,
score_threshold=request.score_threshold,
use_reranking=request.use_reranking,
use_compression=request.use_compression,
use_query_expansion=request.use_query_expansion,
max_context_tokens=500,
hf_client=hf_client
)
# Convert to dict format for compatibility
context_used = [
{
"id": doc.id,
"confidence": doc.confidence,
"metadata": doc.metadata
}
for doc in documents
]
rag_stats = stats
# Format context using advanced RAG formatter
context_text = advanced_rag.format_context_for_llm(documents)
else:
# Use basic RAG (original implementation)
query_embedding = embedding_service.encode_text(request.message)
results = qdrant_service.search(
query_embedding=query_embedding,
limit=request.top_k,
score_threshold=request.score_threshold
)
context_used = results
# Build context text (basic format)
context_text = "\n\nRelevant Context:\n"
for i, doc in enumerate(context_used, 1):
doc_text = doc["metadata"].get("text", "")
confidence = doc["confidence"]
context_text += f"\n[{i}] (Confidence: {confidence:.2f})\n{doc_text}\n"
# Build context text (basic format)
context_text = "\n\nRelevant Context:\n"
for i, doc in enumerate(context_used, 1):
doc_text = doc["metadata"].get("text", "")
confidence = doc["confidence"]
context_text += f"\n[{i}] (Confidence: {confidence:.2f})\n{doc_text}\n"
# Build system message with context
if request.use_rag and context_used:
if request.use_advanced_rag:
# Use advanced prompt builder
system_message = advanced_rag.build_rag_prompt(
query=request.message,
context=context_text,
system_message=request.system_message
)
else:
# Basic prompt
system_message = f"{request.system_message}\n{context_text}\n\nPlease use the above context to answer the user's question when relevant."
# Build system message with context
if request.use_rag and context_used:
if request.use_advanced_rag:
# Use advanced prompt builder
system_message = advanced_rag.build_rag_prompt(
query=request.message,
context=context_text,
system_message=request.system_message
)
else:
# Basic prompt
system_message = f"{request.system_message}\n{context_text}\n\nPlease use the above context to answer the user's question when relevant."
else:
system_message = request.system_message
# Use token from request or fallback to env
token = request.hf_token or hf_token
# Generate response
if not token:
response = f"""[LLM Response Placeholder]
Context retrieved: {len(context_used)} documents
User question: {request.message}
To enable actual LLM generation:
1. Set HUGGINGFACE_TOKEN environment variable, OR
2. Pass hf_token in request body
Example:
{{
"message": "Your question",
"hf_token": "hf_xxxxxxxxxxxxx"
}}
"""
else:
try:
client = InferenceClient(
token=hf_token,
model="openai/gpt-oss-20b"
)
# Build messages
messages = [
{"role": "system", "content": system_message},
{"role": "user", "content": request.message}
]
# Generate response
response = ""
for msg in client.chat_completion(
messages,
max_tokens=request.max_tokens,
stream=True,
temperature=request.temperature,
top_p=request.top_p,
):
choices = msg.choices
if len(choices) and choices[0].delta.content:
response += choices[0].delta.content
except Exception as e:
response = f"Error generating response with LLM: {str(e)}\n\nContext was retrieved successfully, but LLM generation failed."
# Save to history
chat_data = {
"user_message": request.message,
"assistant_response": response,
"context_used": context_used,
"timestamp": datetime.utcnow()
}
chat_history_collection.insert_one(chat_data)
# ============================================
# CAG: Save to Cache (if RAG was used)
# ============================================
if cag_service and request.use_rag and context_used and response:
try:
cag_service.save_to_cache(
query=request.message,
response=response,
context_used=context_used,
rag_stats=rag_stats
)
except Exception as cache_error:
print(f"Warning: Failed to save to cache: {cache_error}")
return ChatResponse(
response=response,
context_used=context_used,
timestamp=datetime.utcnow().isoformat(),
rag_stats=rag_stats
timestamp=datetime.utcnow().isoformat(),
rag_stats=rag_stats
)
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.post("/documents", response_model=AddDocumentResponse)
async def add_document(request: AddDocumentRequest):
"""
Add document to knowledge base
Body:
- text: Document text
- metadata: Additional metadata (optional)
Returns:
- success: True/False
- doc_id: MongoDB document ID
- message: Status message
"""
try:
# Save to MongoDB
doc_data = {
"text": request.text,
"metadata": request.metadata or {},
"created_at": datetime.utcnow()
}
result = documents_collection.insert_one(doc_data)
doc_id = str(result.inserted_id)
# Generate embedding
embedding = embedding_service.encode_text(request.text)
# Index to Qdrant
qdrant_service.index_data(
doc_id=doc_id,
embedding=embedding,
metadata={
"text": request.text,
"source": "api",
**(request.metadata or {})
}
)
return AddDocumentResponse(
success=True,
doc_id=doc_id,
message=f"Document added successfully with ID: {doc_id}"
)
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.post("/rag/search", response_model=List[SearchResponse])
async def rag_search(
query: str = Form(...),
top_k: int = Form(5),
score_threshold: Optional[float] = Form(0.5)
):
"""
Search in knowledge base
Body:
- query: Search query
- top_k: Number of results (default: 5)
- score_threshold: Minimum score (default: 0.5)
Returns:
- results: List of matching documents
"""
try:
# Generate query embedding
query_embedding = embedding_service.encode_text(query)
# Search in Qdrant
results = qdrant_service.search(
query_embedding=query_embedding,
limit=top_k,
score_threshold=score_threshold
)
return [
SearchResponse(
id=result["id"],
confidence=result["confidence"],
metadata=result["metadata"]
)
for result in results
]
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.get("/history")
async def get_history(limit: int = 10, skip: int = 0):
"""
Get chat history
Query params:
- limit: Number of messages to return (default: 10)
- skip: Number of messages to skip (default: 0)
Returns:
- history: List of chat messages
"""
try:
history = list(
chat_history_collection
.find({}, {"_id": 0})
.sort("timestamp", -1)
.skip(skip)
.limit(limit)
)
# Convert datetime to string
for msg in history:
if "timestamp" in msg:
msg["timestamp"] = msg["timestamp"].isoformat()
return {
"history": history,
"total": chat_history_collection.count_documents({})
}
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.delete("/documents/{doc_id}")
async def delete_document_from_kb(doc_id: str):
"""
Delete document from knowledge base
Args:
- doc_id: Document ID (MongoDB ObjectId)
Returns:
- success: True/False
- message: Status message
"""
try:
# Delete from MongoDB
result = documents_collection.delete_one({"_id": doc_id})
# Delete from Qdrant
if result.deleted_count > 0:
qdrant_service.delete_by_id(doc_id)
return {"success": True, "message": f"Document {doc_id} deleted from knowledge base"}
else:
raise HTTPException(status_code=404, detail=f"Document {doc_id} not found")
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.post("/upload-pdf", response_model=UploadPDFResponse)
async def upload_pdf(
file: UploadFile = File(...),
document_id: Optional[str] = Form(None),
title: Optional[str] = Form(None),
description: Optional[str] = Form(None),
category: Optional[str] = Form(None)
):
"""
Upload and index PDF file into knowledge base
Body (multipart/form-data):
- file: PDF file (required)
- document_id: Custom document ID (optional, auto-generated if not provided)
- title: Document title (optional)
- description: Document description (optional)
- category: Document category (optional, e.g., "user_guide", "faq")
Returns:
- success: True/False
- document_id: Document ID
- filename: Original filename
- chunks_indexed: Number of chunks created
- message: Status message
Example:
```bash
curl -X POST "http://localhost:8000/upload-pdf" \
-F "file=@user_guide.pdf" \
-F "title=Hướng dẫn sử dụng ChatbotRAG" \
-F "category=user_guide"
```
"""
try:
# Validate file type
if not file.filename.endswith('.pdf'):
raise HTTPException(status_code=400, detail="Only PDF files are allowed")
# Generate document ID if not provided
if not document_id:
from datetime import datetime
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
document_id = f"pdf_{timestamp}"
# Read PDF bytes
pdf_bytes = await file.read()
# Prepare metadata
metadata = {}
if title:
metadata['title'] = title
if description:
metadata['description'] = description
if category:
metadata['category'] = category
# Index PDF
result = pdf_indexer.index_pdf_bytes(
pdf_bytes=pdf_bytes,
document_id=document_id,
filename=file.filename,
document_metadata=metadata
)
return UploadPDFResponse(
success=True,
document_id=result['document_id'],
filename=result['filename'],
chunks_indexed=result['chunks_indexed'],
message=f"PDF '{file.filename}' đã được index thành công với {result['chunks_indexed']} chunks"
)
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error uploading PDF: {str(e)}")
@app.get("/documents/pdf")
async def list_pdf_documents():
"""
List all PDF documents in knowledge base
Returns:
- documents: List of PDF documents with metadata
"""
try:
docs = list(documents_collection.find(
{"type": "pdf"},
{"_id": 0}
))
return {"documents": docs, "total": len(docs)}
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.delete("/documents/pdf/{document_id}")
async def delete_pdf_document(document_id: str):
"""
Delete PDF document and all its chunks from knowledge base
Args:
- document_id: Document ID
Returns:
- success: True/False
- message: Status message
"""
try:
# Get document info
doc = documents_collection.find_one({"document_id": document_id, "type": "pdf"})
if not doc:
raise HTTPException(status_code=404, detail=f"PDF document {document_id} not found")
# Delete all chunks from Qdrant
chunk_ids = doc.get('chunk_ids', [])
for chunk_id in chunk_ids:
try:
qdrant_service.delete_by_id(chunk_id)
except:
pass # Chunk might already be deleted
# Delete from MongoDB
documents_collection.delete_one({"document_id": document_id})
return {
"success": True,
"message": f"PDF document {document_id} and {len(chunk_ids)} chunks deleted"
}
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.post("/upload-pdf-multimodal", response_model=UploadPDFResponse)
async def upload_pdf_multimodal(
file: UploadFile = File(...),
document_id: Optional[str] = Form(None),
title: Optional[str] = Form(None),
description: Optional[str] = Form(None),
category: Optional[str] = Form(None)
):
"""
Upload PDF with text and image URLs (for user guides with screenshots)
This endpoint is optimized for PDFs containing:
- Text instructions
- Image URLs (http://... or https://...)
- Markdown images: 
- HTML images: <img src="url">
The system will:
1. Extract text from PDF
2. Detect all image URLs in the text
3. Link images to their corresponding text chunks
4. Store image URLs in metadata
5. Return images along with text during chat
Body (multipart/form-data):
- file: PDF file (required)
- document_id: Custom document ID (optional, auto-generated if not provided)
- title: Document title (optional)
- description: Document description (optional)
- category: Document category (optional, e.g., "user_guide", "tutorial")
Returns:
- success: True/False
- document_id: Document ID
- filename: Original filename
- chunks_indexed: Number of chunks created
- message: Status message (includes image count)
Example:
```bash
curl -X POST "http://localhost:8000/upload-pdf-multimodal" \
-F "file=@user_guide_with_images.pdf" \
-F "title=Hướng dẫn có ảnh minh họa" \
-F "category=user_guide"
```
Example Response:
```json
{
"success": true,
"document_id": "pdf_20251029_150000",
"filename": "user_guide_with_images.pdf",
"chunks_indexed": 25,
"message": "PDF 'user_guide_with_images.pdf' indexed with 25 chunks and 15 images"
}
```
"""
try:
# Validate file type
if not file.filename.endswith('.pdf'):
raise HTTPException(status_code=400, detail="Only PDF files are allowed")
# Generate document ID if not provided
if not document_id:
from datetime import datetime
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
document_id = f"pdf_multimodal_{timestamp}"
# Read PDF bytes
pdf_bytes = await file.read()
# Prepare metadata
metadata = {'type': 'multimodal'}
if title:
metadata['title'] = title
if description:
metadata['description'] = description
if category:
metadata['category'] = category
# Index PDF with multimodal parser
result = multimodal_pdf_indexer.index_pdf_bytes(
pdf_bytes=pdf_bytes,
document_id=document_id,
filename=file.filename,
document_metadata=metadata
)
return UploadPDFResponse(
success=True,
document_id=result['document_id'],
filename=result['filename'],
chunks_indexed=result['chunks_indexed'],
message=f"PDF '{file.filename}' indexed successfully with {result['chunks_indexed']} chunks and {result.get('images_found', 0)} images"
)
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error uploading multimodal PDF: {str(e)}")
@app.post("/upload-pdf", response_model=UploadPDFResponse)
async def upload_pdf(
file: UploadFile = File(...),
document_id: Optional[str] = Form(None),
title: Optional[str] = Form(None),
description: Optional[str] = Form(None),
category: Optional[str] = Form(None)
):
"""
Upload and index PDF file into knowledge base
Body (multipart/form-data):
- file: PDF file (required)
- document_id: Custom document ID (optional, auto-generated if not provided)
- title: Document title (optional)
- description: Document description (optional)
- category: Document category (optional, e.g., "user_guide", "faq")
Returns:
- success: True/False
- document_id: Document ID
- filename: Original filename
- chunks_indexed: Number of chunks created
- message: Status message
Example:
```bash
curl -X POST "http://localhost:8000/upload-pdf" \
-F "file=@user_guide.pdf" \
-F "title=Hướng dẫn sử dụng ChatbotRAG" \
-F "category=user_guide"
```
"""
try:
# Validate file type
if not file.filename.endswith('.pdf'):
raise HTTPException(status_code=400, detail="Only PDF files are allowed")
# Generate document ID if not provided
if not document_id:
from datetime import datetime
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
document_id = f"pdf_{timestamp}"
# Read PDF bytes
pdf_bytes = await file.read()
# Prepare metadata
metadata = {}
if title:
metadata['title'] = title
if description:
metadata['description'] = description
if category:
metadata['category'] = category
# Index PDF
result = pdf_indexer.index_pdf_bytes(
pdf_bytes=pdf_bytes,
document_id=document_id,
filename=file.filename,
document_metadata=metadata
)
return UploadPDFResponse(
success=True,
document_id=result['document_id'],
filename=result['filename'],
chunks_indexed=result['chunks_indexed'],
message=f"PDF '{file.filename}' đã được index thành công với {result['chunks_indexed']} chunks"
)
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error uploading PDF: {str(e)}")
@app.get("/documents/pdf")
async def list_pdf_documents():
"""
List all PDF documents in knowledge base
Returns:
- documents: List of PDF documents with metadata
"""
try:
docs = list(documents_collection.find(
{"type": "pdf"},
{"_id": 0}
))
return {"documents": docs, "total": len(docs)}
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.delete("/documents/pdf/{document_id}")
async def delete_pdf_document(document_id: str):
"""
Delete PDF document and all its chunks from knowledge base
Args:
- document_id: Document ID
Returns:
- success: True/False
- message: Status message
"""
try:
# Get document info
doc = documents_collection.find_one({"document_id": document_id, "type": "pdf"})
if not doc:
raise HTTPException(status_code=404, detail=f"PDF document {document_id} not found")
# Delete all chunks from Qdrant
chunk_ids = doc.get('chunk_ids', [])
for chunk_id in chunk_ids:
try:
qdrant_service.delete_by_id(chunk_id)
except:
pass # Chunk might already be deleted
# Delete from MongoDB
documents_collection.delete_one({"document_id": document_id})
return {
"success": True,
"message": f"PDF document {document_id} and {len(chunk_ids)} chunks deleted"
}
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error: {str(e)}")
@app.post("/upload-pdf-multimodal", response_model=UploadPDFResponse)
async def upload_pdf_multimodal(
file: UploadFile = File(...),
document_id: Optional[str] = Form(None),
title: Optional[str] = Form(None),
description: Optional[str] = Form(None),
category: Optional[str] = Form(None)
):
"""
Upload PDF with text and image URLs (for user guides with screenshots)
This endpoint is optimized for PDFs containing:
- Text instructions
- Image URLs (http://... or https://...)
- Markdown images: 
- HTML images: <img src="url">
The system will:
1. Extract text from PDF
2. Detect all image URLs in the text
3. Link images to their corresponding text chunks
4. Store image URLs in metadata
5. Return images along with text during chat
Body (multipart/form-data):
- file: PDF file (required)
- document_id: Custom document ID (optional, auto-generated if not provided)
- title: Document title (optional)
- description: Document description (optional)
- category: Document category (optional, e.g., "user_guide", "tutorial")
Returns:
- success: True/False
- document_id: Document ID
- filename: Original filename
- chunks_indexed: Number of chunks created
- message: Status message (includes image count)
Example:
```bash
curl -X POST "http://localhost:8000/upload-pdf-multimodal" \
-F "file=@user_guide_with_images.pdf" \
-F "title=Hướng dẫn có ảnh minh họa" \
-F "category=user_guide"
```
Example Response:
```json
{
"success": true,
"document_id": "pdf_20251029_150000",
"filename": "user_guide_with_images.pdf",
"chunks_indexed": 25,
"message": "PDF 'user_guide_with_images.pdf' indexed with 25 chunks and 15 images"
}
```
"""
try:
# Validate file type
if not file.filename.endswith('.pdf'):
raise HTTPException(status_code=400, detail="Only PDF files are allowed")
# Generate document ID if not provided
if not document_id:
from datetime import datetime
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
document_id = f"pdf_multimodal_{timestamp}"
# Read PDF bytes
pdf_bytes = await file.read()
# Prepare metadata
metadata = {'type': 'multimodal'}
if title:
metadata['title'] = title
if description:
metadata['description'] = description
if category:
metadata['category'] = category
# Index PDF with multimodal parser
result = multimodal_pdf_indexer.index_pdf_bytes(
pdf_bytes=pdf_bytes,
document_id=document_id,
filename=file.filename,
document_metadata=metadata
)
return UploadPDFResponse(
success=True,
document_id=result['document_id'],
filename=result['filename'],
chunks_indexed=result['chunks_indexed'],
message=f"PDF '{file.filename}' indexed successfully with {result['chunks_indexed']} chunks and {result.get('images_found', 0)} images"
)
except HTTPException:
raise
except Exception as e:
raise HTTPException(status_code=500, detail=f"Error uploading multimodal PDF: {str(e)}")
if __name__ == "__main__":
import uvicorn
uvicorn.run(
app,
host="0.0.0.0",
port=8000,
log_level="info"
)
|