Spaces:
Running
Running
File size: 78,896 Bytes
1c3c064 75e21d7 f5bf3a2 1c3c064 f5bf3a2 2d11264 1c3c064 f5bf3a2 ad25f3a f5bf3a2 ad25f3a 3f6e731 ad25f3a 3f6e731 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 75e21d7 f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 ad25f3a f5bf3a2 75e21d7 f5bf3a2 75e21d7 00a56c1 e5af422 8b09b87 00a56c1 834c402 00a56c1 1c3c064 44be9b7 1c3c064 44be9b7 1c3c064 3f6e731 ad25f3a 3f6e731 ad25f3a 3f6e731 ad25f3a 3f6e731 ad25f3a 3f6e731 f5bf3a2 ad25f3a f5bf3a2 1c3c064 d824a9d 1c3c064 00a56c1 1c3c064 d824a9d 1c3c064 00a56c1 1c3c064 00a56c1 1c3c064 00a56c1 1c3c064 f5bf3a2 1c3c064 f5bf3a2 1c3c064 f5bf3a2 1c3c064 f5bf3a2 1c3c064 b5c0b05 8b09b87 b5c0b05 8b09b87 b5c0b05 8b09b87 b5c0b05 1c3c064 8b09b87 1c3c064 bcd379e 1c3c064 b5c0b05 00a56c1 1c3c064 f5bf3a2 1c3c064 d824a9d 1c3c064 d824a9d 1c3c064 f5bf3a2 ad25f3a 1c3c064 00a56c1 dd66841 1c3c064 f5bf3a2 dd66841 1c3c064 dd66841 f5bf3a2 1c3c064 8b09b87 1c3c064 00a56c1 8b09b87 f5bf3a2 8b09b87 00a56c1 8b09b87 80369d7 00a56c1 8b09b87 f247ad6 8b09b87 00a56c1 8b09b87 00a56c1 1c3c064 f5bf3a2 1c3c064 d824a9d 1c3c064 f5bf3a2 1c3c064 ae870fb 1c3c064 ae870fb 1c3c064 ae870fb 1c3c064 d824a9d 1c3c064 d824a9d 1c3c064 d824a9d 1c3c064 f5bf3a2 1c3c064 d824a9d 1c3c064 d824a9d 1c3c064 44be9b7 1c3c064 d824a9d 44be9b7 d824a9d 44be9b7 1c3c064 44be9b7 1c3c064 44be9b7 1c3c064 f5bf3a2 1c3c064 44be9b7 00a56c1 1c3c064 2d11264 f5bf3a2 1c3c064 2d11264 1c3c064 f5bf3a2 ad25f3a 1c3c064 dd66841 1c3c064 f5bf3a2 1c3c064 dd66841 f5bf3a2 1c3c064 2d11264 1c3c064 2f01cc6 2d11264 1c3c064 2d11264 1c3c064 2d11264 2f01cc6 2d11264 2f01cc6 2d11264 1c3c064 f5bf3a2 1c3c064 2f01cc6 1c3c064 2d11264 1c3c064 f5bf3a2 1c3c064 834c402 f5bf3a2 834c402 f5bf3a2 ad25f3a 834c402 dd66841 834c402 f5bf3a2 dd66841 834c402 dd66841 f5bf3a2 834c402 f5bf3a2 834c402 f5bf3a2 834c402 f5bf3a2 834c402 2d11264 86371bb 2d11264 f5bf3a2 86371bb 1c3c064 f5bf3a2 d824a9d 75e21d7 dd66841 f5bf3a2 75e21d7 1c3c064 54a9e8f 1c3c064 b5c0b05 5bb0fb3 b5c0b05 1c3c064 00a56c1 1c3c064 44be9b7 1c3c064 44be9b7 1c3c064 44be9b7 00a56c1 44be9b7 1c3c064 679cf3c 1c3c064 44be9b7 679cf3c 44be9b7 679cf3c 44be9b7 b5c0b05 44be9b7 b5c0b05 44be9b7 679cf3c 44be9b7 b5c0b05 679cf3c b5c0b05 679cf3c b5c0b05 679cf3c b5c0b05 679cf3c 44be9b7 679cf3c 44be9b7 1c3c064 f5bf3a2 54a9e8f ad25f3a 1c3c064 44be9b7 1c3c064 44be9b7 d824a9d 44be9b7 1c3c064 2d11264 1c3c064 2d11264 86371bb 2d11264 86371bb 2d11264 86371bb 2d11264 86371bb 1c3c064 86371bb 2d11264 86371bb 2d11264 f5bf3a2 86371bb 2d11264 86371bb 2d11264 f5bf3a2 2d11264 1c3c064 86371bb 1c3c064 2d11264 f5bf3a2 54a9e8f ad25f3a 1c3c064 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 ae870fb 834c402 f5bf3a2 54a9e8f ad25f3a 834c402 ae870fb 1c3c064 75e21d7 f5bf3a2 7141319 75e21d7 7141319 75e21d7 dd66841 1c3c064 f5bf3a2 8571394 1c3c064 | 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 | # -*- coding: utf-8 -*-
"""
人力V助手 (JinrikiHelper) 云端 Web UI
基于 Gradio 6.2.0 构建
专为云端部署优化：上传 → 处理 → 下载
作者：TNOT
"""
import gradio as gr
import logging
import os
import sys
import json
import tempfile
import zipfile
import shutil
import uuid
import threading
import time
import re
from pathlib import Path
from dataclasses import dataclass
from typing import Optional, List, Dict, Tuple, Any
# 配置日志
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
handlers=[logging.StreamHandler(sys.stdout)]
)
logger = logging.getLogger(__name__)
@dataclass
class TaskToken:
"""任务令牌（用于取消/超时/状态跟踪）"""
task_id: str
session_id: str
kind: str
started_at: float
last_touch: float
cancel_event: threading.Event
class CloudTaskManager:
"""云端任务管理器：独立队列并发 + 会话取消 + 滞留回收
支持四个独立的并发队列：
- make: 综合任务（VAD→Whisper→MFA→打包）包含最多资源消耗
- whisper: 单独的语音识别任务
- mfa: 单独的 MFA 对齐任务
- export: 音源导出任务
"""
def __init__(self,
max_make_jobs: int = 5,
max_whisper_jobs: int = 0,
max_mfa_jobs: int = 2,
max_export_jobs: int = 10,
max_job_seconds: int = 7200):
"""初始化任务管理器
Args:
max_make_jobs: 综合任务（make）的最大并发数（包含 Whisper+MFA）
max_whisper_jobs: 单独 Whisper 任务的最大并发数（0表示不独立限制）
max_mfa_jobs: 单独 MFA 对齐的最大并发数
max_export_jobs: 音源导出的最大并发数
max_job_seconds: 单个任务的最大运行时间（秒）
"""
self.max_make_jobs = max(1, int(max_make_jobs))
self.max_whisper_jobs = max(0, int(max_whisper_jobs)) # 0 表示不限制
self.max_mfa_jobs = max(1, int(max_mfa_jobs))
self.max_export_jobs = max(1, int(max_export_jobs))
self.max_job_seconds = max(60, int(max_job_seconds))
self._lock = threading.Lock()
self._tasks: Dict[str, TaskToken] = {}
self._session_index: Dict[str, set] = {}
# 按任务类型统计并发数
self._task_counts: Dict[str, int] = {
"make": 0,
"whisper": 0,
"mfa": 0,
"export": 0,
}
self._watchdog = threading.Thread(
target=self._watchdog_loop,
name="cloud-task-watchdog",
daemon=True
)
self._watchdog.start()
def _get_max_for_kind(self, kind: str) -> int:
"""根据任务类型获取最大并发数"""
if kind == "make":
return self.max_make_jobs
elif kind == "whisper":
return self.max_whisper_jobs if self.max_whisper_jobs > 0 else 999 # 0 表示不限制
elif kind == "mfa":
return self.max_mfa_jobs
elif kind == "export":
return self.max_export_jobs
else:
return 1 # 未知类型默认 1
def start_task(self, kind: str, session_id: Optional[str]) -> Optional[TaskToken]:
sid = (session_id or "").strip() or uuid.uuid4().hex[:8]
now = time.time()
with self._lock:
# 检查该类型是否达到上限
current = self._task_counts.get(kind, 0)
max_allowed = self._get_max_for_kind(kind)
if current >= max_allowed:
return None
task_id = uuid.uuid4().hex[:10]
token = TaskToken(
task_id=task_id,
session_id=sid,
kind=kind,
started_at=now,
last_touch=now,
cancel_event=threading.Event()
)
self._tasks[task_id] = token
self._task_counts[kind] = current + 1
self._session_index.setdefault(sid, set()).add(task_id)
return token
def finish_task(self, task_id: Optional[str]):
if not task_id:
return
with self._lock:
token = self._tasks.pop(task_id, None)
if not token:
return
# 更新任务类型的计数
current = self._task_counts.get(token.kind, 0)
self._task_counts[token.kind] = max(0, current - 1)
session_tasks = self._session_index.get(token.session_id)
if session_tasks:
session_tasks.discard(task_id)
if not session_tasks:
self._session_index.pop(token.session_id, None)
def touch(self, task_id: Optional[str]):
if not task_id:
return
now = time.time()
with self._lock:
token = self._tasks.get(task_id)
if token:
token.last_touch = now
def should_cancel(self, task_id: Optional[str]) -> bool:
if not task_id:
return False
with self._lock:
token = self._tasks.get(task_id)
return bool(token and token.cancel_event.is_set())
def cancel_session(self, session_id: Optional[str]) -> int:
sid = (session_id or "").strip()
if not sid:
return 0
cancelled = 0
with self._lock:
for task_id in list(self._session_index.get(sid, set())):
token = self._tasks.get(task_id)
if token and not token.cancel_event.is_set():
token.cancel_event.set()
cancelled += 1
return cancelled
def get_status_text(self) -> str:
with self._lock:
counts = dict(self._task_counts) # 复制当前计数
running = [f"{t.kind}:{t.task_id}" for t in self._tasks.values()]
# 格式化队列状态
make_status = f"制作: {counts.get('make', 0)}/{self.max_make_jobs}"
whisper_status = f"识别: {counts.get('whisper', 0)}/{self.max_whisper_jobs if self.max_whisper_jobs > 0 else '∞'}"
mfa_status = f"对齐: {counts.get('mfa', 0)}/{self.max_mfa_jobs}"
export_status = f"导出: {counts.get('export', 0)}/{self.max_export_jobs}"
total_active = sum(counts.values())
status_line = f"【任务队列】 {make_status} | {whisper_status} | {mfa_status} | {export_status}"
if running:
detail = " | ".join(running[:3])
if len(running) > 3:
detail += f" ...共{len(running)}个"
return f"{status_line}\\n【运行中】 {detail}"
return status_line
def _watchdog_loop(self):
while True:
time.sleep(5)
now = time.time()
with self._lock:
for token in self._tasks.values():
runtime = now - token.started_at
if runtime > self.max_job_seconds and not token.cancel_event.is_set():
token.cancel_event.set()
logger.warning(
"任务超时，已标记取消: %s (%s, %.1fs)",
token.task_id, token.kind, runtime
)
def get_concurrency_status() -> str:
"""获取当前并发状态文本"""
return TASK_MANAGER.get_status_text()
def sanitize_source_name(name: Optional[str]) -> Optional[str]:
"""净化用户输入的音源名，避免路径逃逸与非法文件名。"""
cleaned = re.sub(r"[^0-9A-Za-z_\-\u4e00-\u9fff]", "_", (name or "").strip())
cleaned = cleaned.strip("._- ")
if not cleaned:
return None
return cleaned[:64]
def safe_gradio_handler(func):
"""
Gradio 处理函数的安全包装器
捕获所有异常并返回友好的错误信息，避免 Gradio 显示默认的"错误"状态
"""
import functools
import traceback
@functools.wraps(func)
def wrapper(*args, **kwargs):
try:
return func(*args, **kwargs)
except Exception as e:
# 记录完整的异常堆栈
error_trace = traceback.format_exc()
logger.error(f"处理函数 {func.__name__} 发生异常:\n{error_trace}")
# 根据函数返回值数量返回错误信息
# 检查函数的类型注解来确定返回值数量
annotations = getattr(func, '__annotations__', {})
return_type = annotations.get('return', None)
error_msg = f"❌ 系统错误: {str(e)}"
error_detail = f"异常类型: {type(e).__name__}\n详情: {str(e)}"
# 根据函数名判断返回值数量
if func.__name__ == 'process_make_voicebank':
return error_msg, error_detail, None, None
elif func.__name__ == 'process_export_voicebank':
return error_msg, error_detail, None
elif func.__name__ == 'collect_and_export':
return error_msg, error_detail, None
elif func.__name__ == 'process_mfa_realign':
return error_msg, error_detail, None
else:
# 默认返回单个错误消息
return error_msg
return wrapper
# 项目根目录
BASE_DIR = Path(__file__).parent.parent.absolute()
class CloudConfig:
"""云端配置"""
# 临时工作目录
TEMP_BASE = tempfile.gettempdir()
# 模型目录（云端使用项目内目录）
MODELS_DIR = str(BASE_DIR / "models")
MFA_DIR = str(BASE_DIR / "models" / "mfa")
# 支持的音频格式
AUDIO_EXTENSIONS = ('.wav', '.mp3', '.flac', '.ogg', '.m4a')
# Whisper 模型选项（含速度说明）
WHISPER_MODELS = {
"whisper-small (快速，约4秒/句)": "openai/whisper-small",
"whisper-medium (精准，约12秒/句)": "openai/whisper-medium"
}
# 语言选项
LANGUAGES = ["chinese", "japanese"]
# 独立队列并发限制（针对升级后硬件配置的推荐配置）
# 支持通过环境变量覆盖默认值
#
# 资源消耗估算（每任务）：
# - make (VAD→Whisper→MFA→打包): 6-10GB, 10-30分钟
# - whisper (单独识别): 3-5GB, 4-12秒/句
# - mfa (单独对齐, 自动多核): 2-4GB, 5-20秒/句
# - export (导出, I/O密集): 1-2GB, 1-5秒/句
#
MAX_MAKE_JOBS = int(os.environ.get("JINRIKI_MAX_MAKE_JOBS", "5"))
MAX_WHISPER_JOBS = int(os.environ.get("JINRIKI_MAX_WHISPER_JOBS", "0")) # 0=不限制（默认）
MAX_MFA_JOBS = int(os.environ.get("JINRIKI_MAX_MFA_JOBS", "2"))
MAX_EXPORT_JOBS = int(os.environ.get("JINRIKI_MAX_EXPORT_JOBS", "10"))
# 任务超时时间（秒）
MAX_JOB_SECONDS = int(os.environ.get("JINRIKI_MAX_JOB_SECONDS", "7200")) # 2小时
TASK_MANAGER = CloudTaskManager(
max_make_jobs=CloudConfig.MAX_MAKE_JOBS,
max_whisper_jobs=CloudConfig.MAX_WHISPER_JOBS,
max_mfa_jobs=CloudConfig.MAX_MFA_JOBS,
max_export_jobs=CloudConfig.MAX_EXPORT_JOBS,
max_job_seconds=CloudConfig.MAX_JOB_SECONDS
)
def create_temp_workspace() -> str:
"""创建临时工作空间"""
workspace_id = str(uuid.uuid4())[:8]
workspace = os.path.join(CloudConfig.TEMP_BASE, f"jinriki_{workspace_id}")
os.makedirs(workspace, exist_ok=True)
return workspace
def cleanup_workspace(workspace: str):
"""清理工作空间"""
if workspace and os.path.exists(workspace):
try:
shutil.rmtree(workspace)
logger.info(f"已清理工作空间: {workspace}")
except Exception as e:
logger.warning(f"清理工作空间失败: {e}")
def create_zip(source_dir: str, zip_name: str) -> Optional[str]:
"""打包目录为 zip（使用 uuid 避免多用户冲突）"""
if not os.path.isdir(source_dir):
logger.warning(f"打包失败: 目录不存在 {source_dir}")
return None
try:
unique_id = str(uuid.uuid4())[:8]
zip_path = os.path.join(CloudConfig.TEMP_BASE, f"{zip_name}_{unique_id}.zip")
with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zf:
file_count = 0
for root, dirs, files in os.walk(source_dir):
for file in files:
file_path = os.path.join(root, file)
arcname = os.path.relpath(file_path, source_dir)
zf.write(file_path, arcname)
file_count += 1
logger.info(f"打包完成: {zip_path} ({file_count} 个文件)")
return zip_path
except Exception as e:
logger.error(f"打包失败: {e}", exc_info=True)
return None
def extract_zip(zip_path: str, target_dir: str) -> Tuple[bool, str]:
"""安全解压 zip 文件（防 Zip Slip）"""
try:
target_real = os.path.realpath(target_dir)
os.makedirs(target_real, exist_ok=True)
with zipfile.ZipFile(zip_path, 'r') as zf:
for info in zf.infolist():
member = info.filename.replace('\\\\', '/')
if member.startswith('/') or member.startswith('..'):
return False, f"解压失败: 非法路径 {member}"
if '..' in Path(member).parts:
return False, f"解压失败: 非法路径 {member}"
out_path = os.path.realpath(os.path.join(target_real, member))
if out_path != target_real and not out_path.startswith(target_real + os.sep):
return False, f"解压失败: 路径越界 {member}"
if info.is_dir():
os.makedirs(out_path, exist_ok=True)
continue
os.makedirs(os.path.dirname(out_path), exist_ok=True)
with zf.open(info, 'r') as src, open(out_path, 'wb') as dst:
shutil.copyfileobj(src, dst)
return True, "解压成功"
except Exception as e:
return False, f"解压失败: {e}"
def scan_mfa_models() -> Dict[str, List[str]]:
"""扫描 MFA 模型"""
result = {"acoustic": [], "dictionary": []}
if not os.path.exists(CloudConfig.MFA_DIR):
return result
for f in os.listdir(CloudConfig.MFA_DIR):
if f.endswith('.zip'):
result["acoustic"].append(f)
elif f.endswith('.dict') or f.endswith('.txt'):
result["dictionary"].append(f)
return result
def check_mfa_available() -> bool:
"""检查 MFA 是否可用"""
from src.mfa_runner import check_mfa_available as _check
return _check()
def cancel_current_session_tasks(session_id: Optional[str]) -> str:
"""取消当前会话的运行中任务"""
cancelled = TASK_MANAGER.cancel_session(session_id)
if cancelled <= 0:
return "⚠️ 当前会话没有可取消任务"
return f"🛑 已请求取消 {cancelled} 个任务"
# ==================== 制作音源功能 ====================
def get_audio_duration(file_path: str) -> Optional[float]:
"""
获取音频文件时长（秒）
返回: 时长秒数，失败返回 None
"""
try:
import wave
import contextlib
# 对于 WAV 文件，使用 wave 模块快速获取时长
if file_path.lower().endswith('.wav'):
with contextlib.closing(wave.open(file_path, 'r')) as f:
frames = f.getnframes()
rate = f.getframerate()
return frames / float(rate)
# 对于其他格式，使用 pydub（如果可用）
try:
from pydub import AudioSegment
audio = AudioSegment.from_file(file_path)
return len(audio) / 1000.0 # 毫秒转秒
except ImportError:
# pydub 不可用，尝试使用 librosa
try:
import librosa
duration = librosa.get_duration(path=file_path)
return duration
except ImportError:
logger.warning(f"无法获取音频时长，缺少 pydub 或 librosa: {file_path}")
return None
except Exception as e:
logger.warning(f"获取音频时长失败 {file_path}: {e}")
return None
# 云端音频时长限制（秒）
MAX_AUDIO_DURATION_SECONDS = 600 # 10分钟
def validate_audio_upload(files) -> Tuple[bool, str, List[str]]:
"""
验证上传的音频文件
返回: (是否有效, 消息, 文件路径列表)
"""
if not files:
return False, "请上传音频文件", []
valid_files = []
for f in files:
if hasattr(f, 'name'):
path = f.name
else:
path = str(f)
if path.lower().endswith(CloudConfig.AUDIO_EXTENSIONS):
valid_files.append(path)
if not valid_files:
return False, f"未找到有效音频文件，支持格式: {', '.join(CloudConfig.AUDIO_EXTENSIONS)}", []
return True, f"找到 {len(valid_files)} 个音频文件", valid_files
def validate_audio_duration(file_paths: List[str]) -> Tuple[bool, str, List[str]]:
"""
验证音频文件时长，过滤超时文件
返回: (是否全部通过, 消息, 有效文件列表)
"""
valid_files = []
rejected_files = []
max_minutes = MAX_AUDIO_DURATION_SECONDS / 60
for path in file_paths:
duration = get_audio_duration(path)
filename = os.path.basename(path)
if duration is None:
# 无法获取时长，允许通过（后续处理可能会失败）
valid_files.append(path)
logger.warning(f"无法检测时长，允许通过: {filename}")
elif duration > MAX_AUDIO_DURATION_SECONDS:
duration_min = duration / 60
rejected_files.append(f"{filename} ({duration_min:.1f}分钟)")
else:
valid_files.append(path)
if rejected_files:
if not valid_files:
# 全部被拒绝
return False, f"所有音频超过{max_minutes:.0f}分钟限制: {', '.join(rejected_files)}", []
else:
# 部分被拒绝
msg = f"已过滤 {len(rejected_files)} 个超时音频（>{max_minutes:.0f}分钟）: {', '.join(rejected_files[:3])}"
if len(rejected_files) > 3:
msg += f" 等{len(rejected_files)}个"
return True, msg, valid_files
return True, "", valid_files
@safe_gradio_handler
def process_make_voicebank(
audio_files,
source_name: str,
language: str,
whisper_model: str,
session_id: Optional[str] = None,
progress=gr.Progress()
) -> Tuple[str, str, Optional[str], Optional[str]]:
"""
制作音源：上传音频 → VAD切片 → Whisper转录 → MFA对齐 → 打包下载
返回: (状态, 日志, 下载文件路径, 会话存储的音源包路径)
"""
task = TASK_MANAGER.start_task("make", session_id)
if not task:
return f"❌ 服务繁忙，制作任务队列已达上限（{CloudConfig.MAX_MAKE_JOBS}）", "", None, None
logs = []
workspace = None
def log(msg):
logs.append(msg)
logger.info(msg)
TASK_MANAGER.touch(task.task_id)
def is_cancelled() -> bool:
return TASK_MANAGER.should_cancel(task.task_id)
try:
try:
# 导入依赖（放在 try 块内以捕获导入错误）
from src.pipeline import PipelineConfig, VoiceBankPipeline
except Exception as e:
logger.error(f"导入模块失败: {e}", exc_info=True)
return f"❌ 系统错误: 模块加载失败", str(e), None, None
# 验证输入
sanitized_name = sanitize_source_name(source_name)
if not sanitized_name:
return "❌ 请输入音源名称", "", None, None
source_name = sanitized_name
valid, msg, file_paths = validate_audio_upload(audio_files)
if not valid:
return f"❌ {msg}", "", None, None
log(f"📁 {msg}")
# 检查音频时长限制
valid, duration_msg, file_paths = validate_audio_duration(file_paths)
if not valid:
return f"❌ {duration_msg}", "", None, None
if duration_msg:
log(f"⚠️ {duration_msg}")
# 创建临时工作空间
workspace = create_temp_workspace()
log(f"🔧 创建工作空间: {workspace}")
if is_cancelled():
return "⚠️ 任务已取消", "任务在开始处理前被取消", None, None
# 准备输入目录
input_dir = os.path.join(workspace, "input")
bank_dir = os.path.join(workspace, "bank")
os.makedirs(input_dir, exist_ok=True)
os.makedirs(bank_dir, exist_ok=True)
# 复制音频文件到输入目录
progress(0.05, desc="复制音频文件...")
copied_count = 0
for src_path in file_paths:
if is_cancelled():
return "⚠️ 任务已取消", "\n".join(logs), None, None
# 检查源文件是否存在
if not os.path.exists(src_path):
log(f"⚠️ 文件不存在或已被清理: {src_path}")
continue
try:
dst_path = os.path.join(input_dir, os.path.basename(src_path))
shutil.copy2(src_path, dst_path)
copied_count += 1
except Exception as e:
log(f"⚠️ 复制文件失败 {os.path.basename(src_path)}: {e}")
if copied_count == 0:
return "❌ 无法访问上传的文件，请重新上传", "\n".join(logs), None, None
log(f"📋 已复制 {copied_count}/{len(file_paths)} 个文件到工作目录")
# 获取 MFA 模型路径
mfa_models = scan_mfa_models()
dict_path = None
acoustic_path = None
if mfa_models["dictionary"]:
# 根据语言选择字典
for d in mfa_models["dictionary"]:
if language == "japanese" and "japanese" in d.lower():
dict_path = os.path.join(CloudConfig.MFA_DIR, d)
break
elif language == "chinese" and "mandarin" in d.lower():
dict_path = os.path.join(CloudConfig.MFA_DIR, d)
break
if not dict_path:
dict_path = os.path.join(CloudConfig.MFA_DIR, mfa_models["dictionary"][0])
if mfa_models["acoustic"]:
for a in mfa_models["acoustic"]:
if language == "japanese" and "japanese" in a.lower():
acoustic_path = os.path.join(CloudConfig.MFA_DIR, a)
break
elif language == "chinese" and "mandarin" in a.lower():
acoustic_path = os.path.join(CloudConfig.MFA_DIR, a)
break
if not acoustic_path:
acoustic_path = os.path.join(CloudConfig.MFA_DIR, mfa_models["acoustic"][0])
# 配置流水线
whisper_model_name = CloudConfig.WHISPER_MODELS.get(whisper_model, "openai/whisper-small")
config = PipelineConfig(
source_name=source_name,
input_path=input_dir,
output_base_dir=bank_dir,
models_dir=CloudConfig.MODELS_DIR,
whisper_model=whisper_model_name,
mfa_dict_path=dict_path,
mfa_model_path=acoustic_path,
language=language,
cancel_checker=is_cancelled
)
pipeline = VoiceBankPipeline(config, log)
# 步骤0: VAD切片 + Whisper转录
progress(0.1, desc="VAD切片 + Whisper转录...")
log("\n" + "=" * 50)
log("【步骤1】VAD切片 + Whisper转录")
success, msg, slices = pipeline.step0_preprocess()
if not success:
return f"❌ 预处理失败: {msg}", "\n".join(logs), None, None
log(f"✅ {msg}")
if is_cancelled():
return "⚠️ 任务已取消", "\n".join(logs), None, None
# 步骤1: MFA对齐
progress(0.6, desc="MFA语音对齐...")
log("\n" + "=" * 50)
log("【步骤2】MFA语音对齐")
mfa_success = False
if check_mfa_available():
success, msg = pipeline.step1_mfa_align()
if not success:
log(f"⚠️ MFA对齐失败: {msg}")
log("继续导出（无TextGrid）...")
else:
log(f"✅ {msg}")
mfa_success = True
else:
log("⚠️ MFA不可用，跳过对齐步骤")
# MFA 不可用时的提示
if not mfa_success:
log("")
log("💡 提示：音源包中缺少 TextGrid 对齐数据")
log(" 如需导出 UTAU oto.ini，请前往「MFA补对齐」页面进行修复")
# 打包结果
progress(0.9, desc="打包结果...")
log("\n" + "=" * 50)
log("【打包结果】")
source_dir = os.path.join(bank_dir, source_name)
zip_name = f"{source_name}_音源数据"
zip_path = create_zip(source_dir, zip_name)
if zip_path:
log(f"📦 已打包: {os.path.basename(zip_path)}")
progress(1.0, desc="完成")
# 返回路径到会话状态，供导出页面使用
return "✅ 音源制作完成", "\n".join(logs), zip_path, zip_path
else:
return "❌ 打包失败", "\n".join(logs), None, None
except Exception as e:
logger.error(f"制作音源失败: {e}", exc_info=True)
return f"❌ 处理失败: {e}", "\n".join(logs), None, None
finally:
cleanup_workspace(workspace)
TASK_MANAGER.finish_task(task.task_id)
# ==================== 导出音源功能 ====================
def get_last_made_voicebank(session_voicebank: Optional[str]) -> Tuple[Optional[str], str]:
"""
获取当前会话制作的音源包
参数:
session_voicebank: 会话状态中存储的音源包路径
返回: (文件路径, 信息消息)
"""
if session_voicebank and os.path.exists(session_voicebank):
valid, msg, name = validate_voicebank_zip_path(session_voicebank)
if valid:
return session_voicebank, f"✅ 已选择刚制作的音源: {name}"
return None, "❌ 没有找到刚制作的音源，请先在「制作音源」页面制作，或手动上传"
def validate_voicebank_zip_path(zip_path: str) -> Tuple[bool, str, Optional[str]]:
"""
验证音源压缩包路径
返回: (是否有效, 消息, 音源名称)
"""
if not zip_path or not os.path.exists(zip_path):
return False, "文件不存在", None
if not zip_path.lower().endswith('.zip'):
return False, "请上传 .zip 格式的压缩包", None
try:
with zipfile.ZipFile(zip_path, 'r') as zf:
names = zf.namelist()
has_wav = any(n.endswith('.wav') for n in names)
has_slices = any('slices/' in n for n in names)
has_textgrid = any('textgrid/' in n for n in names)
if not has_wav:
return False, "压缩包中未找到 .wav 音频文件", None
source_name = None
if 'meta.json' in names:
try:
with zf.open('meta.json') as mf:
meta = json.load(mf)
source_name = meta.get('source_name')
except:
pass
if not source_name:
source_name = Path(zip_path).stem.replace('_音源数据', '')
source_name = sanitize_source_name(source_name) or "voicebank"
info_parts = []
if has_slices:
wav_count = len([n for n in names if 'slices/' in n and n.endswith('.wav')])
info_parts.append(f"切片: {wav_count} 个")
if has_textgrid:
tg_count = len([n for n in names if 'textgrid/' in n and n.endswith('.TextGrid')])
info_parts.append(f"TextGrid: {tg_count} 个")
info = " | ".join(info_parts) if info_parts else "有效的音源包"
return True, f"✅ {info}", source_name
except zipfile.BadZipFile:
return False, "无效的 zip 文件", None
except Exception as e:
return False, f"验证失败: {e}", None
def validate_voicebank_zip(zip_file) -> Tuple[bool, str, Optional[str]]:
"""
验证上传的音源压缩包
返回: (是否有效, 消息, 音源名称)
"""
if not zip_file:
return False, "请上传音源压缩包", None
zip_path = zip_file.name if hasattr(zip_file, 'name') else str(zip_file)
return validate_voicebank_zip_path(zip_path)
@safe_gradio_handler
def process_export_voicebank(
zip_file,
plugin_name: str,
options_json: str,
session_id: Optional[str] = None,
progress=gr.Progress()
) -> Tuple[str, str, Optional[str]]:
"""
导出音源：上传音源包 → 解压 → 导出 → 打包下载
参数:
zip_file: 上传的音源压缩包
plugin_name: 导出插件名称
options_json: JSON 格式的插件选项
返回: (状态, 日志, 下载文件路径)
"""
task = TASK_MANAGER.start_task("export", session_id)
if not task:
return f"❌ 服务繁忙，导出任务队列已达上限（{CloudConfig.MAX_EXPORT_JOBS}）", "", None
logs = []
workspace = None
def log(msg):
logs.append(msg)
logger.info(msg)
TASK_MANAGER.touch(task.task_id)
def is_cancelled() -> bool:
return TASK_MANAGER.should_cancel(task.task_id)
try:
# 验证输入
valid, msg, source_name = validate_voicebank_zip(zip_file)
if not valid:
return f"❌ {msg}", "", None
source_name = sanitize_source_name(source_name or "") or "voicebank"
log(f"📦 {msg}")
log(f"📝 音源名称: {source_name}")
# 解析选项
try:
options = json.loads(options_json) if options_json else {}
except json.JSONDecodeError:
options = {}
# 创建临时工作空间
workspace = create_temp_workspace()
log(f"🔧 创建工作空间")
if is_cancelled():
return "⚠️ 任务已取消", "任务在开始处理前被取消", None
zip_path = zip_file.name if hasattr(zip_file, 'name') else str(zip_file)
# 解压音源包
progress(0.1, desc="解压音源包...")
bank_dir = os.path.join(workspace, "bank")
source_dir = os.path.join(bank_dir, source_name)
os.makedirs(source_dir, exist_ok=True)
success, msg = extract_zip(zip_path, source_dir)
if not success:
return f"❌ {msg}", "\n".join(logs), None
log(f"📂 已解压到工作目录")
# 检查目录结构，处理可能的嵌套
slices_dir = os.path.join(source_dir, "slices")
if not os.path.exists(slices_dir):
# 可能解压后有额外的一层目录
subdirs = [d for d in os.listdir(source_dir) if os.path.isdir(os.path.join(source_dir, d))]
if len(subdirs) == 1:
nested_dir = os.path.join(source_dir, subdirs[0])
if os.path.exists(os.path.join(nested_dir, "slices")):
# 移动内容到上层
for item in os.listdir(nested_dir):
shutil.move(os.path.join(nested_dir, item), source_dir)
os.rmdir(nested_dir)
# 执行导出
progress(0.3, desc="执行导出...")
log("\n" + "=" * 50)
log(f"【{plugin_name}】")
from src.export_plugins import load_plugins
plugins = load_plugins()
if plugin_name not in plugins:
return f"❌ 未找到插件: {plugin_name}", "\n".join(logs), None
plugin = plugins[plugin_name]
plugin.set_progress_callback(log)
# 添加默认选项
options["clean_temp"] = True
success, msg = plugin.export(source_name, bank_dir, options)
if not success:
return f"❌ 导出失败: {msg}", "\n".join(logs), None
log(f"✅ {msg}")
# 打包导出结果
progress(0.9, desc="打包结果...")
log("\n" + "=" * 50)
log("【打包结果】")
# 根据插件类型确定导出目录和导出标识
if "UTAU" in plugin_name:
export_subdir = "utau_oto"
export_id = "utau_oto_export"
else:
export_subdir = "simple_export"
export_id = "simple_export"
export_dir = os.path.join(workspace, "export", source_name, export_subdir)
# 如果导出目录不存在，尝试其他位置
if not os.path.exists(export_dir):
alt_export = os.path.join(os.path.dirname(bank_dir), "export", source_name, export_subdir)
if os.path.exists(alt_export):
export_dir = alt_export
# 再尝试另一个子目录
if not os.path.exists(export_dir):
other_subdir = "simple_export" if export_subdir == "utau_oto" else "utau_oto"
other_id = "simple_export" if export_id == "utau_oto_export" else "utau_oto_export"
export_dir = os.path.join(workspace, "export", source_name, other_subdir)
if not os.path.exists(export_dir):
alt_export = os.path.join(os.path.dirname(bank_dir), "export", source_name, other_subdir)
if os.path.exists(alt_export):
export_dir = alt_export
export_id = other_id
else:
export_id = other_id
if not os.path.exists(export_dir):
return "❌ 未找到导出结果", "\n".join(logs), None
if is_cancelled():
return "⚠️ 任务已取消", "\n".join(logs), None
# 命名格式: [音源名称]_[插件标识]
zip_name = f"{source_name}_{export_id}"
result_zip = create_zip(export_dir, zip_name)
if result_zip:
# 统计导出文件数
file_count = len([f for f in os.listdir(export_dir) if f.endswith(('.wav', '.ini'))])
log(f"📦 已打包: {file_count} 个文件")
progress(1.0, desc="完成")
return "✅ 导出完成", "\n".join(logs), result_zip
else:
return "❌ 打包失败", "\n".join(logs), None
except Exception as e:
logger.error(f"导出失败: {e}", exc_info=True)
return f"❌ 处理失败: {e}", "\n".join(logs), None
finally:
cleanup_workspace(workspace)
TASK_MANAGER.finish_task(task.task_id)
# ==================== MFA补对齐功能 ====================
def validate_mfa_voicebank(zip_file) -> str:
"""
验证上传的音源包，检查 TextGrid 状态
返回: 状态信息字符串
"""
if not zip_file:
return "⏳ 请上传音源压缩包"
zip_path = zip_file.name if hasattr(zip_file, 'name') else str(zip_file)
if not zip_path.lower().endswith('.zip'):
return "❌ 请上传 .zip 格式的压缩包"
try:
with zipfile.ZipFile(zip_path, 'r') as zf:
names = zf.namelist()
# 统计文件
wav_files = [n for n in names if n.endswith('.wav') and 'slices/' in n]
lab_files = [n for n in names if n.endswith('.lab') and 'slices/' in n]
tg_files = [n for n in names if n.endswith('.TextGrid') and 'textgrid/' in n]
if not wav_files:
return "❌ 未找到 slices 目录下的 .wav 文件"
if not lab_files:
return "❌ 未找到 slices 目录下的 .lab 标注文件"
# 获取音源名称
source_name = None
if 'meta.json' in names:
try:
with zf.open('meta.json') as mf:
meta = json.load(mf)
source_name = meta.get('source_name')
except:
pass
if not source_name:
source_name = Path(zip_path).stem.replace('_音源数据', '')
# 构建状态信息
info_parts = [f"📝 音源: {source_name}"]
info_parts.append(f"🎵 切片: {len(wav_files)} 个 WAV")
info_parts.append(f"📄 标注: {len(lab_files)} 个 LAB")
if tg_files:
coverage = len(tg_files) / len(wav_files) * 100 if wav_files else 0
if coverage >= 100:
info_parts.append(f"✅ TextGrid: {len(tg_files)} 个 (已完整)")
else:
info_parts.append(f"⚠️ TextGrid: {len(tg_files)} 个 ({coverage:.0f}% 覆盖)")
else:
info_parts.append("❌ TextGrid: 无 (需要对齐)")
return " | ".join(info_parts)
except zipfile.BadZipFile:
return "❌ 无效的 zip 文件"
except Exception as e:
return f"❌ 验证失败: {e}"
@safe_gradio_handler
def process_mfa_realign(
zip_file,
language: str,
session_id: Optional[str] = None,
progress=gr.Progress()
) -> Tuple[str, str, Optional[str]]:
"""
MFA 补对齐：为缺少 TextGrid 的音源包执行 MFA 对齐
参数:
zip_file: 上传的音源压缩包
language: 语言选择
返回: (状态, 日志, 下载文件路径)
"""
task = TASK_MANAGER.start_task("mfa", session_id)
if not task:
return f"❌ 服务繁忙，对齐任务队列已达上限（{CloudConfig.MAX_MFA_JOBS}）", "", None
logs = []
workspace = None
def log(msg):
logs.append(msg)
logger.info(msg)
TASK_MANAGER.touch(task.task_id)
def is_cancelled() -> bool:
return TASK_MANAGER.should_cancel(task.task_id)
try:
# 验证输入
if not zip_file:
return "❌ 请上传音源压缩包", "", None
zip_path = zip_file.name if hasattr(zip_file, 'name') else str(zip_file)
# 检查 MFA 是否可用
if not check_mfa_available():
return "❌ MFA 不可用，无法执行对齐", "请检查 MFA 环境配置", None
log("📦 开始处理音源包...")
# 创建临时工作空间
workspace = create_temp_workspace()
log(f"🔧 创建工作空间")
if is_cancelled():
return "⚠️ 任务已取消", "任务在开始处理前被取消", None
# 解压音源包
progress(0.1, desc="解压音源包...")
# 先检查压缩包结构，确定音源名称
source_name = None
with zipfile.ZipFile(zip_path, 'r') as zf:
names = zf.namelist()
# 尝试从 meta.json 获取音源名称
if 'meta.json' in names:
try:
with zf.open('meta.json') as mf:
meta = json.load(mf)
source_name = meta.get('source_name')
except:
pass
if not source_name:
source_name = Path(zip_path).stem.replace('_音源数据', '')
source_name = sanitize_source_name(source_name) or "voicebank"
log(f"📝 音源名称: {source_name}")
# 解压到工作目录
bank_dir = os.path.join(workspace, "bank")
source_dir = os.path.join(bank_dir, source_name)
os.makedirs(source_dir, exist_ok=True)
success, msg = extract_zip(zip_path, source_dir)
if not success:
return f"❌ {msg}", "\n".join(logs), None
log(f"📂 已解压到工作目录")
# 检查目录结构，处理可能的嵌套
slices_dir = os.path.join(source_dir, "slices")
if not os.path.exists(slices_dir):
# 可能解压后有额外的一层目录
subdirs = [d for d in os.listdir(source_dir) if os.path.isdir(os.path.join(source_dir, d))]
if len(subdirs) == 1:
nested_dir = os.path.join(source_dir, subdirs[0])
if os.path.exists(os.path.join(nested_dir, "slices")):
# 移动内容到上层
for item in os.listdir(nested_dir):
shutil.move(os.path.join(nested_dir, item), source_dir)
os.rmdir(nested_dir)
slices_dir = os.path.join(source_dir, "slices")
if not os.path.exists(slices_dir):
return "❌ 未找到 slices 目录", "\n".join(logs), None
# 统计文件
wav_files = [f for f in os.listdir(slices_dir) if f.endswith('.wav')]
lab_files = [f for f in os.listdir(slices_dir) if f.endswith('.lab')]
log(f"🎵 找到 {len(wav_files)} 个 WAV 文件")
log(f"📄 找到 {len(lab_files)} 个 LAB 标注文件")
if not wav_files:
return "❌ slices 目录中没有 WAV 文件", "\n".join(logs), None
if not lab_files:
return "❌ slices 目录中没有 LAB 标注文件", "\n".join(logs), None
# 获取 MFA 模型路径
progress(0.2, desc="加载 MFA 模型...")
mfa_models = scan_mfa_models()
dict_path = None
acoustic_path = None
if mfa_models["dictionary"]:
for d in mfa_models["dictionary"]:
if language == "japanese" and "japanese" in d.lower():
dict_path = os.path.join(CloudConfig.MFA_DIR, d)
break
elif language == "chinese" and "mandarin" in d.lower():
dict_path = os.path.join(CloudConfig.MFA_DIR, d)
break
if not dict_path:
dict_path = os.path.join(CloudConfig.MFA_DIR, mfa_models["dictionary"][0])
if mfa_models["acoustic"]:
for a in mfa_models["acoustic"]:
if language == "japanese" and "japanese" in a.lower():
acoustic_path = os.path.join(CloudConfig.MFA_DIR, a)
break
elif language == "chinese" and "mandarin" in a.lower():
acoustic_path = os.path.join(CloudConfig.MFA_DIR, a)
break
if not acoustic_path:
acoustic_path = os.path.join(CloudConfig.MFA_DIR, mfa_models["acoustic"][0])
if not dict_path or not os.path.exists(dict_path):
return f"❌ 未找到 {language} 语言的字典文件", "\n".join(logs), None
if not acoustic_path or not os.path.exists(acoustic_path):
return f"❌ 未找到 {language} 语言的声学模型", "\n".join(logs), None
log(f"📚 字典: {os.path.basename(dict_path)}")
log(f"🔊 声学模型: {os.path.basename(acoustic_path)}")
# 执行 MFA 对齐
progress(0.3, desc="执行 MFA 对齐...")
log("\n" + "=" * 50)
log("【MFA 强制对齐】")
textgrid_dir = os.path.join(source_dir, "textgrid")
os.makedirs(textgrid_dir, exist_ok=True)
from src.mfa_runner import run_mfa_alignment
success, mfa_msg = run_mfa_alignment(
corpus_dir=slices_dir,
output_dir=textgrid_dir,
dict_path=dict_path,
model_path=acoustic_path,
single_speaker=True,
clean=True,
progress_callback=log,
cancel_checker=is_cancelled
)
if not success:
log(f"❌ MFA 对齐失败: {mfa_msg}")
return "❌ MFA 对齐失败", "\n".join(logs), None
log("✅ MFA 对齐完成")
# 统计生成的 TextGrid 文件
tg_files = [f for f in os.listdir(textgrid_dir) if f.endswith('.TextGrid')]
log(f"📊 生成 {len(tg_files)} 个 TextGrid 文件")
if not tg_files:
return "❌ 未生成任何 TextGrid 文件", "\n".join(logs), None
# 更新 meta.json
meta_path = os.path.join(source_dir, "meta.json")
if os.path.exists(meta_path):
try:
with open(meta_path, 'r', encoding='utf-8') as f:
meta = json.load(f)
meta['textgrid_count'] = len(tg_files)
meta['mfa_realigned'] = True
with open(meta_path, 'w', encoding='utf-8') as f:
json.dump(meta, f, ensure_ascii=False, indent=2)
log("📝 已更新 meta.json")
except Exception as e:
log(f"⚠️ 更新 meta.json 失败: {e}")
# 打包结果
progress(0.9, desc="打包结果...")
log("\n" + "=" * 50)
log("【打包结果】")
zip_name = f"{source_name}_音源数据_已对齐"
result_zip = create_zip(source_dir, zip_name)
if result_zip:
log(f"📦 已打包: {os.path.basename(result_zip)}")
progress(1.0, desc="完成")
return "✅ MFA 补对齐完成", "\n".join(logs), result_zip
else:
return "❌ 打包失败", "\n".join(logs), None
except Exception as e:
logger.error(f"MFA 补对齐失败: {e}", exc_info=True)
return f"❌ 处理失败: {e}", "\n".join(logs), None
finally:
cleanup_workspace(workspace)
TASK_MANAGER.finish_task(task.task_id)
# ==================== 插件选项 UI 生成 ====================
def get_plugin_options_config(plugins: Dict[str, Any]) -> Dict[str, List[Dict]]:
"""
获取所有插件的选项配置
返回: {插件名: [选项配置列表]}
"""
from src.export_plugins.base import OptionType
config = {}
for name, plugin in plugins.items():
options = []
for opt in plugin.get_options():
opt_config = {
"key": opt.key,
"label": opt.label,
"type": opt.option_type.value,
"default": opt.default,
"description": opt.description,
"choices": opt.choices,
"min_value": opt.min_value,
"max_value": opt.max_value,
"step": opt.step,
}
options.append(opt_config)
config[name] = options
return config
def get_default_options_json(plugin_name: str, plugins_config: Dict) -> str:
"""获取插件的默认选项 JSON"""
if plugin_name not in plugins_config:
return "{}"
options = plugins_config[plugin_name]
defaults = {}
for opt in options:
if opt["type"] != "label":
defaults[opt["key"]] = opt["default"]
return json.dumps(defaults, ensure_ascii=False)
def create_dynamic_plugin_options(plugins: Dict[str, Any], plugins_config: Dict) -> Tuple[Dict[str, Any], Dict[str, Any]]:
"""
创建动态插件选项组件
返回:
(组件字典, 收集选项函数)
组件字典结构: {
"container": gr.Column, # 主容器
"groups": {插件名: gr.Group}, # 每个插件的选项组
"components": {插件名: {选项key: 组件}}, # 所有组件
}
"""
from src.export_plugins.base import OptionType
all_groups = {}
all_components = {}
# 为每个插件创建选项组
for plugin_name, options in plugins_config.items():
plugin_components = {}
# 创建该插件的选项组（初始隐藏，第一个插件除外）
is_first = (plugin_name == list(plugins_config.keys())[0])
with gr.Group(visible=is_first) as plugin_group:
# 显示插件描述
if plugin_name in plugins:
gr.Markdown(f"> {plugins[plugin_name].description}")
for opt in options:
opt_type = opt["type"]
key = opt["key"]
label = opt["label"]
default = opt["default"]
description = opt.get("description", "")
choices = opt.get("choices", [])
min_val = opt.get("min_value")
max_val = opt.get("max_value")
step = opt.get("step")
# 根据类型创建对应的 Gradio 组件
if opt_type == "label":
# 纯文本标签
gr.Markdown(f"*{label}*")
continue
elif opt_type == "text":
component = gr.Textbox(
label=label,
value=default or "",
info=description
)
elif opt_type == "number":
component = gr.Number(
label=label,
value=default if default is not None else 0,
minimum=min_val,
maximum=max_val,
step=step or 1,
info=description
)
elif opt_type == "switch":
component = gr.Checkbox(
label=label,
value=bool(default),
info=description
)
elif opt_type == "combo":
component = gr.Dropdown(
label=label,
choices=choices,
value=default if default in choices else (choices[0] if choices else None),
info=description
)
elif opt_type == "multi_select":
component = gr.CheckboxGroup(
label=label,
choices=choices,
value=default if isinstance(default, list) else [],
info=description
)
else:
# 未知类型，使用文本框
component = gr.Textbox(
label=label,
value=str(default) if default else "",
info=description
)
plugin_components[key] = component
all_groups[plugin_name] = plugin_group
all_components[plugin_name] = plugin_components
return all_groups, all_components
def build_options_collector(plugins_config: Dict, all_components: Dict):
"""
构建选项收集函数
返回一个函数，该函数接收插件名和所有组件值，返回选项字典
"""
# 构建组件到选项的映射
component_keys = {}
for plugin_name, components in all_components.items():
component_keys[plugin_name] = list(components.keys())
def collect_options(plugin_name: str, *values) -> Dict[str, Any]:
"""收集当前插件的选项值"""
if plugin_name not in component_keys:
return {}
keys = component_keys[plugin_name]
options = {}
# 计算当前插件的值在 values 中的起始位置
start_idx = 0
for pname in component_keys:
if pname == plugin_name:
break
start_idx += len(component_keys[pname])
# 提取当前插件的值
for i, key in enumerate(keys):
if start_idx + i < len(values):
options[key] = values[start_idx + i]
return options
return collect_options
# ==================== 构建界面 ====================
def create_cloud_ui():
"""创建云端 Gradio 界面"""
# 检查 MFA 状态
mfa_available = check_mfa_available()
mfa_status = "✅ MFA 已就绪" if mfa_available else "⚠️ MFA 不可用（将跳过对齐步骤）"
# 加载导出插件
from src.export_plugins import load_plugins
plugins = load_plugins()
plugin_names = list(plugins.keys()) if plugins else ["简单单字导出"]
with gr.Blocks(
title="人力V助手 (JinrikiHelper)",
theme=gr.themes.Soft()
) as app:
session_id_state = gr.State(value=uuid.uuid4().hex[:8])
# 会话状态：存储当前用户制作的音源包路径
session_voicebank = gr.State(value=None)
# 标题行：左侧标题 + 右侧并发状态
with gr.Row():
with gr.Column(scale=4):
gr.Markdown("# 🎤 人力V助手 (JinrikiHelper)")
with gr.Column(scale=1, min_width=200):
concurrency_display = gr.Markdown(
value=get_concurrency_status(),
elem_id="concurrency-status"
)
concurrency_timer = gr.Timer(2.0)
with gr.Row():
cancel_jobs_btn = gr.Button("🛑 取消当前会话任务", variant="stop")
cancel_jobs_msg = gr.Textbox(label="任务控制", interactive=False, value="")
gr.Markdown("语音数据集处理工具 - 自动化制作语音音源库")
gr.Markdown("> ☁️ 云端版：上传音频 → 自动处理 → 下载结果")
gr.Markdown("""
<div style="background: linear-gradient(135deg, #d4edda 0%, #c3e6cb 100%); border-left: 4px solid #28a745; padding: 10px 14px; border-radius: 8px; margin: 8px 0;">
<p style="margin: 0; color: #155724; line-height: 1.5;">
<strong>📊 处理状态说明：</strong>只有当<strong>进度条弹出并开始滚动</strong>时，才表示正在处理中。<br/>
如果点击按钮后长时间没有进度条出现，可能是正在排队等待，或遇到了问题。
</p>
</div>
""")
with gr.Tabs():
# ==================== 制作音源页 ====================
with gr.Tab("🎵 制作音源"):
gr.Markdown("""
<div style="background: linear-gradient(135deg, #fff3cd 0%, #ffeeba 100%); border-left: 4px solid #ffc107; padding: 12px 16px; border-radius: 8px; margin-bottom: 16px;">
<p style="margin: 0 0 8px 0; font-weight: bold; color: #856404;">⚠️ 温馨提示</p>
<p style="margin: 0; color: #856404; line-height: 1.6;">
<strong style="font-size: 1.1em;">音频质量 >> 音频数量！</strong><br/>
请控制上传音频的数量！经测试，<strong>8 分钟以内的高质量音频已经非常充足</strong>。<br/>
上传过多音频可能导致混入低质量样本，同时也会占用服务器并发资源。<br/>
建议大量音频先人工筛选后再上传，感谢配合！🙏
</p>
</div>
""")
gr.Markdown("### 上传音频文件")
gr.Markdown("支持格式: WAV, MP3, FLAC, OGG, M4A")
gr.Markdown("允许同时拖拽多个文件上传，也可点击上传框的右上角追加文件")
audio_upload = gr.File(
label="上传音频文件",
file_count="multiple",
file_types=["audio"]
)
# 上传状态提示
upload_status = gr.Textbox(
label="上传状态",
value="⏳ 请上传音频文件",
interactive=False
)
with gr.Row():
make_source_name = gr.Textbox(
label="音源名称",
placeholder="my_voice",
info="用于标识输出的音源包"
)
make_language = gr.Dropdown(
choices=CloudConfig.LANGUAGES,
value="chinese",
label="语言"
)
with gr.Row():
make_whisper = gr.Dropdown(
choices=list(CloudConfig.WHISPER_MODELS.keys()),
value=list(CloudConfig.WHISPER_MODELS.keys())[0],
label="Whisper 模型"
)
make_mfa_status = gr.Textbox(
label="MFA 状态",
value=mfa_status,
interactive=False
)
gr.Markdown("""
> ⏱️ **模型速度参考**：small 约 4 秒/句，medium 约 12 秒/句（medium 慢 2-3 倍但更准确）
>
> <span style="color: red;">**small完全够用的，medium费时还容易炸空间，除非实在识别不出来字再用**</span>
""")
make_btn = gr.Button("🚀 开始制作", variant="primary", size="lg", interactive=False)
# 时长估算显示
time_estimate = gr.Markdown(
value="",
visible=False
)
make_status = gr.Textbox(label="状态", interactive=False)
make_log = gr.Textbox(label="处理日志", lines=12, interactive=False)
gr.Markdown("### 下载结果")
make_download = gr.File(label="音源包下载", interactive=False)
gr.Markdown("""
> 💡 处理流程：
> 1. VAD 语音活动检测，自动切分音频
> 2. Whisper 语音识别，生成文本标注
> 3. MFA 强制对齐，生成音素级时间标注
> 4. 打包为 zip 供下载
""")
# 音频上传状态检测
def check_audio_upload(files):
"""检查音频上传状态，返回状态文本、按钮可用性和时长估算"""
if not files:
return "⏳ 请上传音频文件", gr.update(interactive=False), gr.update(value="", visible=False)
valid_count = 0
total_duration = 0.0
for f in files:
path = f.name if hasattr(f, 'name') else str(f)
if path.lower().endswith(CloudConfig.AUDIO_EXTENSIONS):
valid_count += 1
# 计算时长
duration = get_audio_duration(path)
if duration:
total_duration += duration
if valid_count == 0:
return f"❌ 未找到有效音频，支持: {', '.join(CloudConfig.AUDIO_EXTENSIONS)}", gr.update(interactive=False), gr.update(value="", visible=False)
# 格式化总时长
total_minutes = int(total_duration // 60)
total_seconds = int(total_duration % 60)
duration_str = f"{total_minutes}分{total_seconds}秒" if total_minutes > 0 else f"{total_seconds}秒"
# 计算预估处理时间
# 根据实测数据：1分钟音频约产生79个切片，每个切片处理约3.9秒
# 即每分钟音频需要约 79 * 3.9 / 60 ≈ 5.1 分钟处理时间
PROCESS_TIME_RATIO = 5.1 # 处理时间与音频时长的比例
estimated_seconds = total_duration * PROCESS_TIME_RATIO
est_minutes = int(estimated_seconds // 60)
est_seconds = int(estimated_seconds % 60)
if est_minutes >= 60:
est_hours = est_minutes // 60
est_minutes = est_minutes % 60
estimate_str = f"{est_hours}小时{est_minutes}分钟"
elif est_minutes > 0:
estimate_str = f"{est_minutes}分{est_seconds}秒"
else:
estimate_str = f"{est_seconds}秒"
estimate_md = f"> ⏱️ **预估处理时间**：约 {estimate_str}（基于 small 模型，medium 约为 2-3 倍）"
# 根据时长给出不同提示
if total_duration > MAX_AUDIO_DURATION_SECONDS:
return f"⚠️ 已上传 {valid_count} 个音频，总时长 {duration_str}（超过10分钟，部分文件将被过滤）", gr.update(interactive=True), gr.update(value=estimate_md, visible=True)
elif total_duration > 480: # 8分钟
return f"⚠️ 已上传 {valid_count} 个音频，总时长 {duration_str}（建议控制在8分钟内）", gr.update(interactive=True), gr.update(value=estimate_md, visible=True)
else:
return f"✅ 已上传 {valid_count} 个音频，总时长 {duration_str}", gr.update(interactive=True), gr.update(value=estimate_md, visible=True)
audio_upload.change(
fn=check_audio_upload,
inputs=[audio_upload],
outputs=[upload_status, make_btn, time_estimate]
)
make_btn.click(
fn=process_make_voicebank,
inputs=[audio_upload, make_source_name, make_language, make_whisper, session_id_state],
outputs=[make_status, make_log, make_download, session_voicebank],
concurrency_limit=CloudConfig.MAX_MAKE_JOBS
)
# ==================== 导出音源页 ====================
with gr.Tab("📤 导出音源"):
gr.Markdown("### 选择音源包")
# 使用刚制作的音源按钮
use_last_btn = gr.Button("📦 使用刚制作的音源", variant="secondary")
gr.Markdown("或者上传之前制作的音源压缩包（包含 slices 和 textgrid 目录）")
export_upload = gr.File(
label="上传音源包 (.zip)",
file_types=[".zip"]
)
export_info = gr.Textbox(
label="音源信息",
interactive=False,
placeholder="上传后显示音源信息"
)
# 上传后自动验证
def on_upload(file):
if file:
valid, msg, name = validate_voicebank_zip(file)
return msg
return ""
export_upload.change(
fn=on_upload,
inputs=[export_upload],
outputs=[export_info]
)
# 使用刚制作的音源
use_last_btn.click(
fn=get_last_made_voicebank,
inputs=[session_voicebank],
outputs=[export_upload, export_info]
)
gr.Markdown("---")
gr.Markdown("### 导出设置")
# 获取插件选项配置
plugins_config = get_plugin_options_config(plugins)
export_plugin = gr.Dropdown(
choices=plugin_names,
value=plugin_names[0] if plugin_names else None,
label="导出插件"
)
# ===== 动态选项区域 =====
# 为每个插件动态创建选项组件
all_plugin_groups = {}
all_plugin_components = {}
for idx, (pname, poptions) in enumerate(plugins_config.items()):
is_first = (idx == 0)
plugin_components = {}
with gr.Group(visible=is_first) as plugin_group:
# 插件描述
if pname in plugins:
gr.Markdown(f"> {plugins[pname].description}")
# 动态创建选项组件
for opt in poptions:
opt_type = opt["type"]
key = opt["key"]
label = opt["label"]
default = opt["default"]
description = opt.get("description", "")
choices = opt.get("choices", [])
min_val = opt.get("min_value")
max_val = opt.get("max_value")
step = opt.get("step")
if opt_type == "label":
gr.Markdown(f"*{label}*")
continue
elif opt_type == "text":
component = gr.Textbox(
label=label,
value=default or "",
info=description
)
elif opt_type == "number":
component = gr.Number(
label=label,
value=default if default is not None else 0,
minimum=min_val,
maximum=max_val,
step=step or 1,
info=description
)
elif opt_type == "switch":
component = gr.Checkbox(
label=label,
value=bool(default),
info=description
)
elif opt_type == "combo":
component = gr.Dropdown(
label=label,
choices=choices,
value=default if default in choices else (choices[0] if choices else None),
info=description
)
elif opt_type == "multi_select":
component = gr.CheckboxGroup(
label=label,
choices=choices,
value=default if isinstance(default, list) else [],
info=description
)
else:
component = gr.Textbox(
label=label,
value=str(default) if default else "",
info=description
)
plugin_components[key] = component
all_plugin_groups[pname] = plugin_group
all_plugin_components[pname] = plugin_components
# 插件切换时更新选项组可见性
def on_plugin_change(selected_plugin):
"""切换插件时更新选项区域可见性"""
updates = []
for pname in plugins_config.keys():
updates.append(gr.update(visible=(pname == selected_plugin)))
return updates
# 绑定插件切换事件
export_plugin.change(
fn=on_plugin_change,
inputs=[export_plugin],
outputs=list(all_plugin_groups.values())
)
# 收集选项并导出
def collect_and_export(zip_file, plugin_name, session_id, *all_values, progress=gr.Progress()):
"""收集当前插件的选项并执行导出"""
# 根据插件名找到对应的选项配置
if plugin_name not in plugins_config:
return "❌ 未找到插件配置", "", None
# 计算当前插件的值在 all_values 中的位置
start_idx = 0
for pname in plugins_config.keys():
if pname == plugin_name:
break
# 统计该插件的非 label 选项数量
start_idx += sum(1 for opt in plugins_config[pname] if opt["type"] != "label")
# 提取当前插件的选项值
options = {}
current_idx = start_idx
for opt in plugins_config[plugin_name]:
if opt["type"] == "label":
continue
key = opt["key"]
if current_idx < len(all_values):
value = all_values[current_idx]
# 类型转换
if opt["type"] == "number":
value = float(value) if value is not None else opt["default"]
options[key] = value
current_idx += 1
options_json = json.dumps(options, ensure_ascii=False)
return process_export_voicebank(zip_file, plugin_name, options_json, session_id, progress)
export_btn = gr.Button("📤 开始导出", variant="primary", size="lg")
export_status = gr.Textbox(label="状态", interactive=False)
export_log = gr.Textbox(label="处理日志", lines=10, interactive=False)
gr.Markdown("### 下载结果")
export_download = gr.File(label="导出结果下载", interactive=False)
gr.Markdown("""
> 💡 导出说明：
> - 从 TextGrid 提取每个汉字/音节的时间边界
> - 按拼音/罗马音分类，选取最佳样本
> - 导出为适配其他软件的音源格式
""")
# 收集所有插件的所有组件作为输入
all_option_components = []
for pname in plugins_config.keys():
if pname in all_plugin_components:
for opt in plugins_config[pname]:
if opt["type"] != "label" and opt["key"] in all_plugin_components[pname]:
all_option_components.append(all_plugin_components[pname][opt["key"]])
export_btn.click(
fn=collect_and_export,
inputs=[export_upload, export_plugin, session_id_state] + all_option_components,
outputs=[export_status, export_log, export_download],
concurrency_limit=CloudConfig.MAX_EXPORT_JOBS
)
# ==================== MFA补对齐页 ====================
with gr.Tab("🔧 MFA补对齐"):
gr.Markdown("""
## MFA 补对齐
<div style="background: linear-gradient(135deg, #cce5ff 0%, #b8daff 100%); border-left: 4px solid #004085; padding: 12px 16px; border-radius: 8px; margin-bottom: 16px;">
<p style="margin: 0 0 8px 0; font-weight: bold; color: #004085;">ℹ️ 功能说明</p>
<p style="margin: 0; color: #004085; line-height: 1.6;">
如果在「制作音源」时 MFA 不可用或对齐失败，音源包中将缺少 TextGrid 文件。<br/>
此页面可以为已有的音源包补充 MFA 对齐数据，以便后续导出 UTAU oto.ini 等格式。
</p>
</div>
""")
gr.Markdown("### 上传音源包")
gr.Markdown("上传包含 `slices` 目录（.wav + .lab 文件）的音源压缩包")
mfa_upload = gr.File(
label="上传音源包 (.zip)",
file_types=[".zip"]
)
mfa_info = gr.Textbox(
label="音源信息",
interactive=False,
placeholder="上传后显示音源信息和 TextGrid 状态"
)
# 上传后自动验证
mfa_upload.change(
fn=validate_mfa_voicebank,
inputs=[mfa_upload],
outputs=[mfa_info]
)
gr.Markdown("---")
gr.Markdown("### 对齐设置")
with gr.Row():
mfa_language = gr.Dropdown(
choices=CloudConfig.LANGUAGES,
value="chinese",
label="语言",
info="选择与音源匹配的语言"
)
mfa_status_display = gr.Textbox(
label="MFA 状态",
value=mfa_status,
interactive=False
)
mfa_btn = gr.Button("🔧 开始对齐", variant="primary", size="lg")
mfa_process_status = gr.Textbox(label="状态", interactive=False)
mfa_log = gr.Textbox(label="处理日志", lines=10, interactive=False)
gr.Markdown("### 下载结果")
mfa_download = gr.File(label="补齐后的音源包下载", interactive=False)
gr.Markdown("""
> 💡 **处理流程**：
> 1. 解压上传的音源包
> 2. 检测 slices 目录中的 .wav 和 .lab 文件
> 3. 执行 MFA 强制对齐，生成 TextGrid 文件
> 4. 打包完整音源包供下载
""")
mfa_btn.click(
fn=process_mfa_realign,
inputs=[mfa_upload, mfa_language, session_id_state],
outputs=[mfa_process_status, mfa_log, mfa_download],
concurrency_limit=CloudConfig.MAX_MFA_JOBS
)
# ==================== 关于页 ====================
with gr.Tab("ℹ️ 关于"):
gr.Markdown("""
## 人力V助手 (JinrikiHelper)
语音数据集处理工具，用于自动化制作语音音源库。
### 功能特点
- **VAD 切片**: 使用 Silero VAD 自动检测语音片段
- **语音识别**: 使用 Whisper 模型转录文本
- **强制对齐**: 使用 MFA 生成音素级时间标注
- **智能导出**: 按拼音分类，选取最佳样本
### 支持语言
- 中文（普通话）
- 日语
### 使用流程
1. **制作音源**: 上传原始音频 → 自动处理 → 下载音源包
2. **导出音源**: 上传音源包 → 选择导出格式 → 下载导出结果
---
**作者**: TNOT | **协议**: MIT
本工具集成 Montreal Forced Aligner (MIT License)
""")
cancel_jobs_btn.click(
fn=cancel_current_session_tasks,
inputs=[session_id_state],
outputs=[cancel_jobs_msg]
)
# 页面加载时刷新并发状态
app.load(
fn=get_concurrency_status,
outputs=[concurrency_display]
)
concurrency_timer.tick(
fn=get_concurrency_status,
outputs=[concurrency_display]
)
return app
def main():
"""云端入口"""
app = create_cloud_ui()
# 启用队列，实际并发由 CloudTaskManager + 事件并发上限共同控制
app.queue()
app.launch(
server_name="0.0.0.0",
server_port=7860,
share=False,
show_error=True
)
if __name__ == "__main__":
main()
|