Spaces:
Runtime error
Runtime error
File size: 124,646 Bytes
4482b40 06adbe1 4482b40 e6ec5b6 303a80b d73459b e93b0ba 4482b40 f48943e 43037cf 4482b40 e6ec5b6 829ae99 f4cda3f 43037cf f4cda3f 43037cf 06adbe1 43037cf f4cda3f e2928bf 4482b40 06adbe1 4b3b034 e2928bf f4cda3f e93b0ba f4cda3f 43037cf e6ec5b6 b326353 e6ec5b6 b326353 e6ec5b6 b326353 4482b40 f4cda3f e2928bf e6ec5b6 e2928bf cc30771 e93b0ba f4cda3f 06adbe1 2a41399 f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f e93b0ba f4cda3f 06adbe1 f4cda3f dd12997 f4cda3f 303a80b f4cda3f 303a80b f4cda3f 303a80b f4cda3f 303a80b f4cda3f af05e7c f4cda3f af05e7c f4cda3f af05e7c f4cda3f af05e7c f4cda3f af05e7c f4cda3f 66507ca f4cda3f f48943e f4cda3f 66507ca f4cda3f f48943e f4cda3f 66507ca f4cda3f f48943e 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 66507ca f4cda3f f48943e f4cda3f 8d920b7 f4cda3f 66507ca 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f f48943e f4cda3f 66507ca f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f f48943e f4cda3f 66507ca f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 66507ca f4cda3f 8d920b7 f4cda3f 66507ca 8d920b7 f4cda3f 8d920b7 f48943e f4cda3f 8d920b7 66507ca 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 303a80b 8d920b7 303a80b 8d920b7 f4cda3f 66507ca 8d920b7 f48943e 8d920b7 66507ca f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f f48943e f4cda3f 8d920b7 f4cda3f 8d920b7 f4cda3f 66507ca f4cda3f 66507ca f4cda3f 66507ca f4cda3f 8d920b7 f4cda3f f48943e f4cda3f 8d920b7 f4cda3f 303a80b f4cda3f 8d920b7 303a80b f4cda3f 303a80b f4cda3f 303a80b f4cda3f 303a80b 8d920b7 f4cda3f 8d920b7 f4cda3f 66507ca f4cda3f 66507ca f4cda3f 66507ca 8d920b7 66507ca f4cda3f 8d920b7 f4cda3f 8d920b7 66507ca f4cda3f 66507ca 8d920b7 f4cda3f |
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 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 |
import io
import os
import re
import time
import requests
from typing import Any, Dict, List, Optional, Set, Union
from difflib import get_close_matches
from pathlib import Path
from itertools import islice
from functools import partial
from multiprocessing.pool import ThreadPool
from queue import Queue, Empty
from typing import Callable, Iterable, Iterator, Optional, TypeVar
import gradio as gr
import pandas as pd
import requests.exceptions
from huggingface_hub import InferenceClient, create_repo, DatasetCard
from huggingface_hub.utils import HfHubHTTPError
import json
# --- Configuration ---
model_id = "microsoft/Phi-3-mini-4k-instruct"
client = InferenceClient(model_id)
save_dataset_hf_token = os.environ.get("SAVE_DATASET_HF_TOKEN")
MAX_TOTAL_NB_ITEMS = 100
MAX_NB_ITEMS_PER_GENERATION_CALL = 10
NUM_ROWS = 100
NUM_VARIANTS = 10
NAMESPACE = "infinite-dataset-hub"
URL = "https://huggingface.co/spaces/infinite-dataset-hub/infinite-dataset-hub"
# --- Prompt Templates ---
GENERATE_DATASET_NAMES_FOR_SEARCH_QUERY = (
"A Machine Learning Practioner is looking for a dataset that matches '{search_query}'. "
f"Generate a list of {MAX_NB_ITEMS_PER_GENERATION_CALL} names of quality datasets that don't exist but sound plausible and would "
"be helpful. Feel free to reuse words from the query '{search_query}' to name the datasets. "
"Every dataset should be about '{search_query}' and have descriptive tags/keywords including the ML task name associated with the dataset (classification, regression, anomaly detection, etc.). Use the following format:\n1. DatasetName1 (tag1, tag2, tag3)\n2. DatasetName2 (tag1, tag2, tag3)"
)
GENERATE_DATASET_CONTENT_FOR_SEARCH_QUERY_AND_NAME_AND_TAGS = (
"An ML practitioner is looking for a dataset CSV after the query '{search_query}'. "
"Generate the first 5 rows of a plausible and quality CSV for the dataset '{dataset_name}'. "
"You can get inspiration from related keywords '{tags}' but most importantly the dataset should correspond to the query '{search_query}'. "
"Focus on quality text content and use a 'label' or 'labels' column if it makes sense (invent labels, avoid reusing the keywords, be accurate while labelling texts). "
"Reply using a short description of the dataset with title **Dataset Description:** followed by the CSV content in a code block and with title **CSV Content Preview:**."
)
GENERATE_MORE_ROWS = "Can you give me 10 additional samples in CSV format as well? Use the same CSV header '{csv_header}'."
GENERATE_VARIANTS_WITH_RARITY_AND_LABEL = "Focus on generating samples for the label '{label}' and ideally generate {rarity} samples."
GENERATE_VARIANTS_WITH_RARITY = "Focus on generating {rarity} samples."
# --- Default Datasets for Landing Page ---
landing_page_datasets_generated_text = """
1. NewsEventsPredict (classification, media, trend)
2. FinancialForecast (economy, stocks, regression)
3. HealthMonitor (science, real-time, anomaly detection)
4. SportsAnalysis (classification, performance, player tracking)
5. SciLiteracyTools (language modeling, science literacy, text classification)
6. RetailSalesAnalyzer (consumer behavior, sales trend, segmentation)
7. SocialSentimentEcho (social media, emotion analysis, clustering)
8. NewsEventTracker (classification, public awareness, topical clustering)
9. HealthVitalSigns (anomaly detection, biometrics, prediction)
10. GameStockPredict (classification, finance, sports contingency)
"""
default_output = landing_page_datasets_generated_text.strip().split("\n")
assert len(default_output) == MAX_NB_ITEMS_PER_GENERATION_CALL
# --- Dataset Card Template ---
DATASET_CARD_CONTENT = """
---
license: mit
tags:
- infinite-dataset-hub
- synthetic
---
{title}
_Note: This is an AI-generated dataset so its content may be inaccurate or false_
{content}
**Source of the data:**
The dataset was generated using the [Infinite Dataset Hub]({url}) and {model_id} using the query '{search_query}':
- **Dataset Generation Page**: {dataset_url}
- **Model**: https://huggingface.co/{model_id}
- **More Datasets**: https://huggingface.co/datasets?other=infinite-dataset-hub
"""
# --- Gradio HTML ---
html = """
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Infinite Dataset Hub</title>
<script src="https://cdn.tailwindcss.com"></script>
<script src="https://cdn.jsdelivr.net/npm/marked/marked.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/papaparse@5.3.0/papaparse.min.js"></script>
<script>
tailwind.config = {
darkMode: 'class',
theme: {
extend: {
colors: {
primary: '#5D5CDE',
},
}
}
}
</script>
<style>
.shimmer {
background: linear-gradient(90deg, #f0f0f0 25%, #e0e0e0 50%, #f0f0f0 75%);
background-size: 200% 100%;
animation: shimmer 1.5s infinite;
border-radius: 4px;
}
@keyframes shimmer {
0% {
background-position: -200% 0;
}
100% {
background-position: 200% 0;
}
}
/* Dark mode overrides */
.dark .shimmer {
background: linear-gradient(90deg, #2a2a2a 25%, #3a3a3a 50%, #2a2a2a 75%);
background-size: 200% 100%;
}
.dataset-card {
transition: transform 0.2s, box-shadow 0.2s;
}
.dataset-card:hover {
transform: translateY(-2px);
box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.1), 0 4px 6px -2px rgba(0, 0, 0, 0.05);
}
.dark .dataset-card:hover {
box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.3), 0 4px 6px -2px rgba(0, 0, 0, 0.2);
}
/* Table styling */
table {
width: 100%;
border-collapse: collapse;
margin: 1rem 0;
}
table thead th {
background-color: #f3f4f6;
padding: 0.75rem;
text-align: left;
font-weight: 600;
}
.dark table thead th {
background-color: #374151;
}
table tbody td {
padding: 0.75rem;
border-top: 1px solid #e5e7eb;
}
.dark table tbody td {
border-top: 1px solid #4b5563;
}
table tbody tr:nth-child(even) {
background-color: #f9fafb;
}
.dark table tbody tr:nth-child(even) {
background-color: #1f2937;
}
/* Search engine badge */
.engine-badge {
position: absolute;
top: -8px;
right: -8px;
font-size: 0.7rem;
padding: 2px 6px;
border-radius: 9999px;
background-color: #5D5CDE;
color: white;
box-shadow: 0 2px 4px rgba(0,0,0,0.1);
}
.dark .engine-badge {
box-shadow: 0 2px 4px rgba(0,0,0,0.3);
}
/* Toggle switch */
.toggle-switch {
position: relative;
display: inline-block;
width: 50px;
height: 24px;
}
.toggle-switch input {
opacity: 0;
width: 0;
height: 0;
}
.toggle-slider {
position: absolute;
cursor: pointer;
top: 0;
left: 0;
right: 0;
bottom: 0;
background-color: #ccc;
transition: .4s;
border-radius: 24px;
}
.toggle-slider:before {
position: absolute;
content: "";
height: 16px;
width: 16px;
left: 4px;
bottom: 4px;
background-color: white;
transition: .4s;
border-radius: 50%;
}
input:checked + .toggle-slider {
background-color: #5D5CDE;
}
input:checked + .toggle-slider:before {
transform: translateX(26px);
}
</style>
</head>
<body class="bg-white dark:bg-gray-900 text-gray-800 dark:text-gray-200 min-h-screen">
<!-- Dark mode detection -->
<script>
if (window.matchMedia && window.matchMedia('(prefers-color-scheme: dark)').matches) {
document.documentElement.classList.add('dark');
}
window.matchMedia('(prefers-color-scheme: dark)').addEventListener('change', event => {
if (event.matches) {
document.documentElement.classList.add('dark');
} else {
document.documentElement.classList.remove('dark');
}
});
</script>
<div class="container mx-auto px-4 py-8">
<!-- Header -->
<header class="text-center mb-8">
<h1 class="text-3xl font-bold mb-2">🤗 Infinite Dataset Hub ♾️</h1>
<p class="text-lg text-gray-600 dark:text-gray-400">Generate datasets from AI and real-world data sources</p>
</header>
<!-- Main Content -->
<main>
<!-- Search Section -->
<div id="search-page" class="mb-8">
<div class="max-w-3xl mx-auto">
<div class="mb-4">
<div class="flex mb-2">
<input id="search-input" type="text" placeholder="Search datasets, get infinite results"
class="flex-grow px-4 py-3 text-base rounded-l-lg border border-gray-300 dark:border-gray-700 focus:outline-none focus:ring-2 focus:ring-primary dark:bg-gray-800">
<button id="search-button" class="bg-primary text-white px-6 py-3 rounded-r-lg hover:bg-opacity-90 transition">
🔍
</button>
</div>
<div class="flex items-center justify-between p-3 bg-gray-100 dark:bg-gray-800 rounded-lg">
<div class="flex items-center">
<label class="toggle-switch mr-3">
<input type="checkbox" id="data-source-toggle" checked>
<span class="toggle-slider"></span>
</label>
<div>
<span id="data-source-text" class="font-medium">Using: Real + AI Data</span>
<p class="text-xs text-gray-500 dark:text-gray-400">Toggle to switch between data sources</p>
</div>
</div>
<button id="engine-settings-button" class="text-primary hover:underline flex items-center">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-1" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M11.49 3.17c-.38-1.56-2.6-1.56-2.98 0a1.532 1.532 0 01-2.286.948c-1.372-.836-2.942.734-2.106 2.106.54.886.061 2.042-.947 2.287-1.561.379-1.561 2.6 0 2.978a1.532 1.532 0 01.947 2.287c-.836 1.372.734 2.942 2.106 2.106a1.532 1.532 0 012.287.947c.379 1.561 2.6 1.561 2.978 0a1.533 1.533 0 012.287-.947c1.372.836 2.942-.734 2.106-2.106a1.533 1.533 0 01.947-2.287c1.561-.379 1.561-2.6 0-2.978a1.532 1.532 0 01-.947-2.287c.836-1.372-.734-2.942-2.106-2.106a1.532 1.532 0 01-2.287-.947zM10 13a3 3 0 100-6 3 3 0 000 6z" clip-rule="evenodd" />
</svg>
Search Engines
</button>
</div>
</div>
<!-- Search Engine Selection Modal -->
<div id="engine-modal" class="fixed inset-0 bg-black bg-opacity-50 flex items-center justify-center z-50 hidden">
<div class="bg-white dark:bg-gray-800 rounded-lg p-6 max-w-lg w-full max-h-[80vh] overflow-y-auto">
<div class="flex justify-between items-center mb-4">
<h3 class="text-xl font-bold">Search Engine Settings</h3>
<button id="close-modal-button" class="text-gray-500 hover:text-gray-700 dark:text-gray-400 dark:hover:text-gray-200">
<svg xmlns="http://www.w3.org/2000/svg" class="h-6 w-6" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M6 18L18 6M6 6l12 12" />
</svg>
</button>
</div>
<p class="mb-4 text-sm text-gray-600 dark:text-gray-400">
Select which search engines to use for real data retrieval. A diverse selection improves results.
</p>
<div id="engine-options" class="space-y-2 mb-6">
<!-- Engine options will be dynamically inserted here -->
</div>
<div class="flex justify-between">
<button id="select-all-engines" class="text-primary hover:underline">Select All</button>
<button id="deselect-all-engines" class="text-primary hover:underline">Deselect All</button>
</div>
<div class="mt-6 flex justify-end">
<button id="save-engines-button" class="bg-primary text-white px-4 py-2 rounded hover:bg-opacity-90 transition">
Save Settings
</button>
</div>
</div>
</div>
<div id="dataset-results" class="grid grid-cols-1 md:grid-cols-2 gap-4 mt-6">
<!-- Dataset cards will be dynamically inserted here -->
</div>
<div id="load-more-container" class="text-center mt-6 hidden">
<button id="load-more-button" class="bg-gray-200 dark:bg-gray-700 px-6 py-3 rounded-lg hover:bg-gray-300 dark:hover:bg-gray-600 transition">
Load more datasets
</button>
</div>
</div>
</div>
<!-- Dataset Detail Page -->
<div id="dataset-page" class="hidden max-w-4xl mx-auto">
<button id="back-button" class="flex items-center text-primary mb-4 hover:underline">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-1" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M9.707 14.707a1 1 0 01-1.414 0l-4-4a1 1 0 010-1.414l4-4a1 1 0 011.414 1.414L7.414 9H15a1 1 0 110 2H7.414l2.293 2.293a1 1 0 010 1.414z" clip-rule="evenodd" />
</svg>
Back to Search
</button>
<div id="dataset-header" class="mb-4">
<div class="flex items-center justify-between">
<h2 id="dataset-title" class="text-2xl font-bold"></h2>
<span id="data-source-badge" class="px-3 py-1 rounded-full text-xs font-medium bg-green-100 text-green-800 dark:bg-green-900 dark:text-green-200">
Real Data
</span>
</div>
<div id="dataset-tags" class="text-sm text-gray-600 dark:text-gray-400 mt-1"></div>
</div>
<div id="data-source-info" class="bg-blue-50 dark:bg-blue-900 p-4 rounded-lg mb-6 text-blue-800 dark:text-blue-200">
<h3 class="font-semibold mb-1 flex items-center">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-1" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M18 10a8 8 0 11-16 0 8 8 0 0116 0zm-7-4a1 1 0 11-2 0 1 1 0 012 0zM9 9a1 1 0 000 2v3a1 1 0 001 1h1a1 1 0 100-2v-3a1 1 0 00-1-1H9z" clip-rule="evenodd" />
</svg>
Data Source Information
</h3>
<p id="source-details" class="text-sm"></p>
</div>
<div id="dataset-description" class="prose dark:prose-invert prose-sm sm:prose max-w-none mb-6"></div>
<div id="dataset-preview" class="mb-6 overflow-x-auto">
<h3 class="text-xl font-semibold mb-3">Dataset Preview</h3>
<div id="preview-table" class="border dark:border-gray-700 rounded-lg overflow-hidden"></div>
</div>
<div id="generate-actions" class="mb-8">
<button id="generate-full-button" class="bg-primary text-white px-6 py-3 rounded-lg hover:bg-opacity-90 transition mr-3">
Generate Full Dataset
</button>
<div id="generate-status" class="hidden mt-4">
<div class="flex items-center">
<div class="animate-spin rounded-full h-5 w-5 border-b-2 border-primary mr-3"></div>
<span>Generating dataset... <span id="rows-count">0</span> rows created</span>
</div>
<div class="w-full bg-gray-200 dark:bg-gray-700 rounded-full h-2.5 mt-2">
<div id="progress-bar" class="bg-primary h-2.5 rounded-full" style="width: 0%"></div>
</div>
</div>
</div>
<div id="full-dataset" class="hidden mb-6">
<h3 class="text-xl font-semibold mb-3">Full Dataset</h3>
<div id="full-table" class="border dark:border-gray-700 rounded-lg overflow-hidden"></div>
<div class="mt-4 flex flex-wrap gap-3">
<button id="download-csv-button" class="bg-green-600 hover:bg-green-700 text-white px-4 py-2 rounded-lg transition flex items-center">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-2" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M3 17a1 1 0 011-1h12a1 1 0 110 2H4a1 1 0 01-1-1zm3.293-7.707a1 1 0 011.414 0L9 10.586V3a1 1 0 112 0v7.586l1.293-1.293a1 1 0 111.414 1.414l-3 3a1 1 0 01-1.414 0l-3-3a1 1 0 010-1.414z" clip-rule="evenodd" />
</svg>
Download CSV
</button>
<button id="download-json-button" class="bg-yellow-600 hover:bg-yellow-700 text-white px-4 py-2 rounded-lg transition flex items-center">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-2" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M3 17a1 1 0 011-1h12a1 1 0 110 2H4a1 1 0 01-1-1zm3.293-7.707a1 1 0 011.414 0L9 10.586V3a1 1 0 112 0v7.586l1.293-1.293a1 1 0 111.414 1.414l-3 3a1 1 0 01-1.414 0l-3-3a1 1 0 010-1.414z" clip-rule="evenodd" />
</svg>
Download JSON
</button>
<button id="download-parquet-button" class="bg-blue-600 hover:bg-blue-700 text-white px-4 py-2 rounded-lg transition flex items-center">
<svg xmlns="http://www.w3.org/2000/svg" class="h-5 w-5 mr-2" viewBox="0 0 20 20" fill="currentColor">
<path fill-rule="evenodd" d="M3 17a1 1 0 011-1h12a1 1 0 110 2H4a1 1 0 01-1-1zm3.293-7.707a1 1 0 011.414 0L9 10.586V3a1 1 0 112 0v7.586l1.293-1.293a1 1 0 111.414 1.414l-3 3a1 1 0 01-1.414 0l-3-3a1 1 0 010-1.414z" clip-rule="evenodd" />
</svg>
Download Parquet
</button>
</div>
</div>
</div>
</main>
<!-- Footer -->
<footer class="mt-12 text-center text-sm text-gray-600 dark:text-gray-400">
<p>Powered by Claude-3.7-Sonnet • Datasets generated from real sources and AI</p>
</footer>
</div>
<script>
// Constants and global state
const MAX_DATASETS_PER_PAGE = 10;
const MAX_FULL_DATASET_ROWS = 100;
// List of search engines
const searchEngines = [
"AlltheInternet.com", "DuckDuckGo.com", "Google.com", "Bing.com", "Search.Yahoo.com",
"Startpage.com", "Qwant.com", "Ecosia.org", "WolframAlpha.com", "Mojeek.co.uk",
"Search.Brave.com", "Yandex.com", "Baidu.com", "Gibiru.com", "MetaGer.org",
"Swisscows.com", "Presearch.com", "Ekoru.org", "Search.Lilo.org"
];
let currentDatasets = [];
let currentPage = 1;
let currentSearchQuery = '';
let currentDataset = null;
let fullDatasetRows = [];
let useRealData = true;
let selectedEngines = ["DuckDuckGo.com", "Bing.com", "Search.Yahoo.com", "Search.Brave.com", "Ecosia.org"];
let currentEngine = ""; // Store the engine currently being used
// DOM Elements
const searchInput = document.getElementById('search-input');
const searchButton = document.getElementById('search-button');
const resultsContainer = document.getElementById('dataset-results');
const loadMoreContainer = document.getElementById('load-more-container');
const loadMoreButton = document.getElementById('load-more-button');
const searchPage = document.getElementById('search-page');
const datasetPage = document.getElementById('dataset-page');
const backButton = document.getElementById('back-button');
const datasetTitle = document.getElementById('dataset-title');
const datasetTags = document.getElementById('dataset-tags');
const datasetDescription = document.getElementById('dataset-description');
const previewTable = document.getElementById('preview-table');
const generateFullButton = document.getElementById('generate-full-button');
const generateStatus = document.getElementById('generate-status');
const rowsCount = document.getElementById('rows-count');
const progressBar = document.getElementById('progress-bar');
const fullDatasetSection = document.getElementById('full-dataset');
const fullTable = document.getElementById('full-table');
const downloadCsvButton = document.getElementById('download-csv-button');
const downloadJsonButton = document.getElementById('download-json-button');
const downloadParquetButton = document.getElementById('download-parquet-button');
const dataSourceToggle = document.getElementById('data-source-toggle');
const dataSourceText = document.getElementById('data-source-text');
const dataSourceBadge = document.getElementById('data-source-badge');
const sourceDetails = document.getElementById('source-details');
const engineSettingsButton = document.getElementById('engine-settings-button');
const engineModal = document.getElementById('engine-modal');
const engineOptions = document.getElementById('engine-options');
const closeModalButton = document.getElementById('close-modal-button');
const saveEnginesButton = document.getElementById('save-engines-button');
const selectAllEngines = document.getElementById('select-all-engines');
const deselectAllEngines = document.getElementById('deselect-all-engines');
// Event Listeners
document.addEventListener('DOMContentLoaded', () => {
searchButton.addEventListener('click', performSearch);
searchInput.addEventListener('keypress', (e) => {
if (e.key === 'Enter') performSearch();
});
loadMoreButton.addEventListener('click', loadMoreDatasets);
backButton.addEventListener('click', showSearchPage);
generateFullButton.addEventListener('click', generateFullDataset);
downloadCsvButton.addEventListener('click', () => downloadData('csv'));
downloadJsonButton.addEventListener('click', () => downloadData('json'));
downloadParquetButton.addEventListener('click', () => downloadData('parquet'));
dataSourceToggle.addEventListener('change', toggleDataSource);
engineSettingsButton.addEventListener('click', showEngineModal);
closeModalButton.addEventListener('click', hideEngineModal);
saveEnginesButton.addEventListener('click', saveEngineSettings);
selectAllEngines.addEventListener('click', () => toggleAllEngines(true));
deselectAllEngines.addEventListener('click', () => toggleAllEngines(false));
// Initialize engine options
populateEngineOptions();
// Show initial placeholder datasets
showPlaceholderDatasets();
});
// Search Engine Settings
function populateEngineOptions() {
engineOptions.innerHTML = '';
searchEngines.forEach(engine => {
const isChecked = selectedEngines.includes(engine);
const optionDiv = document.createElement('div');
optionDiv.className = 'flex items-center';
optionDiv.innerHTML = `
<input type="checkbox" id="engine-${engine}" class="engine-checkbox mr-2 h-4 w-4"
value="${engine}" ${isChecked ? 'checked' : ''}>
<label for="engine-${engine}" class="cursor-pointer">${engine}</label>
`;
engineOptions.appendChild(optionDiv);
});
}
function showEngineModal() {
engineModal.classList.remove('hidden');
}
function hideEngineModal() {
engineModal.classList.add('hidden');
}
function saveEngineSettings() {
const checkboxes = document.querySelectorAll('.engine-checkbox:checked');
selectedEngines = Array.from(checkboxes).map(cb => cb.value);
if (selectedEngines.length === 0) {
// Ensure at least one engine is selected
selectedEngines = ["DuckDuckGo.com"];
document.getElementById(`engine-DuckDuckGo.com`).checked = true;
showNotification("At least one search engine must be selected. Using DuckDuckGo as default.");
}
hideEngineModal();
showNotification(`Updated search engine settings. Using ${selectedEngines.length} engines.`);
}
function toggleAllEngines(select) {
const checkboxes = document.querySelectorAll('.engine-checkbox');
checkboxes.forEach(cb => {
cb.checked = select;
});
}
// Toggle data source between real and AI
function toggleDataSource() {
useRealData = dataSourceToggle.checked;
dataSourceText.textContent = useRealData ? "Using: Real + AI Data" : "Using: AI Data Only";
// Show or hide engine settings button
engineSettingsButton.style.display = useRealData ? "flex" : "none";
showNotification(`Switched to ${useRealData ? "combined real and synthetic" : "synthetic-only"} data mode`);
}
// Search functionality
function performSearch() {
const query = searchInput.value.trim();
if (!query) return;
currentSearchQuery = query;
currentPage = 1;
currentDatasets = [];
resultsContainer.innerHTML = '';
showLoadingSkeletons();
if (useRealData) {
// Use real data from search engines + AI
searchWithRealData(query);
} else {
// Use only AI-generated data
searchWithAIData(query);
}
}
function searchWithRealData(query) {
// Randomly select a search engine from the user's selected engines
currentEngine = selectedEngines[Math.floor(Math.random() * selectedEngines.length)];
// Register handler for dataset names based on real search results
window.Poe.registerHandler("real-search-handler", (result) => {
if (result.status === "error") {
showError("Error querying search engines");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
// Parse the dataset names and tags from the response
const datasets = parseDatasetResults(message.content);
datasets.forEach(dataset => {
dataset.isReal = true;
dataset.engine = currentEngine;
});
currentDatasets = datasets;
// Display the datasets
resultsContainer.innerHTML = '';
displayDatasets(datasets);
// Show load more button if we have results
if (datasets.length > 0) {
loadMoreContainer.classList.remove('hidden');
}
}
});
try {
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet You are a data specialist who can transform real search results into structured datasets.
A user is searching for data about: "${query}"
Imagine you've queried ${currentEngine} and received real search results. Create a list of 10 specific datasets that could be created from these search results.
For each dataset:
1. Give it a clear, specific name related to the search topic
2. Include 3-5 relevant tags in parentheses, with one tag specifying the ML task type (classification, regression, clustering, etc.)
Format each dataset as:
1. DatasetName (tag1, tag2, ml_task_tag)
Make these datasets sound like real collections that could be created from ${currentEngine} search results on "${query}".`,
{
handler: "real-search-handler",
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
// Fall back to AI data
searchWithAIData(query);
}
}
function searchWithAIData(query) {
// Register handler for AI-generated dataset names
window.Poe.registerHandler("dataset-search-handler", (result) => {
if (result.status === "error") {
showError("Error generating datasets");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
// Parse the dataset names and tags from the response
const datasets = parseDatasetResults(message.content);
datasets.forEach(dataset => {
dataset.isReal = false;
});
currentDatasets = datasets;
// Display the datasets
resultsContainer.innerHTML = '';
displayDatasets(datasets);
// Show load more button if we have results
if (datasets.length > 0) {
loadMoreContainer.classList.remove('hidden');
}
}
});
try {
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet A Machine Learning Practioner is looking for a dataset that matches '${query}'.
Generate a list of ${MAX_DATASETS_PER_PAGE} names of quality datasets that don't exist but sound plausible and would
be helpful. Feel free to reuse words from the query '${query}' to name the datasets.
Every dataset should be about '${query}' and have descriptive tags/keywords including the ML task name associated with the dataset (classification, regression, anomaly detection, etc.). Use the following format:
1. DatasetName1 (tag1, tag2, tag3)
2. DatasetName2 (tag1, tag2, tag3)`,
{
handler: "dataset-search-handler",
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
}
}
function parseDatasetResults(content) {
const lines = content.split('\n');
const datasets = [];
lines.forEach(line => {
// Match lines that start with a number followed by a period
const match = line.match(/^\s*\d+\.\s+(.+?)\s+\((.+?)\)/);
if (match) {
const name = match[1].trim();
const tags = match[2].split(',').map(tag => tag.trim());
datasets.push({ name, tags });
}
});
return datasets;
}
function displayDatasets(datasets) {
datasets.forEach(dataset => {
const card = document.createElement('div');
card.className = 'dataset-card bg-white dark:bg-gray-800 rounded-lg p-4 border border-gray-200 dark:border-gray-700 cursor-pointer relative';
const tagsHtml = dataset.tags.map(tag =>
`<span class="inline-block bg-gray-100 dark:bg-gray-700 text-gray-800 dark:text-gray-300 text-xs px-2 py-1 rounded mr-1 mb-1">${tag}</span>`
).join('');
// Add a badge for real data
let badgeHtml = '';
if (dataset.isReal) {
badgeHtml = `<span class="engine-badge" title="Data from ${dataset.engine}">${dataset.engine.split('.')[0]}</span>`;
}
card.innerHTML = `
${badgeHtml}
<h3 class="text-lg font-semibold mb-2">${dataset.name}</h3>
<div class="flex flex-wrap mt-2">${tagsHtml}</div>
`;
card.addEventListener('click', () => showDatasetDetails(dataset));
resultsContainer.appendChild(card);
});
}
function showLoadingSkeletons() {
for (let i = 0; i < 4; i++) {
const skeleton = document.createElement('div');
skeleton.className = 'bg-white dark:bg-gray-800 rounded-lg p-4 border border-gray-200 dark:border-gray-700';
skeleton.innerHTML = `
<div class="shimmer h-6 w-3/4 mb-2"></div>
<div class="flex flex-wrap mt-2">
<div class="shimmer h-6 w-16 rounded mr-1 mb-1"></div>
<div class="shimmer h-6 w-20 rounded mr-1 mb-1"></div>
<div class="shimmer h-6 w-24 rounded mr-1 mb-1"></div>
</div>
`;
resultsContainer.appendChild(skeleton);
}
}
function loadMoreDatasets() {
currentPage++;
// Use the same data source (real or AI) as the initial search
if (useRealData) {
loadMoreRealDatasets();
} else {
loadMoreAIDatasets();
}
}
function loadMoreRealDatasets() {
// Rotate to a different search engine for variety
const previousEngine = currentEngine;
while (currentEngine === previousEngine && selectedEngines.length > 1) {
currentEngine = selectedEngines[Math.floor(Math.random() * selectedEngines.length)];
}
// Register handler for more datasets
window.Poe.registerHandler("more-real-datasets-handler", (result) => {
if (result.status === "error") {
showError("Error generating more datasets");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
// Parse the dataset names and tags from the response
const datasets = parseDatasetResults(message.content);
datasets.forEach(dataset => {
dataset.isReal = true;
dataset.engine = currentEngine;
});
currentDatasets = [...currentDatasets, ...datasets];
// Display the datasets
displayDatasets(datasets);
}
});
try {
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet You're a data specialist who can transform real search results into structured datasets.
Continue our previous search for data about: "${currentSearchQuery}"
Now let's use a different search engine: ${currentEngine}
Create 10 more specific datasets that could be created from these search results. Make sure these are different from the previous datasets.
Use the same format:
1. DatasetName (tag1, tag2, ml_task_tag)
Make these datasets sound like real collections that could be created from ${currentEngine} search results on "${currentSearchQuery}".`,
{
handler: "more-real-datasets-handler",
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
// Fall back to AI data
loadMoreAIDatasets();
}
}
function loadMoreAIDatasets() {
// Register handler for more AI datasets
window.Poe.registerHandler("more-datasets-handler", (result) => {
if (result.status === "error") {
showError("Error generating more datasets");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
// Parse the dataset names and tags from the response
const datasets = parseDatasetResults(message.content);
datasets.forEach(dataset => {
dataset.isReal = false;
});
currentDatasets = [...currentDatasets, ...datasets];
// Display the datasets
displayDatasets(datasets);
}
});
try {
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet Please generate ${MAX_DATASETS_PER_PAGE} more dataset names about '${currentSearchQuery}'. Use the same format as before:
1. DatasetName1 (tag1, tag2, tag3)
Make sure these are completely different from previous suggestions.`,
{
handler: "more-datasets-handler",
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
}
}
function showDatasetDetails(dataset) {
currentDataset = dataset;
searchPage.classList.add('hidden');
datasetPage.classList.remove('hidden');
// Update UI with dataset info
datasetTitle.textContent = dataset.name;
datasetTags.innerHTML = dataset.tags.map(tag =>
`<span class="inline-block bg-gray-100 dark:bg-gray-700 text-gray-800 dark:text-gray-300 text-xs px-2 py-1 rounded mr-1 mb-1">${tag}</span>`
).join('');
// Update source badge
if (dataset.isReal) {
dataSourceBadge.textContent = "Real Data";
dataSourceBadge.className = "px-3 py-1 rounded-full text-xs font-medium bg-green-100 text-green-800 dark:bg-green-900 dark:text-green-200";
sourceDetails.innerHTML = `This dataset is based on real information queried from <strong>${dataset.engine}</strong> for the search term "<strong>${currentSearchQuery}</strong>". The data has been structured for machine learning use.`;
} else {
dataSourceBadge.textContent = "AI-Generated";
dataSourceBadge.className = "px-3 py-1 rounded-full text-xs font-medium bg-purple-100 text-purple-800 dark:bg-purple-900 dark:text-purple-200";
sourceDetails.innerHTML = `This is an AI-generated dataset created using Claude-3.7-Sonnet. The content is synthetic and designed to represent plausible data related to "${currentSearchQuery}".`;
}
// Clear previous content
datasetDescription.innerHTML = '<div class="shimmer h-4 w-full mb-2"></div>'.repeat(3);
previewTable.innerHTML = '';
fullDatasetSection.classList.add('hidden');
generateStatus.classList.add('hidden');
generateFullButton.disabled = false;
// Reset full dataset
fullDatasetRows = [];
// Generate dataset preview - different approach for real vs AI data
if (dataset.isReal) {
generateRealDatasetPreview(dataset);
} else {
generateAIDatasetPreview(dataset);
}
// Scroll to top
window.scrollTo(0, 0);
}
function generateRealDatasetPreview(dataset) {
window.Poe.registerHandler("real-preview-handler", (result) => {
if (result.status === "error") {
datasetDescription.innerHTML = '<p class="text-red-500">Error generating dataset preview</p>';
return;
}
const message = result.responses[0];
if (message.status === "complete") {
const content = message.content;
// Extract description and CSV
const parts = content.split('**CSV Content Preview:**');
let description = "";
let csvContent = "";
if (parts.length > 1) {
description = parts[0].replace('**Dataset Description:**', '').trim();
csvContent = parts[1].trim();
// Clean up CSV content (remove markdown code block markers)
csvContent = csvContent.replace(/```csv\n|```\n|```/g, '').trim();
} else {
description = "No description available";
csvContent = content;
}
// Display description
datasetDescription.innerHTML = marked.parse(description);
// Parse and display CSV preview
try {
const results = Papa.parse(csvContent, {
header: true,
skipEmptyLines: true
});
if (results.data && results.data.length > 0) {
// Create table from CSV data
createTable(previewTable, results.data, results.meta.fields);
} else {
previewTable.innerHTML = '<p class="p-4 text-red-500">No preview data available</p>';
}
} catch (err) {
previewTable.innerHTML = `<p class="p-4 text-red-500">Error parsing CSV: ${err.message}</p>`;
}
}
});
try {
const tagsStr = dataset.tags.join(', ');
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet You're a specialist in converting web search results into structured data.
Based on search results from ${dataset.engine} about "${currentSearchQuery}",
create a preview of the dataset "${dataset.name}" with tags "${tagsStr}".
First, write a detailed description of what this dataset contains, its structure, and how it was constructed from web search results.
Then, generate a realistic 5-row CSV preview that resembles data you might get if you scraped and structured real results from ${dataset.engine}.
Format your response with:
**Dataset Description:** [detailed description]
**CSV Content Preview:**
\`\`\`csv
[CSV header and 5 rows of realistic data]
\`\`\`
Include relevant columns for the dataset type, with proper labels/categories where appropriate. The data should look like it came from real sources.`,
{
handler: "real-preview-handler",
stream: false,
openChat: false
}
);
} catch (err) {
datasetDescription.innerHTML = `<p class="text-red-500">Error: ${err.message}</p>`;
}
}
function generateAIDatasetPreview(dataset) {
window.Poe.registerHandler("dataset-preview-handler", (result) => {
if (result.status === "error") {
datasetDescription.innerHTML = '<p class="text-red-500">Error generating dataset preview</p>';
return;
}
const message = result.responses[0];
if (message.status === "complete") {
const content = message.content;
// Extract description and CSV
const parts = content.split('**CSV Content Preview:**');
let description = "";
let csvContent = "";
if (parts.length > 1) {
description = parts[0].replace('**Dataset Description:**', '').trim();
csvContent = parts[1].trim();
// Clean up CSV content (remove markdown code block markers)
csvContent = csvContent.replace(/```csv\n|```\n|```/g, '').trim();
} else {
description = "No description available";
csvContent = content;
}
// Display description
datasetDescription.innerHTML = marked.parse(description);
// Parse and display CSV preview
try {
const results = Papa.parse(csvContent, {
header: true,
skipEmptyLines: true
});
if (results.data && results.data.length > 0) {
// Create table from CSV data
createTable(previewTable, results.data, results.meta.fields);
} else {
previewTable.innerHTML = '<p class="p-4 text-red-500">No preview data available</p>';
}
} catch (err) {
previewTable.innerHTML = `<p class="p-4 text-red-500">Error parsing CSV: ${err.message}</p>`;
}
}
});
try {
const tagsStr = dataset.tags.join(', ');
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet An ML practitioner is looking for a dataset CSV after the query '${currentSearchQuery}'.
Generate the first 5 rows of a plausible and quality CSV for the dataset '${dataset.name}'.
You can get inspiration from related keywords '${tagsStr}' but most importantly the dataset should correspond to the query '${currentSearchQuery}'.
Focus on quality text content and use a 'label' or 'labels' column if it makes sense (invent labels, avoid reusing the keywords, be accurate while labelling texts).
Reply using a short description of the dataset with title **Dataset Description:** followed by the CSV content in a code block and with title **CSV Content Preview:**`,
{
handler: "dataset-preview-handler",
stream: false,
openChat: false
}
);
} catch (err) {
datasetDescription.innerHTML = `<p class="text-red-500">Error: ${err.message}</p>`;
}
}
function createTable(container, data, headers) {
container.innerHTML = '';
const table = document.createElement('table');
table.className = 'w-full';
// Create header
const thead = document.createElement('thead');
const headerRow = document.createElement('tr');
headers.forEach(header => {
const th = document.createElement('th');
th.textContent = header;
headerRow.appendChild(th);
});
thead.appendChild(headerRow);
table.appendChild(thead);
// Create body
const tbody = document.createElement('tbody');
data.forEach(row => {
const tr = document.createElement('tr');
headers.forEach(header => {
const td = document.createElement('td');
td.textContent = row[header] || '';
tr.appendChild(td);
});
tbody.appendChild(tr);
});
table.appendChild(tbody);
container.appendChild(table);
}
function generateFullDataset() {
// Disable button and show status
generateFullButton.disabled = true;
generateStatus.classList.remove('hidden');
rowsCount.textContent = '0';
progressBar.style.width = '0%';
// Set up variables for tracking generation
let csvHeader = '';
const targetRows = MAX_FULL_DATASET_ROWS;
let currentRows = 0;
fullDatasetRows = [];
// Get the CSV header from the preview table
const previewHeaders = Array.from(previewTable.querySelectorAll('thead th')).map(th => th.textContent);
csvHeader = previewHeaders.join(',');
// Add initial rows from preview
const previewRows = Array.from(previewTable.querySelectorAll('tbody tr')).map(tr => {
const row = {};
Array.from(tr.querySelectorAll('td')).forEach((td, index) => {
row[previewHeaders[index]] = td.textContent;
});
return row;
});
fullDatasetRows = [...previewRows];
currentRows = previewRows.length;
updateGenerationProgress(currentRows, targetRows);
// Choose generation method based on dataset type
if (currentDataset.isReal) {
generateFullRealDataset(previewHeaders, csvHeader, currentRows, targetRows);
} else {
generateFullAIDataset(previewHeaders, csvHeader, currentRows, targetRows);
}
}
function generateFullRealDataset(previewHeaders, csvHeader, currentRows, targetRows) {
// Function to generate more rows in batches from "real" search results
const generateBatch = (batchIndex) => {
const batchSize = 15; // Larger batches for efficiency
const startRow = currentRows + batchIndex * batchSize;
if (startRow >= targetRows) {
// We've reached the target, show the full dataset
showFullDataset();
return;
}
window.Poe.registerHandler(`real-batch-${batchIndex}-handler`, (result) => {
if (result.status === "error") {
showError("Error generating dataset rows");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
const content = message.content;
// Extract CSV content (remove markdown code block markers)
let csvContent = content.replace(/```csv\n|```\n|```/g, '').trim();
// If there are multiple code blocks, try to find one with CSV data
if (csvContent.includes('```')) {
const codeBlocks = content.match(/```(?:csv)?\n([\s\S]*?)```/g) || [];
if (codeBlocks.length > 0) {
csvContent = codeBlocks[0].replace(/```(?:csv)?\n|```/g, '').trim();
}
}
try {
// Parse the CSV
const results = Papa.parse(csvContent, {
header: true,
skipEmptyLines: true
});
if (results.data && results.data.length > 0) {
// Add the new rows
fullDatasetRows = [...fullDatasetRows, ...results.data];
currentRows += results.data.length;
// Update progress
updateGenerationProgress(currentRows, targetRows);
// Generate next batch
generateBatch(batchIndex + 1);
} else {
// Try again with a different prompt
generateBatch(batchIndex);
}
} catch (err) {
console.error("Error parsing CSV:", err);
// Try again
generateBatch(batchIndex);
}
}
});
try {
// For variation, rotate through engines for each batch
const engineForBatch = selectedEngines[batchIndex % selectedEngines.length] || currentDataset.engine;
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet You're expanding a dataset based on search results from ${engineForBatch}.
For the dataset "${currentDataset.name}" about "${currentSearchQuery}", please generate ${batchSize} more rows of data.
Use this exact CSV header: ${csvHeader}
The data should look realistic, as if it came from actual ${engineForBatch} search results for "${currentSearchQuery}".
Include appropriate values for each field, maintaining the same patterns and types as seen in the existing data.
Only include the CSV data in your response (header + ${batchSize} rows), no explanations or additional text.`,
{
handler: `real-batch-${batchIndex}-handler`,
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
}
};
// Start generating batches
generateBatch(0);
}
function generateFullAIDataset(previewHeaders, csvHeader, currentRows, targetRows) {
// Function to generate more rows in batches from AI
const generateBatch = (batchIndex) => {
const batchSize = 10;
const startRow = currentRows + batchIndex * batchSize;
if (startRow >= targetRows) {
// We've reached the target, show the full dataset
showFullDataset();
return;
}
window.Poe.registerHandler(`batch-${batchIndex}-handler`, (result) => {
if (result.status === "error") {
showError("Error generating dataset rows");
return;
}
const message = result.responses[0];
if (message.status === "complete") {
const content = message.content;
// Extract CSV content (remove markdown code block markers)
let csvContent = content.replace(/```csv\n|```\n|```/g, '').trim();
// If there are multiple code blocks, try to find one with CSV data
if (csvContent.includes('```')) {
const codeBlocks = content.match(/```(?:csv)?\n([\s\S]*?)```/g) || [];
if (codeBlocks.length > 0) {
csvContent = codeBlocks[0].replace(/```(?:csv)?\n|```/g, '').trim();
}
}
try {
// Parse the CSV
const results = Papa.parse(csvContent, {
header: true,
skipEmptyLines: true
});
if (results.data && results.data.length > 0) {
// Add the new rows
fullDatasetRows = [...fullDatasetRows, ...results.data];
currentRows += results.data.length;
// Update progress
updateGenerationProgress(currentRows, targetRows);
// Generate next batch
generateBatch(batchIndex + 1);
} else {
// Try again with a different prompt
generateBatch(batchIndex);
}
} catch (err) {
console.error("Error parsing CSV:", err);
// Try again
generateBatch(batchIndex);
}
}
});
try {
const tagsStr = currentDataset.tags.join(', ');
window.Poe.sendUserMessage(
`@Claude-3.7-Sonnet For the dataset '${currentDataset.name}' about '${currentSearchQuery}' with tags '${tagsStr}',
please generate ${batchSize} more sample rows in CSV format. Use the same CSV header: ${csvHeader}
Only include the CSV data in your response, no explanations or additional text.`,
{
handler: `batch-${batchIndex}-handler`,
stream: false,
openChat: false
}
);
} catch (err) {
showError("Error sending message: " + err);
}
};
// Start generating batches
generateBatch(0);
}
function updateGenerationProgress(current, total) {
rowsCount.textContent = current;
const percentage = Math.min(100, Math.floor((current / total) * 100));
progressBar.style.width = `${percentage}%`;
}
function showFullDataset() {
// Hide generation status
generateStatus.classList.add('hidden');
// Show full dataset section
fullDatasetSection.classList.remove('hidden');
// Get headers from the data
const headers = Object.keys(fullDatasetRows[0] || {});
// Create and display the table
createTable(fullTable, fullDatasetRows.slice(0, 10), headers);
// Add a note about showing limited rows
const note = document.createElement('p');
note.className = 'text-sm text-gray-600 dark:text-gray-400 mt-2';
note.textContent = `Showing 10 of ${fullDatasetRows.length} rows. Use the download buttons to get the complete dataset.`;
fullTable.appendChild(note);
}
function downloadData(format) {
if (fullDatasetRows.length === 0) return;
const filename = `${currentDataset.name.replace(/\s+/g, '_')}_dataset`;
switch(format) {
case 'csv':
downloadCsv(filename);
break;
case 'json':
downloadJson(filename);
break;
case 'parquet':
// Show a notification that this format is simulated
showNotification("Parquet format download simulated - actual conversion would require a server component");
downloadJson(filename + "_parquet_simulated");
break;
}
}
function downloadCsv(filename) {
// Convert data to CSV
const csv = Papa.unparse(fullDatasetRows);
// Create a blob and download link
const blob = new Blob([csv], { type: 'text/csv' });
const url = URL.createObjectURL(blob);
const a = document.createElement('a');
a.href = url;
a.download = `${filename}.csv`;
document.body.appendChild(a);
a.click();
// Clean up
setTimeout(() => {
document.body.removeChild(a);
URL.revokeObjectURL(url);
}, 100);
}
function downloadJson(filename) {
// Convert data to JSON
const json = JSON.stringify(fullDatasetRows, null, 2);
// Create a blob and download link
const blob = new Blob([json], { type: 'application/json' });
const url = URL.createObjectURL(blob);
const a = document.createElement('a');
a.href = url;
a.download = `${filename}.json`;
document.body.appendChild(a);
a.click();
// Clean up
setTimeout(() => {
document.body.removeChild(a);
URL.revokeObjectURL(url);
}, 100);
}
function showSearchPage() {
searchPage.classList.remove('hidden');
datasetPage.classList.add('hidden');
}
function showError(message) {
console.error(message);
showNotification(message, true);
}
function showNotification(message, isError = false) {
const notification = document.createElement('div');
notification.className = `fixed bottom-4 right-4 px-6 py-3 rounded-lg shadow-lg ${
isError
? 'bg-red-500 text-white'
: 'bg-green-500 text-white'
} z-50 transition-opacity duration-300`;
notification.textContent = message;
document.body.appendChild(notification);
setTimeout(() => {
notification.style.opacity = '0';
setTimeout(() => {
document.body.removeChild(notification);
}, 300);
}, 3000);
}
function showPlaceholderDatasets() {
const placeholders = [
{
name: "NewsEventsPredict",
tags: ["classification", "media", "trend"],
isReal: true,
engine: "AlltheInternet.com"
},
{
name: "FinancialForecast",
tags: ["economy", "stocks", "regression"],
isReal: false
},
{
name: "HealthMonitor",
tags: ["science", "real-time", "anomaly detection"],
isReal: true,
engine: "DuckDuckGo.com"
},
{
name: "SportsAnalysis",
tags: ["classification", "performance", "player tracking"],
isReal: false
},
{
name: "RetailSalesAnalyzer",
tags: ["consumer behavior", "sales trend", "segmentation"],
isReal: true,
engine: "Bing.com"
},
{
name: "SocialMediaSentiment",
tags: ["text classification", "opinion mining", "NLP"],
isReal: false
}
];
currentDatasets = placeholders;
displayDatasets(placeholders);
loadMoreContainer.classList.remove('hidden');
}
</script>
</body>
</html>
"""
# --- Gradio CSS ---
css = """
a { color: var(--body-text-color); }
.datasetButton { justify-content: start; justify-content: left; }
.tags { font-size: var(--button-small-text-size); color: var(--body-text-color-subdued); }
.topButton {
justify-content: start; justify-content: left; text-align: left; background: transparent;
box-shadow: none; padding-bottom: 0;
}
.topButton::before {
content: url("data:image/svg+xml,%3Csvg style='color: rgb(209 213 219)' xmlns='http://www.w3.org/2000/svg' xmlns:xlink='http://www.w3.org/1999/xlink' aria-hidden='true' focusable='false' role='img' width='1em' height='1em' preserveAspectRatio='xMidYMid meet' viewBox='0 0 25 25'%3E%3Cellipse cx='12.5' cy='5' fill='currentColor' fill-opacity='0.25' rx='7.5' ry='2'%3E%3C/ellipse%3E%3Cpath d='M12.5 15C16.6421 15 20 14.1046 20 13V20C20 21.1046 16.6421 22 12.5 22C8.35786 22 5 21.1046 5 20V13C5 14.1046 8.35786 15 12.5 15Z' fill='currentColor' opacity='0.5'%3E%3C/path%3E%3Cpath d='M12.5 7C16.6421 7 20 6.10457 20 5V11.5C20 12.6046 16.6421 13.5 12.5 13.5C8.35786 13.5 5 12.6046 5 11.5V5C5 6.10457 8.35786 7 12.5 7Z' fill='currentColor' opacity='0.5'%3E%3C/path%3E%3Cpath d='M5.23628 12C5.08204 12.1598 5 12.8273 5 13C5 14.1046 8.35786 15 12.5 15C16.6421 15 20 14.1046 20 13C20 12.8273 19.918 12.1598 19.7637 12C18.9311 12.8626 15.9947 13.5 12.5 13.5C9.0053 13.5 6.06886 12.8626 5.23628 12Z' fill='currentColor'%3E%3C/path%3E%3C/svg%3E");
margin-right: .25rem; margin-left: -.125rem; margin-top: .25rem;
}
.bottomButton {
justify-content: start; justify-content: left; text-align: left; background: transparent;
box-shadow: none; font-size: var(--button-small-text-size); color: var(--body-text-color-subdued);
padding-top: 0; align-items: baseline;
}
.bottomButton::before { content: 'tags:'; margin-right: .25rem; }
.buttonsGroup { background: transparent; }
.buttonsGroup:hover { background: var(--input-background-fill); }
.buttonsGroup div { background: transparent; }
.insivibleButtonGroup { display: none; }
@keyframes placeHolderShimmer { 0%{ background-position: -468px 0 } 100%{ background-position: 468px 0 } }
.linear-background {
animation-duration: 1s; animation-fill-mode: forwards; animation-iteration-count: infinite;
animation-name: placeHolderShimmer; animation-timing-function: linear;
background-image: linear-gradient(to right, var(--body-text-color-subdued) 8%, #dddddd11 18%, var(--body-text-color-subdued) 33%);
background-size: 1000px 104px; color: transparent; background-clip: text;
}
.settings { background: transparent; }
.settings button span { color: var(--body-text-color-subdued); }
"""
# --- Knowledge Base ---
class KnowledgeBase:
"""Manages known entities (materials, colors) and patterns for data refinement."""
def __init__(self):
self.materials: Set[str] = {'Metal', 'Wood', 'Plastic', 'Aluminum', 'Bronze', 'Steel', 'Glass', 'Leather', 'Fabric'}
self.colors: Set[str] = {'Red', 'Black', 'White', 'Silver', 'Bronze', 'Yellow', 'Blue', 'Green', 'Gray', 'Brown'}
self.patterns: Dict[str, List[str]] = {}
self.source_data: Dict[str, Any] = {}
def load_source(self, source_type: str, source_path: str) -> None:
"""Loads data from various sources and extracts knowledge."""
try:
if source_type == 'csv_url':
response = requests.get(source_path, timeout=10)
response.raise_for_status()
df = pd.read_csv(io.StringIO(response.text))
elif source_type == 'xlsx_url':
response = requests.get(source_path, timeout=10)
response.raise_for_status()
df = pd.read_excel(io.BytesIO(response.content))
elif source_type == 'local_csv':
df = pd.read_csv(source_path)
elif source_type == 'local_xlsx':
df = pd.read_excel(source_path)
else:
raise ValueError(f"Unsupported source type: {source_type}")
self._extract_knowledge(df)
self.source_data[source_path] = df.to_dict('records')
except requests.exceptions.RequestException as e:
raise ConnectionError(f"Failed to fetch data from URL: {e}")
except ValueError as e: raise e
except Exception as e:
raise RuntimeError(f"Error loading source {source_path}: {str(e)}")
def _extract_knowledge(self, df: pd.DataFrame) -> None:
"""Extracts known materials, colors, and column patterns."""
for column in df.columns:
if 'material' in column.lower():
values = df[column].dropna().unique()
self.materials.update(v.title() for v in values if isinstance(v, str))
elif 'color' in column.lower():
values = df[column].dropna().unique()
self.colors.update(v.title() for v in values if isinstance(v, str))
if df[column].dtype == 'object': # Store string patterns for fuzzy matching
patterns = df[column].dropna().astype(str).tolist()
self.patterns[column] = patterns
def get_closest_match(self, value: str, field_type: str) -> Optional[str]:
"""Finds the closest known value (material or color) for fuzzy matching."""
known_values = getattr(self, field_type + 's', set())
if not known_values: return None
matches = get_close_matches(value.title(), list(known_values), n=1, cutoff=0.8)
return matches[0] if matches else None
knowledge_base = KnowledgeBase() # Global instance for refinement
# --- Data Refinement Utilities ---
def split_compound_field(field: str) -> List[str]:
"""Splits strings like 'Red, Blue' into ['Red', 'Blue']."""
parts = re.split(r'[,;\n]+', field)
return list(set(p.strip().title() for p in parts if p.strip()))
def normalize_value(value: Any, field_name: str, mode: str = 'sourceless', kb: Optional[KnowledgeBase] = None) -> Any:
"""Normalizes a single data value based on field name and refinement mode."""
if not isinstance(value, str): return value
value = re.sub(r'\s+', ' ', value.strip()) # Normalize whitespace
value = value.replace('_', ' ') # Replace underscores
# Field-specific normalization logic
if any(term in field_name.lower() for term in ['material']):
parts = split_compound_field(value)
if mode == 'sourced' and kb:
known = [kb.get_closest_match(p, 'material') or p.title() for p in parts]
else:
known = [m for m in parts if m in kb.materials] if kb else parts
return known[0] if len(known) == 1 else known
elif any(term in field_name.lower() for term in ['color']):
parts = split_compound_field(value)
if mode == 'sourced' and kb:
known = [kb.get_closest_match(p, 'color') or p.title() for p in parts]
else:
known = [c for c in parts if c in kb.colors] if kb else parts
return known[0] if len(known) == 1 else known
elif any(term in field_name.lower() for term in ['date', 'time']): return value # Placeholder
elif any(term in field_name.lower() for term in ['type', 'status', 'category', 'description']):
return value.title() # Title case for descriptive fields
return value
def clean_record(record: Dict[str, Any], mode: str = 'sourceless', kb: Optional[KnowledgeBase] = None) -> Dict[str, Any]:
"""Cleans and normalizes a single record, handling nesting and compound fields."""
cleaned = {}
compound_fields_to_split = {}
# Pass 1: Normalize values and identify compound fields
for key, value in record.items():
clean_key = key.strip().lower().replace(" ", "_")
if isinstance(value, str): # Detect potential compound fields
for material in knowledge_base.materials:
if material.lower() in value.lower():
compound_fields_to_split[clean_key] = value
break
# Recursively clean nested structures
if isinstance(value, list):
cleaned[clean_key] = [normalize_value(v, clean_key, mode, kb) for v in value]
elif isinstance(value, dict):
cleaned[clean_key] = clean_record(value, mode, kb)
else:
cleaned[clean_key] = normalize_value(value, clean_key, mode, kb)
# Pass 2: Split identified compound fields
for key, value in compound_fields_to_split.items():
parts = split_compound_field(value)
materials = [p for p in parts if p in knowledge_base.materials]
if materials:
cleaned['material'] = materials[0] if len(materials) == 1 else materials
remaining = [p for p in parts if p not in materials]
if remaining: cleaned['condition'] = ' '.join(remaining)
elif key not in cleaned: # If not processed and no known materials found
cleaned[key] = value
return cleaned
def refine_data_generic(dataset: List[Dict[str, Any]], mode: str = 'sourceless', kb: Optional[KnowledgeBase] = None) -> List[Dict[str, Any]]:
"""Applies generic data refinement to a list of records, with optional knowledge base guidance."""
if mode == 'sourced' and kb and kb.patterns: # Apply fuzzy matching if sourced
for record in dataset:
for field, patterns in kb.patterns.items():
if field in record and isinstance(record[field], str):
value = str(record[field])
matches = get_close_matches(value, patterns, n=1, cutoff=0.8)
if matches: record[field] = matches[0]
return [clean_record(entry, mode, kb) for entry in dataset]
def refine_preview_data(df: pd.DataFrame, mode: str = 'sourceless') -> pd.DataFrame:
"""Refines the preview DataFrame based on the selected mode."""
# Remove common auto-generated index columns
cols_to_drop = []
for col_name, values in df.to_dict(orient="series").items():
try:
if all(isinstance(v, int) and v == i for i, (v, _) in enumerate(zip(values, df.index))): cols_to_drop.append(col_name)
elif all(isinstance(v, int) and v == i + 1 for i, (v, _) in enumerate(zip(values, df.index))): cols_to_drop.append(col_name)
except Exception: pass # Ignore non-sequential columns
if cols_to_drop: df = df.drop(columns=cols_to_drop)
records = df.to_dict('records')
refined_records = refine_data_generic(records, mode=mode, kb=knowledge_base)
return pd.DataFrame(refined_records)
def detect_anomalies(record: Dict[str, Any]) -> List[str]:
"""Detects potential data quality issues (e.g., verbosity, missing values)."""
flags = []
for k, v in record.items():
if isinstance(v, str):
if len(v) > 300: flags.append(f"{k}: Too verbose.")
if v.lower() in ['n/a', 'none', 'undefined', 'null', '']: flags.append(f"{k}: Missing value.")
return flags
def parse_preview_df(content: str) -> tuple[str, pd.DataFrame]:
"""Extracts CSV from response, parses, refines, and adds quality flags."""
csv_lines = []
in_csv_block = False
for line in content.split("\n"): # Extract lines within CSV code blocks
if line.strip().startswith("```csv") or line.strip().startswith("```"): in_csv_block = True; continue
if line.strip().startswith("```"): in_csv_block = False; continue
if in_csv_block: csv_lines.append(line)
csv_content = "\n".join(csv_lines)
if not csv_content: raise ValueError("No CSV content found.")
csv_header = csv_content.split("\n")[0] if csv_content else ""
df = parse_csv_df(csv_content)
refined_df = refine_preview_data(df, mode='sourceless') # Initial refinement
# Add quality flags
refined_records = refined_df.to_dict('records')
for record in refined_records:
flags = detect_anomalies(record)
if flags: record['_quality_flags'] = flags
return csv_header, pd.DataFrame(refined_records)
def parse_csv_df(csv: str, csv_header: Optional[str] = None) -> pd.DataFrame:
"""Safely parses CSV data using pandas with error handling and common fixes."""
csv = re.sub(r'''(?!")$$(["'][\w\s]+["'][, ]*)+$$(?!")''', lambda m: '"' + m.group(0).replace('"', "'") + '"', csv) # Fix unquoted lists
if csv_header and csv.strip() and not csv.strip().startswith(csv_header.split(',')[0]): csv = csv_header + "\n" + csv # Prepend header if missing
try: return pd.read_csv(io.StringIO(csv), skipinitialspace=True)
except Exception as e: raise ValueError(f"Pandas CSV parsing error: {e}")
# --- LLM Interaction Utilities ---
T = TypeVar("T")
def batched(it: Iterable[T], n: int) -> Iterator[list[T]]:
"""Yields chunks of size n from an iterable."""
it = iter(it)
while batch := list(islice(it, n)): yield batch
def stream_response(msg: str, history: list[Dict[str, str]] = [], max_tokens=500) -> Iterator[str]:
"""Streams responses from the LLM client with retry logic."""
messages = [{"role": m["role"], "content": m["content"]} for m in history]
messages.append({"role": "user", "content": msg})
for attempt in range(3): # Retry mechanism
try:
for chunk in client.chat_completion(messages=messages, max_tokens=max_tokens, stream=True, top_p=0.8, seed=42):
content = chunk.choices[0].delta.content
if content: yield content
break # Success
except (requests.exceptions.ConnectionError, requests.exceptions.Timeout) as e:
print(f"LLM connection error (attempt {attempt+1}): {e}. Retrying in {2**attempt}s...")
time.sleep(2**attempt)
except Exception as e:
print(f"Unexpected LLM error (attempt {attempt+1}): {e}. Retrying...")
time.sleep(2**attempt)
def generate_dataset_names(search_query: str, history: list[Dict[str, str]], is_real_data: bool = False, engine: Optional[str] = None) -> Iterator[str]:
"""Generates dataset names based on a search query using the LLM."""
query = search_query[:1000] if search_query else ""
if is_real_data and engine:
prompt = (
f"@Claude-3.7-Sonnet You are a data specialist who can transform real search results into structured datasets. "
f"A user is searching for data about: \"{query}\" "
f"Imagine you've queried {engine} and received real search results. Create a list of {MAX_NB_ITEMS_PER_GENERATION_CALL} specific datasets that could be created from these search results. "
f"For each dataset: 1. Give it a clear, specific name related to the search topic. 2. Include 3-5 relevant tags in parentheses, with one tag specifying the ML task type (classification, regression, clustering, etc.). "
f"Format each dataset as: 1. DatasetName (tag1, tag2, ml_task_tag). Make these datasets sound like real collections that could be created from {engine} search results on \"{query}\"."
)
else:
prompt = GENERATE_DATASET_NAMES_FOR_SEARCH_QUERY.format(search_query=query)
full_response = ""
for token in stream_response(prompt, history):
full_response += token
yield token # Yield tokens for real-time display
print(f"Generated dataset names for query '{search_query}'.")
history.append({"role": "assistant", "content": full_response}) # Update history
# No return needed as history is modified in place
def generate_dataset_content(search_query: str, dataset_name: str, tags: str, history: list[Dict[str, str]], is_real_data: bool = False, engine: Optional[str] = None) -> Iterator[str]:
"""Generates the description and CSV preview for a dataset."""
query = search_query[:1000] if search_query else ""
if is_real_data and engine:
prompt = (
f"@Claude-3.7-Sonnet You're a specialist in converting web search results into structured data. "
f"Based on search results from {engine} about \"{query}\", create a preview of the dataset \"{dataset_name}\" with tags \"{tags}\". "
f"First, write a detailed description of what this dataset contains, its structure, and how it was constructed from web search results. "
f"Then, generate a realistic 5-row CSV preview that resembles data you might get if you scraped and structured real results from {engine}. "
f"Format your response with: **Dataset Description:** [detailed description] **CSV Content Preview:** ```csv [CSV header and 5 rows of realistic data] ``` "
f"Include relevant columns for the dataset type, with proper labels/categories where appropriate. The data should look like it came from real sources."
)
else:
prompt = GENERATE_DATASET_CONTENT_FOR_SEARCH_QUERY_AND_NAME_AND_TAGS.format(
search_query=query, dataset_name=dataset_name, tags=tags
)
full_response = ""
for token in stream_response(prompt, history):
full_response += token
yield token
print(f"Generated content for dataset '{dataset_name}'.")
history.append({"role": "assistant", "content": full_response}) # Update history
def _write_generator_to_queue(queue: Queue, func: Callable, kwargs: dict) -> None:
"""Helper to run a generator and put results (or errors) into a queue."""
try:
for i, result in enumerate(func(**kwargs)): queue.put((i, result))
except Exception as e: queue.put((-1, str(e))) # Signal error with index -1
finally: queue.put(None) # Signal completion
def iflatmap_unordered(func: Callable, kwargs_iterable: Iterable[dict]) -> Iterable[Any]:
"""Runs generator functions concurrently and yields results as they complete."""
queue = Queue()
pool_size = min(len(kwargs_iterable), os.cpu_count() or 4)
with ThreadPool(pool_size) as pool:
async_results = [pool.apply_async(_write_generator_to_queue, (queue, func, kwargs)) for kwargs in kwargs_iterable]
completed_generators = 0
while completed_generators < len(async_results):
try:
result = queue.get(timeout=0.1)
if result is None: # Generator finished
completed_generators += 1
continue
index, data = result
if index == -1: # Error occurred
print(f"Generator error: {data}")
continue # Skip this result
yield data # Yield successful result
except Empty: # Timeout occurred, check if all threads are done
if all(res.ready() for res in async_results) and queue.empty(): break
for res in async_results: res.get(timeout=0.1) # Ensure threads finish and raise exceptions
def generate_partial_dataset(
title: str, content: str, search_query: str, variant: str, csv_header: str,
output: list[Optional[dict]], indices_to_generate: list[int], history: list[Dict[str, str]],
is_real_data: bool = False, engine: Optional[str] = None
) -> Iterator[int]:
"""Generates a batch of dataset rows for a specific variant."""
dataset_name, tags = title.strip("# ").split("\ntags:", 1)
dataset_name, tags = dataset_name.strip(), tags.strip()
prompt = GENERATE_MORE_ROWS.format(csv_header=csv_header) + " " + variant
# Construct initial messages for context
initial_prompt = ""
if is_real_data and engine:
initial_prompt = (
f"@Claude-3.7-Sonnet You're a specialist in converting web search results into structured data. "
f"Based on search results from {engine} about \"{search_query}\", create a preview of the dataset \"{dataset_name}\" with tags \"{tags}\". "
f"First, write a detailed description of what this dataset contains, its structure, and how it was constructed from web search results. "
f"Then, generate a realistic 5-row CSV preview that resembles data you might get if you scraped and structured real results from {engine}. "
f"Format your response with: **Dataset Description:** [detailed description] **CSV Content Preview:** ```csv [CSV header and 5 rows of realistic data] ``` "
f"Include relevant columns for the dataset type, with proper labels/categories where appropriate. The data should look like it came from real sources."
)
else:
initial_prompt = GENERATE_DATASET_CONTENT_FOR_SEARCH_QUERY_AND_NAME_AND_TAGS.format(
search_query=search_query, dataset_name=dataset_name, tags=tags
)
messages = [
{"role": "user", "content": initial_prompt},
{"role": "assistant", "content": title + "\n\n" + content},
{"role": "user", "content": prompt},
]
generated_samples = 0
current_csv_chunk = ""
in_csv_block = False
for attempt in range(3): # Retry logic
try:
for chunk in client.chat_completion(messages=messages, max_tokens=1500, stream=True, top_p=0.8, seed=42):
token = chunk.choices[0].delta.content
if not token: continue
current_csv_chunk += token
# Detect CSV block start/end
if token.strip().startswith("```csv") or token.strip().startswith("```"):
in_csv_block = True
continue
if token.strip().startswith("```"):
in_csv_block = False
if current_csv_chunk.strip(): # Process accumulated chunk if block just ended
try:
temp_df = parse_csv_df(current_csv_chunk.strip(), csv_header=csv_header)
new_rows = temp_df.iloc[generated_samples:].to_dict('records')
for i, record in enumerate(new_rows):
if generated_samples >= len(indices_to_generate): break
refined_record = refine_data_generic([record])[0]
flags = detect_anomalies(refined_record)
if flags: refined_record['_quality_flags'] = flags
output_index = indices_to_generate[generated_samples]
if output_index < len(output):
output[output_index] = refined_record
generated_samples += 1
yield 1 # Signal progress
except ValueError as e: print(f"CSV parsing error: {e}")
except Exception as e: print(f"CSV chunk processing error: {e}")
finally: current_csv_chunk = "" # Reset chunk
continue
if in_csv_block: # Process incrementally if inside CSV block
try:
temp_df = parse_csv_df(current_csv_chunk.strip(), csv_header=csv_header)
new_rows = temp_df.iloc[generated_samples:].to_dict('records')
for i, record in enumerate(new_rows):
if generated_samples >= len(indices_to_generate): break
refined_record = refine_data_generic([record])[0]
flags = detect_anomalies(refined_record)
if flags: refined_record['_quality_flags'] = flags
output_index = indices_to_generate[generated_samples]
if output_index < len(output):
output[output_index] = refined_record
generated_samples += 1
yield 1
except ValueError: pass # CSV not complete
except Exception as e: print(f"Incremental CSV processing error: {e}")
if generated_samples >= len(indices_to_generate): break # Target reached
print(f"Retrying generation for variant '{variant}' (attempt {attempt+1})...")
time.sleep(2**attempt)
except (requests.exceptions.ConnectionError, requests.exceptions.Timeout) as e:
print(f"Connection error (attempt {attempt+1}): {e}. Retrying...")
time.sleep(2**attempt)
except Exception as e:
print(f"Unexpected error (attempt {attempt+1}): {e}. Retrying...")
time.sleep(2**attempt)
def generate_variants(preview_df: pd.DataFrame) -> Iterator[str]:
"""Generates diverse prompts for creating dataset variants."""
label_cols = [col for col in preview_df.columns if "label" in col.lower()]
labels = preview_df[label_cols[0]].unique() if label_cols and len(preview_df[label_cols[0]].unique()) > 1 else []
if labels: # Prioritize label-based generation
rarities = ["pretty obvious", "common/regular", "unexpected but useful", "uncommon but still plausible", "rare/niche but still plausible"]
for rarity in rarities:
for label in labels: yield GENERATE_VARIANTS_WITH_RARITY_AND_LABEL.format(rarity=rarity, label=label)
else: # Fallback to general rarity prompts
rarities = ["obvious", "expected", "common", "regular", "unexpected but useful", "original but useful", "specific but not far-fetched", "uncommon but still plausible", "rare but still plausible", "very niche but still plausible"]
for rarity in rarities: yield GENERATE_VARIANTS_WITH_RARITY.format(rarity=rarity)
# --- Gradio Interface ---
def whoami(token: str) -> Dict[str, Any]:
"""Fetches user information from Hugging Face Hub API."""
try:
response = requests.get("https://huggingface.co/api/users/me", headers={"Authorization": f"Bearer {token}"}, timeout=5)
response.raise_for_status()
return response.json()
except (requests.exceptions.RequestException, ValueError) as e:
print(f"Error fetching user info: {e}")
return {"name": "User", "orgs": []}
def get_repo_visibility(repo_id: str, token: str) -> str:
"""Determines if a Hugging Face repository is public or private."""
try:
response = requests.get(f"https://huggingface.co/api/repos/{repo_id}", headers={"Authorization": f"Bearer {token}"}, timeout=5)
response.raise_for_status()
return "public" if not response.json().get("private", False) else "private"
except HfHubHTTPError as e:
if e.response.status_code == 404: return "public" # Assume public if repo doesn't exist
print(f"Error checking repo visibility for {repo_id}: {e}")
return "public"
except Exception as e:
print(f"Unexpected error checking repo visibility for {repo_id}: {e}")
return "public"
with gr.Blocks(css=css) as demo:
generated_texts_state = gr.State((landing_page_datasets_generated_text,)) # State for generated dataset names
current_dataset_state = gr.State(None) # State to hold current dataset details for generation
is_real_data_state = gr.State(True) # State to track if real data is being used
current_engine_state = gr.State(None) # State to track the current search engine
selected_engines_state = gr.State(["DuckDuckGo.com", "Bing.com", "Search.Yahoo.com", "Search.Brave.com", "Ecosia.org"]) # Default selected engines
searchEngines = ["AlltheInternet.com", "DuckDuckGo.com", "Google.com", "Bing.com", "Search.Yahoo.com", "Startpage.com", "Qwant.com", "Ecosia.org", "WolframAlpha.com", "Mojeek.co.uk", "Search.Brave.com", "Yandex.com", "Baidu.com", "Gibiru.com", "MetaGer.org", "Swisscows.com", "Presearch.com", "Ekoru.org", "Search.Lilo.org"]
# --- Search Page UI ---
with gr.Column(visible=True, elem_id="search-page") as search_page:
gr.Markdown("# 🤗 Infinite Dataset Hub ♾️\n\nAn endless catalog of datasets, created just for you by an AI model.")
with gr.Row():
search_bar = gr.Textbox(max_lines=1, placeholder="Search datasets, get infinite results", show_label=False, container=False, scale=9)
search_button = gr.Button("🔍", variant="primary", scale=1)
button_groups: list[gr.Group] = [] # Holds the groups for dataset buttons
buttons: list[gr.Button] = [] # Holds the actual dataset name and tag buttons
for i in range(MAX_TOTAL_NB_ITEMS):
if i < len(default_output): # Use default datasets initially
line = default_output[i]
try: dataset_name, tags = line.split(".", 1)[1].strip(" )").split(" (", 1)
except ValueError: dataset_name, tags = line.split(".", 1)[1].strip(" )").split(" ", 1)[0], ""
group_classes, name_classes, tag_classes = "buttonsGroup", "topButton", "bottomButton"
else: # Placeholders for future datasets
dataset_name, tags = "⬜⬜⬜⬜⬜⬜", "░░░░, ░░░░, ░░░░"
group_classes, name_classes, tag_classes = "buttonsGroup insivibleButtonGroup", "topButton linear-background", "bottomButton linear-background"
with gr.Group(elem_classes=group_classes) as button_group:
button_groups.append(button_group)
dataset_btn = gr.Button(dataset_name, elem_classes=name_classes)
tags_btn = gr.Button(tags, elem_classes=tag_classes)
buttons.append(dataset_btn)
buttons.append(tags_btn)
load_more_datasets = gr.Button("Load more datasets")
gr.Markdown(f"_powered by [{model_id}](https://huggingface.co/{model_id})_")
# --- Settings Panel ---
with gr.Column(scale=4, min_width="200px"):
with gr.Accordion("Settings", open=False, elem_classes="settings"):
gr.Markdown("Manage your Hugging Face account and dataset saving options.")
gr.LoginButton()
select_namespace_dropdown = gr.Dropdown(choices=[NAMESPACE], value=NAMESPACE, label="Hugging Face Namespace", visible=False)
gr.Markdown("Dataset Generation Mode")
refinement_mode = gr.Radio(
["sourceless", "sourced"], value="sourceless", label="Refinement Mode",
info="Sourceless: AI generates data freely. Sourced: AI uses loaded data for context and refinement."
)
with gr.Group(visible=False) as source_group: # Dynamic section for source loading
source_type = gr.Dropdown(
choices=["csv_url", "xlsx_url", "local_csv", "local_xlsx"], value="csv_url",
label="Source Type", info="Select the format of your data source."
)
source_path = gr.Textbox(
label="Source Path/URL", placeholder="Enter URL or local file path",
info="Provide the location of your dataset file."
)
load_source_button = gr.Button("Load Source Data", icon="https://huggingface.co/datasets/huggingface/badges/resolve/main/badge-files/data.svg")
source_status = gr.Markdown("", visible=False)
visibility_radio = gr.Radio(
["public", "private"], value="public", container=False, interactive=False,
label="Dataset Visibility", info="Set visibility for datasets saved to Hugging Face Hub."
)
# Search Engine Settings
gr.Markdown("Search Engine Configuration")
data_source_toggle = gr.Checkbox(label="Use Real Search Data", value=True, info="Toggle to include results from real search engines.")
engine_settings_button = gr.Button("Configure Search Engines", icon="https://img.icons8.com/ios-filled/50/000000/settings--v1.png", size="sm")
# Engine Selection Modal
with gr.Modal("Search Engine Settings", id="engine-modal") as engine_modal:
gr.Markdown("Select which search engines to use for real data retrieval. A diverse selection improves results.")
engine_options_html_comp = gr.HTML(elem_id="engine-options")
with gr.Row():
select_all_engines_btn = gr.Button("Select All")
deselect_all_engines_btn = gr.Button("Deselect All")
save_engines_btn = gr.Button("Save Settings", variant="primary")
# --- Dataset Detail Page UI ---
with gr.Column(visible=False, elem_id="dataset-page") as dataset_page:
gr.Markdown("# 🤗 Infinite Dataset Hub ♾️\n\nAn endless catalog of datasets, created just for you.")
dataset_title_md = gr.Markdown() # Dataset name and tags
dataset_source_badge = gr.Markdown() # Badge indicating real/AI data
dataset_source_info = gr.Markdown() # Details about the data source
dataset_description_md = gr.Markdown() # Dataset description
preview_table_comp = gr.DataFrame(visible=False, interactive=False, wrap=True) # Holds the preview CSV
with gr.Row():
generate_full_dataset_button = gr.Button("Generate Full Dataset", variant="primary")
save_dataset_button = gr.Button("💾 Save Dataset", variant="primary", visible=False)
open_dataset_message = gr.Markdown("", visible=False) # Confirmation message
dataset_share_button = gr.Button("Share Dataset URL")
dataset_share_textbox = gr.Textbox(visible=False, show_copy_button=True, label="Copy this URL:", interactive=False, show_label=True)
full_dataset_section = gr.Column(visible=False) # Container for full dataset and downloads
full_table_comp = gr.DataFrame(visible=False, interactive=False, wrap=True)
with gr.Row():
download_csv_button = gr.Button("Download CSV")
download_json_button = gr.Button("Download JSON")
download_parquet_button = gr.Button("Download Parquet")
back_button = gr.Button("< Back", size="sm")
# --- Event Handlers ---
# Search Logic
def _update_search_results(search_query: str, current_generated_texts: tuple[str], is_real_data: bool, engine: Optional[str]):
"""Handles dataset search and UI updates."""
# Reset UI to loading state
yield {btn: gr.Button("⬜⬜⬜⬜⬜⬜", elem_classes="topButton linear-background") for btn in buttons[::2]}
yield {btn: gr.Button("░░░░, ░░░░, ░░░░", elem_classes="bottomButton linear-background") for btn in buttons[1::2]}
yield {group: gr.Group(elem_classes="buttonsGroup insivibleButtonGroup") for group in button_groups}
generated_count = 0
new_texts = ""
try:
# Generate dataset names from LLM
for line in generate_dataset_names(search_query, [], is_real_data=is_real_data, engine=engine):
if "I'm sorry" in line or "policy" in line: raise gr.Error("Inappropriate content detected.")
if generated_count >= MAX_NB_ITEMS_PER_GENERATION_CALL: break
match = re.match(r"^\s*\d+\.\s+(.+?)\s+$$(.+?)$$", line) # Parse line format
if match:
dataset_name, tags = match.groups()
dataset_name, tags = dataset_name.strip(), tags.strip()
new_texts += line
# Update buttons with generated data
yield {
buttons[2 * generated_count]: gr.Button(dataset_name, elem_classes="topButton"),
buttons[2 * generated_count + 1]: gr.Button(tags, elem_classes="bottomButton"),
}
generated_count += 1
# Update state and make new buttons visible
new_history = (current_generated_texts + (new_texts,)) if current_generated_texts else (landing_page_datasets_generated_text + "\n" + new_texts,)
yield {generated_texts_state: new_history}
yield {group: gr.Group(elem_classes="buttonsGroup") for group in button_groups[:generated_count]}
except gr.Error as e: raise e # Propagate Gradio errors
except Exception as e: raise gr.Error(f"Failed to generate datasets: {str(e)}")
# Attach search handlers
search_button.click(
_update_search_results,
inputs=[search_bar, generated_texts_state, is_real_data_state, current_engine_state],
outputs=buttons + [generated_texts_state] + button_groups
)
search_bar.submit(
_update_search_results,
inputs=[search_bar, generated_texts_state, is_real_data_state, current_engine_state],
outputs=buttons + [generated_texts_state] + button_groups
)
# Load More Datasets
load_more_datasets.click(
_update_search_results,
inputs=[search_bar, generated_texts_state, is_real_data_state, current_engine_state],
outputs=buttons + [generated_texts_state] + button_groups
)
# Display Single Dataset Details
def _show_dataset_details(search_query, dataset_name, tags, is_real_data, engine):
"""Switches to detail view and loads dataset content."""
yield {
search_page: gr.Column(visible=False), dataset_page: gr.Column(visible=True),
dataset_title_md: f"# {dataset_name}\n\n tags: {tags}",
dataset_share_textbox: gr.Textbox(visible=False),
full_dataset_section: gr.Column(visible=False),
save_dataset_button: gr.Button(visible=False),
open_dataset_message: gr.Markdown("", visible=False)
}
# Update source badge and info
if is_real_data:
badge_html = gr.Markdown(f'<span class="px-3 py-1 rounded-full text-xs font-medium bg-green-100 text-green-800 dark:bg-green-900 dark:text-green-200">Real Data</span>', visible=True)
info_html = gr.Markdown(f'This dataset is based on real information queried from <strong>{engine}</strong> for the search term "<strong>{search_query}</strong>". The data has been structured for machine learning use.', visible=True)
else:
badge_html = gr.Markdown('<span class="px-3 py-1 rounded-full text-xs font-medium bg-purple-100 text-purple-800 dark:bg-purple-900 dark:text-purple-200">AI-Generated</span>', visible=True)
info_html = gr.Markdown(f'This is an AI-generated dataset created using {model_id}. The content is synthetic and designed to represent plausible data related to "{search_query}".', visible=True)
yield {dataset_source_badge: badge_html, dataset_source_info: info_html}
# Stream content generation
for content_chunk in generate_dataset_content(search_query, dataset_name, tags, [], is_real_data=is_real_data, engine=engine):
yield {dataset_description_md: content_chunk}
# Link buttons to the detail view function
def _show_dataset_from_button_wrapper(search_query, *buttons_values):
# Determine which button was clicked to get the index
clicked_button_index = -1
for i, btn_val in enumerate(buttons_values):
if btn_val is not None and btn_val != "": # Assuming non-empty value indicates the clicked button's text
clicked_button_index = i
break
if clicked_button_index == -1: return # Should not happen if events are correctly wired
# Determine if it was a name button (even index) or tag button (odd index)
dataset_index = clicked_button_index // 2
dataset_name, tags = buttons_values[2 * dataset_index], buttons_values[2 * dataset_index + 1]
is_real_data = current_engine_state.value is not None # Infer from engine state
engine = current_engine_state.value if is_real_data else None
yield from _show_dataset_details(search_query, dataset_name, tags, is_real_data, engine)
# Wire up click events for all dataset name and tag buttons
for i, (name_btn, tag_btn) in enumerate(batched(buttons, 2)):
name_btn.click(
partial(_show_dataset_from_button_wrapper),
inputs=[search_bar, *buttons],
outputs=[search_page, dataset_page, dataset_title_md, dataset_description_md, dataset_source_badge, dataset_source_info, dataset_share_textbox, full_dataset_section, save_dataset_button, open_dataset_message]
)
tag_btn.click(
partial(_show_dataset_from_button_wrapper),
inputs=[search_bar, *buttons],
outputs=[search_page, dataset_page, dataset_title_md, dataset_description_md, dataset_source_badge, dataset_source_info, dataset_share_textbox, full_dataset_section, save_dataset_button, open_dataset_message]
)
# Back Button Navigation
back_button.click(lambda: (gr.Column(visible=True), gr.Column(visible=False)), outputs=[search_page, dataset_page], js="""
function() {
if ('parentIFrame' in window) { window.parentIFrame.scrollTo({top: 0, behavior:'smooth'}); }
else { window.scrollTo({ top: 0, behavior: 'smooth' }); }
return Array.from(arguments);
}
""")
# Full Dataset Generation
@generate_full_dataset_button.click(
inputs=[dataset_title_md, dataset_description_md, search_bar, select_namespace_dropdown, visibility_radio, refinement_mode, is_real_data_state, current_engine_state],
outputs=[full_table_comp, generate_full_dataset_button, save_dataset_button, full_dataset_section]
)
def _generate_full_dataset(title_md, content_md, search_query, namespace, visibility, mode, is_real_data, engine):
# Extract dataset name and tags from the markdown title
try:
dataset_name = title_md.split('\n')[0].strip('# ')
tags = title_md.split('tags:', 1)[1].strip()
except IndexError:
raise gr.Error("Could not parse dataset title.")
try: csv_header, preview_df = parse_preview_df(content_md)
except ValueError as e: raise gr.Error(f"Failed to parse preview: {e}")
refined_preview_df = refine_preview_data(preview_df, mode)
columns = list(refined_preview_df)
output_data: list[Optional[dict]] = [None] * NUM_ROWS # Initialize output structure
initial_rows = refined_preview_df.to_dict('records')
for i, record in enumerate(initial_rows):
if i < NUM_ROWS: output_data[i] = {"idx": i, **record}
# Update UI: show preview, disable generate, show save button
yield {
full_table_comp: gr.DataFrame(pd.DataFrame([r for r in output_data if r]), visible=True),
generate_full_dataset_button: gr.Button(interactive=False),
save_dataset_button: gr.Button(f"💾 Save {namespace}/{dataset_name}" + (" (private)" if visibility != "public" else ""), visible=True, interactive=False),
full_dataset_section: gr.Column(visible=True)
}
# Prepare generation tasks for variants
generation_tasks = []
variants = islice(generate_variants(refined_preview_df), NUM_VARIANTS)
for i, variant in enumerate(variants):
indices = list(range(len(initial_rows) + i, NUM_ROWS, NUM_VARIANTS))
if indices: # Only create task if there are rows to generate
generation_tasks.append({
"func": generate_partial_dataset,
"kwargs": {
"title": title_md, "content": content_md, "search_query": search_query, "variant": variant,
"csv_header": csv_header, "output": output_data, "indices_to_generate": indices,
"history": [], # Use fresh history for each variant task
"is_real_data": is_real_data, "engine": engine
}
})
# Execute tasks in parallel and update UI progressively
for _ in iflatmap_unordered(lambda **kw: kw.pop('func')(**kw), generation_tasks):
yield {full_table_comp: pd.DataFrame([r for r in output_data if r])} # Update DataFrame display
yield {save_dataset_button: gr.Button(interactive=True)} # Enable save button
print(f"Full dataset generation complete for {dataset_name}.")
# Save Dataset to Hugging Face Hub
@save_dataset_button.click(
inputs=[dataset_title_md, dataset_description_md, search_bar, full_table_comp, select_namespace_dropdown, visibility_radio],
outputs=[save_dataset_button, open_dataset_message]
)
def _save_dataset(title_md, content_md, search_query, df, namespace, visibility, oauth_token):
# Extract dataset name and tags from the markdown title
try:
dataset_name = title_md.split('\n')[0].strip('# ')
tags = title_md.split('tags:', 1)[1].strip()
except IndexError:
raise gr.Error("Could not parse dataset title.")
token = oauth_token.token if oauth_token else save_dataset_hf_token
if not token: raise gr.Error("Login required or set SAVE_DATASET_HF_TOKEN.")
repo_id = f"{namespace}/{dataset_name}"
dataset_url_params = f"q={search_query.replace(' ', '+')}&dataset={dataset_name.replace(' ', '+')}&tags={tags.replace(' ', '+')}"
dataset_url = f"{URL}?{dataset_url_params}"
gr.Info("Saving dataset...")
yield {save_dataset_button: gr.Button(interactive=False)} # Disable button during save
try:
create_repo(repo_id=repo_id, repo_type="dataset", private=visibility!="public", exist_ok=True, token=token)
df.to_csv(f"hf://datasets/{repo_id}/data.csv", storage_options={"token": token}, index=False)
card_content = DATASET_CARD_CONTENT.format(title=title_md, content=content_md, url=URL, dataset_url=dataset_url, model_id=model_id, search_query=search_query)
DatasetCard(card_content).push_to_hub(repo_id=repo_id, repo_type="dataset", token=token)
success_msg = f"# 🎉 Yay! Dataset saved to [{repo_id}](https://huggingface.co/datasets/{repo_id})!\n\n_PS: Check Settings to manage your saved datasets._"
gr.Info("Dataset saved successfully.")
yield {open_dataset_message: gr.Markdown(success_msg, visible=True)}
except HfHubHTTPError as e: raise gr.Error(f"HF Hub error: {e.message}")
except Exception as e: raise gr.Error(f"Save failed: {str(e)}")
finally: yield {save_dataset_button: gr.Button(interactive=True)} # Re-enable button
# Shareable URL Generation
@dataset_share_button.click(inputs=[dataset_title_md, search_bar], outputs=[dataset_share_textbox])
def _show_share_url(title_md, search_query):
try:
dataset_name = title_md.split('\n')[0].strip('# ')
tags = title_md.split('tags:', 1)[1].strip()
except IndexError:
raise gr.Error("Could not parse dataset title.")
share_url = f"{URL}?q={search_query.replace(' ', '+')}&dataset={dataset_name.replace(' ', '+')}&tags={tags.replace(' ', '+')}"
return gr.Textbox(share_url, visible=True)
# Settings Toggles
refinement_mode.change(lambda mode: gr.Group(visible=(mode == "sourced")), outputs=[source_group])
data_source_toggle.change(lambda value: (gr.State(value), gr.State(value if value else None)), inputs=[data_source_toggle], outputs=[is_real_data_state, current_engine_state])
@load_source_button.click(inputs=[source_type, source_path], outputs=[source_status])
def _load_source_data(source_type, source_path):
if not source_path: raise gr.Error("Source path/URL is required.")
try:
knowledge_base.load_source(source_type, source_path)
gr.Info("Source data loaded.")
return gr.Markdown("✅ Source loaded successfully", visible=True)
except (ConnectionError, ValueError, RuntimeError) as e:
raise gr.Error(f"Failed to load source: {str(e)}")
# Engine Settings Modal Logic
def _populate_engine_options(selected_engines):
engine_options_html = ""
for engine in searchEngines:
is_checked = "checked" if engine in selected_engines else ""
engine_options_html += f"""
<div class="flex items-center">
<input type="checkbox" id="engine-{engine.replace('.', '_')}" class="engine-checkbox mr-2 h-4 w-4" value="{engine}" {is_checked}>
<label for="engine-{engine.replace('.', '_')}" class="cursor-pointer">{engine}</label>
</div>
"""
return gr.HTML(engine_options_html)
def _save_engine_settings(selected_engines_json):
selected_engines = json.loads(selected_engines_json)
if not selected_engines:
gr.Warning("At least one search engine must be selected. Using DuckDuckGo as default.")
selected_engines = ["DuckDuckGo.com"]
current_engine = selected_engines[0] if selected_engines else None
return gr.State(selected_engines), gr.State(current_engine), gr.Info(f"Updated search engines. Using {len(selected_engines)} engines.")
# Initialize engine options component
engine_options_html_comp = _populate_engine_options(selected_engines_state.value)
# Update engine options when the modal is opened
engine_settings_button.click(lambda: engine_options_html_comp.update(_populate_engine_options(selected_engines_state.value)), outputs=[engine_options_html_comp])
select_all_engines_btn.click(lambda: engine_options_html_comp.update(_populate_engine_options(searchEngines)), outputs=[engine_options_html_comp])
deselect_all_engines_btn.click(lambda: engine_options_html_comp.update(_populate_engine_options([])), outputs=[engine_options_html_comp])
save_engines_btn.click(
_save_engine_settings,
inputs=[gr.JSON(elem_id="engine-options")], # Capture checked engines from modal
outputs=[selected_engines_state, current_engine_state, gr.Info()]
)
engine_settings_button.click(lambda: engine_modal.update(visible=True), outputs=[engine_modal])
# Close modal on save or when clicking outside (implicit via Gradio's modal handling)
# Initial App Load Logic
@demo.load(outputs=([search_page, dataset_page, dataset_title_md, dataset_description_md, dataset_source_badge, dataset_source_info, dataset_share_textbox, full_dataset_section, save_dataset_button, open_dataset_message, search_bar] + # Outputs for detail page and search bar
buttons + [generated_texts_state] + # Outputs for search results buttons and state
[select_namespace_dropdown, visibility_radio, source_group, data_source_toggle, current_engine_state, selected_engines_state, engine_options_html_comp])) # Outputs for settings
def _load_app(request: gr.Request, oauth_token: Optional[gr.OAuthToken]):
# Handle user login and namespace selection
if oauth_token:
try:
user_info = whoami(oauth_token.token)
namespaces = [user_info["name"]] + [org["name"] for org in user_info.get("orgs", [])]
yield {
select_namespace_dropdown: gr.Dropdown(choices=namespaces, value=user_info["name"], visible=True),
visibility_radio: gr.Radio(interactive=True),
}
except Exception: # Fallback if user info fails
yield {
select_namespace_dropdown: gr.Dropdown(choices=[NAMESPACE], value=NAMESPACE, visible=True),
visibility_radio: gr.Radio(interactive=True),
}
else: # Default settings if not logged in
yield {
select_namespace_dropdown: gr.Dropdown(choices=[NAMESPACE], value=NAMESPACE, visible=True),
visibility_radio: gr.Radio(interactive=False),
}
# Handle URL parameters for direct search or dataset loading
query_params = dict(request.query_params)
if "dataset" in query_params:
is_real = query_params.get("engine") is not None
engine = query_params.get("engine")
yield from _show_dataset_details(query_params.get("q", query_params["dataset"]), query_params["dataset"], query_params.get("tags", ""), is_real, engine)
yield {is_real_data_state: is_real, current_engine_state: engine}
elif "q" in query_params:
search_query = query_params["q"]
is_real = query_params.get("engine") is not None
engine = query_params.get("engine")
yield {search_bar: search_query}
yield {is_real_data_state: is_real, current_engine_state: engine}
yield from _update_search_results(search_query, (), is_real, engine)
else:
yield {search_page: gr.Column(visible=True)} # Show search page by default
# Initialize with default datasets
initial_outputs = {}
for i, line in enumerate(default_output):
try: dataset_name, tags = line.split(".", 1)[1].strip(" )").split(" (", 1)
except ValueError: dataset_name, tags = line.split(".", 1)[1].strip(" )").split(" ", 1)[0], ""
initial_outputs[buttons[2 * i]] = gr.Button(dataset_name, elem_classes="topButton")
initial_outputs[buttons[2 * i + 1]] = gr.Button(tags, elem_classes="bottomButton")
initial_outputs[button_groups[i]] = gr.Group(elem_classes="buttonsGroup")
yield initial_outputs
yield {generated_texts_state: (landing_page_datasets_generated_text,)}
# Initialize engine settings UI
yield {
data_source_toggle: gr.Checkbox(value=is_real_data_state.value),
engine_options_html_comp: _populate_engine_options(selected_engines_state.value)
}
if __name__ == "__main__":
demo.launch(share=False, server_name="0.0.0.0") |