File size: 270,353 Bytes
3d9b03b | 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 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625 4626 4627 4628 4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 4723 4724 4725 4726 4727 4728 4729 4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 4767 4768 4769 4770 4771 4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 | import os
import sys
import re
import random
import math
# THIS IS THE FIX - PART 1
os.environ['GRADIO_SUPPRESS_PROGRESS'] = 'true'
# THIS IS THE FIX - PART 2: Clean up console logs from Gradio
import logging
logging.getLogger('gradio').setLevel(logging.ERROR)
import cv2
import numpy as np
import gradio as gr
from gradio import Progress
import shutil
import subprocess
from PIL import Image, ImageDraw, ImageFont, ImageOps, ImageEnhance
from datetime import datetime
from threading import Lock
import base64
import io
# --- Dependency Check ---
try:
from controlnet_aux import (
CannyDetector, MLSDdetector, HEDdetector,
LineartDetector, OpenposeDetector, NormalBaeDetector
)
from gradio_client import Client
from rembg import remove
import librosa
# NEW: Added for the audio chopping feature
from pydub import AudioSegment
from pydub.silence import split_on_silence
except ImportError as e:
print("="*80)
print(f"ERROR: Missing dependency -> {e}")
print("Please install all required packages by running:")
print("pip install -r requirements.txt")
print("(Note: The new feature requires 'pydub'. Make sure it's in your requirements file.)")
print("="*80)
sys.exit(1)
# --- AI Model Dependency Check ---
try:
import whisper
except ImportError:
print("="*80)
print("WARNING: 'openai-whisper' not installed. The Transcription tab will be disabled.")
print("To enable it, run: pip install -U openai-whisper")
print("="*80)
whisper = None
# --- Slo-Mo & Enhance AI Dependency Check (SIMPLIFIED) ---
try:
from rife_ncnn_vulkan_python import Rife
ENHANCE_AI_AVAILABLE = True
except ImportError:
print("="*80)
print("WARNING: 'rife-ncnn-vulkan-python' not found.")
print("The AI-Enhanced option in 'Slo-Mo & Enhance' will be disabled.")
print("To enable it, run: pip install rife-ncnn-vulkan-python")
print("="*80)
Rife = None
ENHANCE_AI_AVAILABLE = False
# --- Global Variables & Setup ---
TEMP_DIR = "temp_gradio"
os.makedirs(TEMP_DIR, exist_ok=True)
model_load_lock = Lock()
loaded_detectors = {}
whisper_model = None
whisper_model_name = ""
rife_model = None
# REMOVED realesrgan_model
# --- Default Presets for Transfer Tab (Flat Dictionary) ---
DEFAULT_LINK_PRESETS = {
# Text To Image
"FLUX.1-schnell (black-forest-labs)": "https://huggingface.co/spaces/black-forest-labs/FLUX.1-schnell",
"FLUX.1-schnell (Rooc)": "https://huggingface.co/spaces/Rooc/FLUX.1-schnell",
"FLUX.1-schnell (evalstate)": "https://huggingface.co/spaces/evalstate/flux1_schnell",
"FLUX.1-schnell (hysts-mcp)": "https://huggingface.co/spaces/hysts-mcp/FLUX.1-schnell",
"FLUX.1-schnell (cbensimon)": "https://huggingface.co/spaces/cbensimon/FLUX-1-schnell-mcp",
"FLUX.1-dev": "https://huggingface.co/spaces/black-forest-labs/FLUX.1-dev",
"FLUX.1-dev-quantized": "https://huggingface.co/spaces/multimodalart/FLUX.1-dev-quantized",
"FLUX.1-dev_NotASI": "https://huggingface.co/spaces/NotASI/FLUX.1-dev",
"FLUX.1-dev_hysts": "https://huggingface.co/spaces/hysts-mcp/FLUX.1-dev",
"HiDream-I1-Dev": "https://huggingface.co/spaces/HiDream-ai/HiDream-I1-Dev",
"UnfilteredAI-NSFW-gen-v2": "https://huggingface.co/spaces/armen425221356/UnfilteredAI-NSFW-gen-v2_self_parms",
"InfiniteYou-FLUX": "https://huggingface.co/spaces/ByteDance/InfiniteYou-FLUX",
"Stable Diffusion 3.5 Large (arad1367)": "https://huggingface.co/spaces/arad1367/Stable_Diffusion_3_5_Large_Customized",
"Stable Diffusion 3.5 Large Turbo (doevent)": "https://huggingface.co/spaces/doevent/stable-diffusion-3.5-large-turbo",
# Virtual Try-On & Character
"OutfitAnyone": "https://huggingface.co/spaces/HumanAIGC/OutfitAnyone",
"Kolors Virtual Try-On": "https://huggingface.co/spaces/Kwai-Kolors/Kolors-Virtual-Try-On",
"Miragic Virtual Try-On": "https://huggingface.co/spaces/Miragic-AI/Miragic-Virtual-Try-On",
"OutfitAnyway": "https://huggingface.co/spaces/selfit-camera/OutfitAnyway",
"IDM-VTON": "https://huggingface.co/spaces/yisol/IDM-VTON",
"InstantCharacter": "https://huggingface.co/spaces/InstantX/InstantCharacter",
"InstantID": "https://huggingface.co/spaces/InstantX/InstantID",
# AI Lip-Sync & Talking Avatars
"LivePortrait": "https://huggingface.co/spaces/Han-123/LivePortrait",
"LivePortrait (CPU)": "https://huggingface.co/spaces/K00B404/LivePortrait_cpu",
"D-ID Live Portrait AI": "https://www.d-id.com/liveportrait-4/",
"Synthesia Avatars": "https://www.synthesia.io/features/avatars",
"Papercup": "https://www.papercup.com/",
"Hedra": "https://www.hedra.com",
"LemonSlice": "https://lemonslice.com",
"Vozo AI": "https://www.vozo.ai/lip-sync",
"Gooey AI Lipsync": "https://gooey.ai/Lipsync",
"Sync.so": "https://sync.so",
"LipDub AI": "https://www.lipdub.ai",
"Magic Hour": "https://magichour.ai",
"Lifelike AI": "https://www.lifelikeai.io",
"DeepMotion": "https://www.deepmotion.com",
"Elai.io": "https://elai.io",
"Rephrase.ai": "https://www.rephrase.ai",
"Colossyan": "https://www.colossyan.com",
"HeyGen (Movio)": "https://www.heygen.com",
"Murf Studio": "https://murf.ai",
# Image Editing & Upscaling
"FLUX Fill/Outpaint": "https://huggingface.co/spaces/multimodalart/flux-fill-outpaint",
"ReSize Image Outpainting": "https://huggingface.co/spaces/VIDraft/ReSize-Image-Outpainting",
"IC-Light (Relighting)": "https://huggingface.co/spaces/lllyasviel/IC-Light",
"IC-Light v2-vary": "https://huggingface.co/spaces/lllyasviel/iclight-v2-vary",
"Kontext Relight": "https://huggingface.co/spaces/kontext-community/kontext-relight",
"SUPIR Upscaler": "https://huggingface.co/spaces/Fabrice-TIERCELIN/SUPIR",
# Video Generation & FramePacks
"Framepacks (atunc29)": "https://huggingface.co/spaces/atunc29/Framepacks",
"Framepack i2v (ginigen)": "https://huggingface.co/spaces/ginigen/framepack-i2v",
"Framepack i2v (beowcow)": "https://huggingface.co/spaces/beowcow/framepack-i2v",
"Framepack i2v (lisonallen)": "https://huggingface.co/spaces/lisonallen/framepack-i2v",
"FramePack F1 (Latyrine)": "https://huggingface.co/spaces/Latyrine/FramePack-F1",
"FramePack F1 (linoyts)": "https://huggingface.co/spaces/linoyts/FramePack-F1",
"FramePack Rotate (tori29umai)": "https://huggingface.co/spaces/tori29umai/FramePack_rotate_landscape",
"FramePack Rotate (bep40)": "https://huggingface.co/spaces/bep40/FramePack_rotate_landscape",
"FramePack Rotate (VIDraft)": "https://huggingface.co/spaces/VIDraft/FramePack_rotate_landscape",
"Framepack-H111 (rahul7star)": "https://huggingface.co/spaces/rahul7star/Framepack-H111",
"FLUX.1 Kontext Dev": "https://huggingface.co/spaces/black-forest-labs/FLUX.1-Kontext-Dev",
"Wan2-1-fast": "https://huggingface.co/spaces/multimodalart/wan2-1-fast",
"LTX-video-distilled": "https://huggingface.co/spaces/Lightricks/ltx-video-distilled",
"RunwayML": "https://app.runwayml.com/video-tools/teams/rinaabdine1/ai-tools/generate",
"Pika Labs": "https://pika.art/",
"Kling AI": "https://app.klingai.com/global/image-to-video/frame-mode",
# Video Interpolation & Slow Motion
"RIFE (remzloev)": "https://huggingface.co/spaces/remzloev/Rife",
"VFI Converter (Agung1453)": "https://huggingface.co/spaces/Agung1453/Video-Frame-Interpolation-Converter",
"ZeroGPU Upscaler/Interpolation": "https://huggingface.co/spaces/inoculatemedia/zerogpu-upscaler-interpolation",
"Frame Interpolation (meta-artem)": "https://huggingface.co/spaces/meta-artem/frame-interpolation",
"Video Frame Interpolation (guardiancc)": "https://huggingface.co/spaces/guardiancc/video_frame_interpolation",
"Video Frame Interpolation (freealise)": "https://huggingface.co/spaces/freealise/video_frame_interpolation",
"Framer (wwen1997)": "https://huggingface.co/spaces/wwen1997/Framer",
"Inter4k VideoInterpolator": "https://huggingface.co/spaces/vimleshc57/Inter4k_VideoInterpolator",
# AnimateDiff & Advanced Animation
"AnimateDiff Lightning (ByteDance)": "https://huggingface.co/spaces/ByteDance/AnimateDiff-Lightning",
"AnimateDiff Lightning (SahaniJi)": "https://huggingface.co/spaces/SahaniJi/AnimateDiff-Lightning",
"AnimateDiff (fatima14)": "https://huggingface.co/spaces/fatima14/AnimateDiff",
"AnimateDiff Video Gen (faizanR)": "https://huggingface.co/spaces/faizanR/animatediff-video-generator",
"Text-to-Animation Fast (MisterProton)": "https://huggingface.co/spaces/MisterProton/text-to-Animation-Fast-AnimateDiff",
"Text-to-Animation Fast (Rowdy013)": "https://huggingface.co/spaces/Rowdy013/text-to-Animation-Fast",
# StyleGAN & Portrait Motion
"StyleGAN-Human Interpolation (hysts)": "https://huggingface.co/spaces/hysts/StyleGAN-Human-Interpolation",
"StyleGAN-Human (Gradio-Blocks)": "https://huggingface.co/spaces/Gradio-Blocks/StyleGAN-Human",
# Film & Style Models
"MGM-Film-Diffusion (tonyassi)": "https://huggingface.co/spaces/tonyassi/MGM-Film-Diffusion",
"CineDiffusion (takarajordan)": "https://huggingface.co/spaces/takarajordan/CineDiffusion",
"FLUX Film Foto (MartsoBodziu1994)": "https://huggingface.co/spaces/MartsoBodziu1994/alvdansen-flux_film_foto",
"FLUX Style Shaping": "https://huggingface.co/spaces/multimodalart/flux-style-shaping",
"Film (Stijnijzelenberg)": "https://huggingface.co/spaces/Stijnijzelenberg/film",
"Film Eras (abbiewoodbridge)": "https://huggingface.co/spaces/abbiewoodbridge/Film_Eras",
"Film Genre Classifier (Rezuwan)": "https://huggingface.co/spaces/Rezuwan/film_genre_classifier",
"RunwayML (Faizbulbul)": "https://huggingface.co/spaces/Faizbulbul/Runwaymlfaiz",
# Text-to-3D
"Step1X-3D": "https://huggingface.co/spaces/stepfun-ai/Step1X-3D",
"TRELLIS TextTo3D (PUM4CH3N)": "https://huggingface.co/spaces/PUM4CH3N/TRELLIS_TextTo3D",
"TRELLIS TextTo3D (cavargas10)": "https://huggingface.co/spaces/cavargas10/TRELLIS-Texto3D",
"TRELLIS TextTo3D (dkatz2391)": "https://huggingface.co/spaces/dkatz2391/TRELLIS_TextTo3D_Try2",
"Sparc3D": "https://huggingface.co/spaces/ilcve21/Sparc3D",
"Hunyuan3D-2.1": "https://huggingface.co/spaces/tencent/Hunyuan3D-2.1",
# Image Captioning & Interrogation
"BLIP-2 (hysts)": "https://huggingface.co/spaces/hysts/BLIP2",
"BLIP-3o": "https://huggingface.co/spaces/BLIP3o/blip-3o",
"Blip-Dalle3 (DarwinAnim8or)": "https://huggingface.co/spaces/DarwinAnim8or/Blip-Dalle3",
"BLIP API (Jonu1)": "https://huggingface.co/spaces/Jonu1/blip-image-captioning-api",
"BLIP API (muxiddin19)": "https://huggingface.co/spaces/muxiddin19/blip-image-captioning-api",
# Diffusion & Sketching Tools
"DiffSketcher (SVGRender)": "https://huggingface.co/spaces/SVGRender/DiffSketcher",
"Diffusion WikiArt (kaupane)": "https://huggingface.co/spaces/kaupane/diffusion-wikiart",
"Diffusers Image Fill (OzzyGT)": "https://huggingface.co/spaces/OzzyGT/diffusers-image-fill",
"Diffusers Fast Inpaint (OzzyGT)": "https://huggingface.co/spaces/OzzyGT/diffusers-fast-inpaint",
# Audio & Voice Tools
"ThinkSound (FunAudioLLM)": "https://huggingface.co/spaces/FunAudioLLM/ThinkSound",
"TTS Unlimited (NihalGazi)": "https://huggingface.co/spaces/NihalGazi/Text-To-Speech-Unlimited",
"Voice Clon (tonyassi)": "https://huggingface.co/spaces/tonyassi/voice-clon",
# Scripting & Writing Tools
"SKRIPTZ (skylinkd)": "https://huggingface.co/spaces/skylinkd/SKRIPTZ",
# AI Frameworks & Platforms
"Hugging Face Hub": "https://huggingface.co",
"Hugging Face Transformers": "https://huggingface.co/docs/transformers/en/index",
"Hugging Face Inference API": "https://huggingface.co/inference-api/",
# Miscellaneous Video Tools
"SpatialTrackerV2 (Yuxihenry)": "https://huggingface.co/spaces/Yuxihenry/SpatialTrackerV2",
"MTVCraft (BAAI)": "https://huggingface.co/spaces/BAAI/MTVCraft",
# Miscellaneous Tools
"EBSynth (NihalGazi)": "https://huggingface.co/spaces/NihalGazi/EBSynth",
"MoodSpace (huzey)": "https://huggingface.co/spaces/huzey/MoodSpace",
"TR0N (Layer6)": "https://huggingface.co/spaces/Layer6/TR0N",
"TUTOR (nathannarrik)": "https://huggingface.co/spaces/nathannarrik/TUTOR",
"Sport Model 1 (CHEN11102)": "https://huggingface.co/spaces/CHEN11102/sportmodel1",
"VBench Leaderboard (Vchitect)": "https://huggingface.co/spaces/Vchitect/VBench_Leaderboard",
}
# --- Model Loading ---
DETECTOR_CONFIG = {
"Canny": {"class": CannyDetector, "args": {}},
"Lineart": {"class": LineartDetector, "args": {"pretrained_model_or_path": "lllyasviel/Annotators"}},
"MLSD": {"class": MLSDdetector, "args": {"pretrained_model_or_path": "lllyasviel/Annotators"}},
"OpenPose": {"class": OpenposeDetector, "args": {"pretrained_model_or_path": "lllyasviel/Annotators"}},
"NormalBAE": {"class": NormalBaeDetector, "args": {"pretrained_model_or_path": "lllyasviel/Annotators"}},
"SoftEdge (HED)": {"class": HEDdetector, "args": {"pretrained_model_or_path": "lllyasviel/Annotators"}},
}
def get_detector(name):
with model_load_lock:
if name not in loaded_detectors:
print(f"Loading {name} model...")
config = DETECTOR_CONFIG[name]
if "pretrained_model_or_path" in config["args"]:
detector_class = config["class"]
loaded_detectors[name] = detector_class.from_pretrained(**config["args"])
else:
loaded_detectors[name] = config["class"](**config["args"])
print(f"{name} model loaded.")
return loaded_detectors[name]
def load_whisper_model(model_name="base"):
global whisper_model, whisper_model_name
if whisper:
with model_load_lock:
if whisper_model is None or whisper_model_name != model_name:
print(f"Loading Whisper model '{model_name}'... (This may download files on first run)")
whisper_model = whisper.load_model(model_name, device="cpu")
whisper_model_name = model_name
print("Whisper model loaded.")
return whisper_model
return None
def load_enhance_ai_models():
"""Load RIFE model if it is not already loaded."""
global rife_model
if not ENHANCE_AI_AVAILABLE:
return
with model_load_lock:
if rife_model is None:
print("Loading RIFE model for frame interpolation...")
rife_model = Rife(gpuid=0, model="rife-v4.6", num_threads=4, tta_mode=False)
print("RIFE model loaded.")
get_detector("Canny") # Pre-load Canny detector
# --- Utility Functions ---
def parse_color(color_str):
"""
Parses a color string from Gradio's ColorPicker.
It can handle hex strings ('#RRGGBB') or the problematic
rgba float format ('rgba(r,g,b,a)').
Returns a tuple (r, g, b) for PIL.
"""
if not isinstance(color_str, str):
return color_str # Should already be a tuple or other valid format
if color_str.startswith('rgba'):
parts = re.findall(r"[\d\.]+", color_str)
if len(parts) >= 3:
return (int(float(parts[0])), int(float(parts[1])), int(float(parts[2])))
# Handle standard hex '#RRGGBB'
if color_str.startswith('#'):
hex_color = color_str.lstrip('#')
return tuple(int(hex_color[i:i+2], 16) for i in (0, 2, 4))
return color_str
def rotate_image(image, rotation):
if rotation == "90 Degrees Clockwise":
return cv2.rotate(image, cv2.ROTATE_90_CLOCKWISE)
elif rotation == "90 Degrees Counter-Clockwise":
return cv2.rotate(image, cv2.ROTATE_90_COUNTERCLOCKWISE)
elif rotation == "180 Degrees":
return cv2.rotate(image, cv2.ROTATE_180)
return image
def manipulate_image(image, operation):
if image is None:
raise gr.Error("Please upload an image first.")
if operation == "Invert Colors":
return cv2.bitwise_not(image)
elif operation == "Flip Horizontal":
return cv2.flip(image, 1)
elif operation == "Flip Vertical":
return cv2.flip(image, 0)
elif operation == "Rotate 90° Right":
return cv2.rotate(image, cv2.ROTATE_90_CLOCKWISE)
elif operation == "Rotate 90° Left":
return cv2.rotate(image, cv2.ROTATE_90_COUNTERCLOCKWISE)
else:
return image
def manipulate_video(video_path, operation):
if not video_path:
raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"manipulated_video_{timestamp}.mp4")
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise gr.Error("Error opening video file.")
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = cap.get(cv2.CAP_PROP_FPS)
if fps == 0: fps = 30
frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
out_width, out_height = width, height
if operation in ["Rotate 90° Right", "Rotate 90° Left"]:
out_width, out_height = height, width
writer = cv2.VideoWriter(output_video_path, fourcc, fps, (out_width, out_height))
for _ in range(frame_count):
ret, frame = cap.read()
if not ret:
break
processed_frame = manipulate_image(frame, operation)
writer.write(processed_frame)
cap.release()
writer.release()
return output_video_path
def get_media_duration(media_path):
if not media_path or not os.path.exists(media_path):
return 0.0
# --- METHOD 1: Fast Metadata Probe (for well-formed files) ---
try:
cmd = ["ffprobe", "-v", "error", "-show_entries", "format=duration", "-of", "default=noprint_wrappers=1:nokey=1", media_path]
result = subprocess.run(cmd, capture_output=True, text=True, check=True, timeout=10)
return float(result.stdout.strip())
except Exception:
# This method failed, likely due to malformed metadata. Proceed to the robust method.
pass
# --- METHOD 2: Robust Full Scan (for problematic files) ---
print(f"Warning: Fast duration check failed for {os.path.basename(media_path)}. Performing robust scan (this may take a moment)...")
try:
cmd = ["ffmpeg", "-i", media_path, "-f", "null", "-"]
# We need to capture stderr, where ffmpeg writes its progress
result = subprocess.run(cmd, capture_output=True, text=True, timeout=60)
# Search for the final 'time=' stamp in ffmpeg's output
matches = re.findall(r"time=(\d{2}):(\d{2}):(\d{2})\.(\d{2})", result.stderr)
if matches:
last_match = matches[-1]
hours, minutes, seconds, hundredths = map(int, last_match)
total_seconds = (hours * 3600) + (minutes * 60) + seconds + (hundredths / 100.0)
print(f"Robust scan successful. Detected duration: {total_seconds:.2f}s")
return total_seconds
else:
# If even this fails, the file is likely very corrupt
print(f"Error: Robust duration scan also failed for {os.path.basename(media_path)}.")
return 0.0
except Exception as e:
print(f"An unexpected error occurred during robust scan for {media_path}: {e}")
return 0.0
def get_video_dimensions(video_path):
if not video_path: return 0, 0
try:
cap = cv2.VideoCapture(video_path)
if not cap.isOpened(): return 0, 0
width, height = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
cap.release()
return width, height
except Exception: return 0, 0
def get_video_fps(video_path):
if not video_path: return 24.0
try:
cap = cv2.VideoCapture(video_path)
if not cap.isOpened(): return 24.0
fps = cap.get(cv2.CAP_PROP_FPS)
cap.release()
return fps if fps > 0 else 24.0
except Exception: return 24.0
def has_audio_stream(video_path):
"""Checks if a video file has at least one audio stream."""
if not video_path:
return False
try:
cmd = [
"ffprobe", "-v", "error", "-select_streams", "a",
"-show_entries", "stream=codec_type", "-of",
"default=noprint_wrappers=1:nokey=1", video_path
]
result = subprocess.run(cmd, capture_output=True, text=True, check=True)
return result.stdout.strip() != ""
except (subprocess.CalledProcessError, FileNotFoundError):
return False
def run_ffmpeg_command(cmd, desc="Processing with FFMPEG..."):
try:
print(f"Running FFMPEG command: {' '.join(cmd)}")
process = subprocess.run(
cmd,
capture_output=True,
text=True,
encoding='utf-8',
check=False
)
if process.returncode != 0:
full_output = f"--- FFMPEG & GRADIO ERROR LOG ---\n\n" \
f"FFMPEG COMMAND:\n{' '.join(cmd)}\n\n" \
f"FFMPEG STDERR:\n{process.stderr}\n\n" \
f"FFMPEG STDOUT:\n{process.stdout}"
raise subprocess.CalledProcessError(process.returncode, cmd, output=full_output)
except subprocess.CalledProcessError as e:
raise gr.Error(f"FFMPEG failed!\n\nDetails:\n{e.output}")
except FileNotFoundError:
raise gr.Error("FFMPEG not found. Please ensure ffmpeg is installed and in your system's PATH.")
def batch_image_processor(files, processing_function, job_name, **kwargs):
if not files: raise gr.Error("Please upload at least one image.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"{job_name}_{timestamp}"); os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for file_obj in files:
try:
base, _ = os.path.splitext(os.path.basename(file_obj.name))
if job_name == "zoom_videos":
output_filename = f"{base}.mp4"
elif job_name == "bg_removed":
output_filename = f"{base}.png"
elif job_name == "cropped":
output_filename = f"{base}_cropped.png"
else:
output_filename = os.path.basename(file_obj.name)
output_path = os.path.join(job_temp_dir, output_filename)
processing_function(input_path=file_obj.name, output_path=output_path, **kwargs)
output_paths.append(output_path)
except Exception as e:
print(f"Skipping file {file_obj.name} due to error: {e}")
continue
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No images could be processed from the batch.")
zip_base_name = os.path.join(TEMP_DIR, f"{job_name}_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths, zip_path, job_temp_dir
def process_batch_images_with_detector(files, detector_name):
detector = get_detector(detector_name)
def apply_detector(input_path, output_path, **kwargs):
with Image.open(input_path).convert("RGB") as img:
processed = detector(img, detect_resolution=512, image_resolution=1024)
processed.save(output_path)
output_paths, zip_path, _ = batch_image_processor(files, apply_detector, f"controlnet_{detector_name}")
return output_paths, zip_path
def process_video_with_detector(video_path, detector_name):
if not video_path: raise gr.Error("Please upload a video first.")
detector = get_detector(detector_name)
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"job_{timestamp}")
input_frames_dir, output_frames_dir = os.path.join(job_temp_dir, "input_frames"), os.path.join(job_temp_dir, "output_frames")
os.makedirs(input_frames_dir, exist_ok=True); os.makedirs(output_frames_dir, exist_ok=True)
output_video_path = os.path.join(TEMP_DIR, f"{detector_name.lower()}_output_{timestamp}.mp4")
cap = cv2.VideoCapture(video_path)
frame_count, frame_rate = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)), get_video_fps(video_path)
for i in range(frame_count):
success, frame = cap.read()
if not success: break
cv2.imwrite(os.path.join(input_frames_dir, f"frame_{i:05d}.png"), frame)
cap.release()
input_files = sorted(os.listdir(input_frames_dir))
for filename in input_files:
with Image.open(os.path.join(input_frames_dir, filename)).convert("RGB") as image:
result_pil = detector(image, detect_resolution=512, image_resolution=1024)
result_np = cv2.cvtColor(np.array(result_pil), cv2.COLOR_RGB2BGR)
cv2.imwrite(os.path.join(output_frames_dir, filename), result_np)
cmd = ["ffmpeg", "-framerate", str(frame_rate), "-i", os.path.join(output_frames_dir, "frame_%05d.png"), "-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path]
run_ffmpeg_command(cmd, "Compiling Video")
shutil.rmtree(job_temp_dir)
return output_video_path
def extract_first_last_frame(video_path):
if not video_path:
raise gr.Error("Please upload a video first.")
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
raise gr.Error("Failed to open video file.")
frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
if frame_count < 1:
cap.release()
raise gr.Error("Video has no frames.")
# Set position to the first frame and read it
cap.set(cv2.CAP_PROP_POS_FRAMES, 0)
success, first_frame_img = cap.read()
if not success:
cap.release()
raise gr.Error("Could not read the first frame.")
# --- FIX for Last Frame (Robust Method) ---
# Direct seeking to frame_count - 1 can be unreliable.
# This method seeks near the end and then reads sequentially to find the true last frame.
last_frame_img = None
# Start checking from a few frames before the reported end to be safe.
start_frame_for_last = max(1, frame_count - 10)
cap.set(cv2.CAP_PROP_POS_FRAMES, start_frame_for_last)
# Loop through the last few frames to ensure we get the very last one
while True:
success, frame = cap.read()
if not success:
break
last_frame_img = frame
cap.release()
# If the loop fails (e.g., for very short videos), fall back to using the first frame as the last.
if last_frame_img is None:
last_frame_img = first_frame_img
# --- FIX for saving with proper extension ---
# The function now saves the images to temporary files with correct names (.png) and returns the paths.
# Gradio's Gallery will display these files, and downloading them will use the correct filename.
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
first_frame_path = os.path.join(TEMP_DIR, f"first_frame_{timestamp}.png")
last_frame_path = os.path.join(TEMP_DIR, f"last_frame_{timestamp}.png")
# Convert from OpenCV's BGR format to RGB before saving with the PIL library
Image.fromarray(cv2.cvtColor(first_frame_img, cv2.COLOR_BGR2RGB)).save(first_frame_path)
Image.fromarray(cv2.cvtColor(last_frame_img, cv2.COLOR_BGR2RGB)).save(last_frame_path)
# Return the list of file paths to be displayed in the gallery
return [first_frame_path, last_frame_path]
# ### --- NEW FEATURE FUNCTION --- ###
def batch_extract_first_last_frames(videos, progress=gr.Progress(track_tqdm=True)):
if not videos:
raise gr.Error("Please upload at least one video.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"batch_fl_frames_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for video_file in progress.tqdm(videos, desc="Processing videos"):
try:
video_path = video_file.name
base_name = os.path.splitext(os.path.basename(video_path))[0]
cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
gr.Warning(f"Skipping '{base_name}': could not open video file.")
continue
frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
if frame_count < 1:
cap.release()
gr.Warning(f"Skipping '{base_name}': video has no frames.")
continue
# First frame
cap.set(cv2.CAP_PROP_POS_FRAMES, 0)
success, first_frame_img = cap.read()
if not success:
cap.release()
gr.Warning(f"Skipping '{base_name}': could not read the first frame.")
continue
# Last frame (robust method)
last_frame_img = None
start_frame_for_last = max(1, frame_count - 10)
cap.set(cv2.CAP_PROP_POS_FRAMES, start_frame_for_last)
while True:
success, frame = cap.read()
if not success:
break
last_frame_img = frame
cap.release()
if last_frame_img is None:
last_frame_img = first_frame_img
# Save frames
first_frame_path = os.path.join(job_temp_dir, f"{base_name}_first.png")
last_frame_path = os.path.join(job_temp_dir, f"{base_name}_last.png")
Image.fromarray(cv2.cvtColor(first_frame_img, cv2.COLOR_BGR2RGB)).save(first_frame_path)
Image.fromarray(cv2.cvtColor(last_frame_img, cv2.COLOR_BGR2RGB)).save(last_frame_path)
output_paths.extend([first_frame_path, last_frame_path])
except Exception as e:
gr.Warning(f"Skipping file {os.path.basename(video_file.name)} due to an error: {e}")
if 'cap' in locals() and cap.isOpened():
cap.release()
continue
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No frames could be extracted from the batch.")
zip_base_name = os.path.join(TEMP_DIR, f"batch_fl_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths, zip_path
def video_to_frames_extractor(video_path, skip_rate, rotation, do_resize, out_w, out_h, out_format, jpg_quality):
if not video_path: raise gr.Error("Please upload a video first.")
if do_resize and (out_w <= 0 or out_h <= 0): raise gr.Error("If resizing, width and height must be positive.")
cap = cv2.VideoCapture(video_path)
if not cap.isOpened(): raise gr.Error("Failed to open video file.")
frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
if frame_count < 1: cap.release(); raise gr.Error("Video appears to have no frames.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"v2f_{timestamp}"); os.makedirs(job_temp_dir, exist_ok=True)
frame_paths = []
saved_count = 0
for i in range(frame_count):
success, frame = cap.read()
if not success: break
if i % skip_rate != 0: continue
frame = rotate_image(frame, rotation)
if do_resize: frame = cv2.resize(frame, (out_w, out_h), interpolation=cv2.INTER_LANCZOS4)
frame_pil = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
file_ext = out_format.lower()
frame_path = os.path.join(job_temp_dir, f"frame_{saved_count:05d}.{file_ext}")
if out_format == "JPG": frame_pil.save(frame_path, quality=jpg_quality)
else: frame_pil.save(frame_path)
frame_paths.append(frame_path)
saved_count += 1
cap.release()
if not frame_paths: shutil.rmtree(job_temp_dir); raise gr.Error("Could not extract any frames.")
zip_base_name = os.path.join(TEMP_DIR, f"frames_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return frame_paths[:100], zip_path
def create_video_from_frames(files, fps, rotation, do_resize, out_w, out_h):
if not files: raise gr.Error("Please upload frame images first.")
if do_resize and (out_w <= 0 or out_h <= 0): raise gr.Error("If resizing, width and height must be positive.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"f2v_{timestamp}"); os.makedirs(job_temp_dir, exist_ok=True)
filenames = []
for i, file in enumerate(files):
ext = os.path.splitext(file.name)[1]
temp_path = os.path.join(job_temp_dir, f"frame_{i:05d}{ext}")
shutil.copy(file.name, temp_path); filenames.append(temp_path)
output_video_path = os.path.join(TEMP_DIR, f"video_from_frames_{timestamp}.mp4")
first_frame_img = rotate_image(cv2.imread(filenames[0]), rotation)
h, w, _ = first_frame_img.shape
if do_resize: w, h = out_w, out_h
w -= w % 2; h -= h % 2
temp_processed_dir = os.path.join(job_temp_dir, "processed"); os.makedirs(temp_processed_dir, exist_ok=True)
for i, filename in enumerate(filenames):
frame = rotate_image(cv2.imread(filename), rotation)
frame = cv2.resize(frame, (w, h), interpolation=cv2.INTER_LANCZOS4)
cv2.imwrite(os.path.join(temp_processed_dir, f"pframe_{i:05d}.png"), frame)
cmd = ["ffmpeg", "-framerate", str(fps), "-i", os.path.join(temp_processed_dir, "pframe_%05d.png"), "-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path]
run_ffmpeg_command(cmd, "Compiling Video")
shutil.rmtree(job_temp_dir)
return output_video_path
def image_to_looping_video(image_array, duration, audio_path=None):
if image_array is None: raise gr.Error("Please upload an image first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
temp_image_path = os.path.join(TEMP_DIR, f"temp_image_{timestamp}.png")
output_video_path = os.path.join(TEMP_DIR, f"looping_video_{timestamp}.mp4")
img = Image.fromarray(image_array)
img.save(temp_image_path)
width, height = img.size
width -= width % 2; height -= height % 2
cmd = ["ffmpeg", "-loop", "1", "-i", temp_image_path]
if audio_path:
cmd.extend(["-i", audio_path, "-c:a", "aac", "-shortest"])
cmd.extend(["-c:v", "libx264", "-t", str(duration), "-pix_fmt", "yuv420p", "-vf", f"scale={width}:{height}", "-y", output_video_path])
run_ffmpeg_command(cmd, "Creating Looping Video...")
os.remove(temp_image_path)
return output_video_path
def create_zoom_videos(files, duration, zoom_ratio, zoom_direction, combine_videos, audio_path=None):
if not files:
raise gr.Error("Please upload at least one image.")
fps = 30
total_frames = int(duration * fps)
zoom_step = (zoom_ratio - 1.0) / total_frames
zoom_coords = {
"Center": "x=iw/2-(iw/zoom)/2:y=ih/2-(ih/zoom)/2", "Top": "x=iw/2-(iw/zoom)/2:y=0", "Bottom": "x=iw/2-(iw/zoom)/2:y=ih-(ih/zoom)",
"Left": "x=0:y=ih/2-(ih/zoom)/2", "Right": "x=iw-(iw/zoom):y=ih/2-(ih/zoom)/2", "Top-Left": "x=0:y=0",
"Top-Right": "x=iw-(iw/zoom):y=0", "Bottom-Left": "x=0:y=ih-(ih/zoom)", "Bottom-Right": "x=iw-(iw/zoom):y=ih-(ih/zoom)",
}
def process_single_image(input_path, output_path, **kwargs):
audio_for_clip = kwargs.get('audio_for_clip')
zoom_filter = (f"scale=3840:-1,zoompan=z='min(zoom+{zoom_step},{zoom_ratio})':{zoom_coords[zoom_direction]}:d={total_frames}:s=1920x1080:fps={fps}")
cmd = ["ffmpeg", "-loop", "1", "-i", input_path]
if audio_for_clip:
cmd.extend(["-i", audio_for_clip, "-c:a", "aac", "-shortest"])
cmd.extend(["-vf", zoom_filter, "-c:v", "libx264", "-t", str(duration), "-pix_fmt", "yuv420p", "-b:v", "5M", "-y", output_path])
run_ffmpeg_command(cmd, f"Creating zoom video for {os.path.basename(input_path)}")
batch_kwargs = {}
if not combine_videos and audio_path:
batch_kwargs['audio_for_clip'] = audio_path
video_paths, zip_path, job_temp_dir = batch_image_processor(files, process_single_image, "zoom_videos", **batch_kwargs)
if not combine_videos:
return video_paths, None, zip_path
if not video_paths:
raise gr.Error("No videos were created to be combined.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
silent_combined_path = os.path.join(job_temp_dir, f"combined_silent_{timestamp}.mp4")
if len(video_paths) > 1:
file_list_path = os.path.join(job_temp_dir, "files.txt")
with open(file_list_path, 'w', encoding='utf-8') as f:
for path in video_paths:
f.write(f"file '{os.path.abspath(path)}'\n")
run_ffmpeg_command(["ffmpeg", "-f", "concat", "-safe", "0", "-i", file_list_path, "-c", "copy", "-y", silent_combined_path], "Combining Videos")
else:
shutil.copy(video_paths[0], silent_combined_path)
if audio_path:
final_video_path = os.path.join(TEMP_DIR, f"combined_audio_{timestamp}.mp4")
run_ffmpeg_command(["ffmpeg", "-i", silent_combined_path, "-i", audio_path, "-c:v", "copy", "-c:a", "aac", "-shortest", "-y", final_video_path], "Adding audio...")
else:
final_video_path = os.path.join(TEMP_DIR, f"combined_final_{timestamp}.mp4")
shutil.move(silent_combined_path, final_video_path)
return None, final_video_path, zip_path
def change_video_speed(video_path, speed_multiplier):
if not video_path: raise gr.Error("Please upload a video first.")
if speed_multiplier <= 0: raise gr.Error("Speed multiplier must be positive.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"speed_change_{timestamp}.mp4")
pts_value = 1 / speed_multiplier
cmd = ["ffmpeg", "-i", video_path, "-filter:v", f"setpts={pts_value}*PTS", "-an", "-y", output_video_path]
run_ffmpeg_command(cmd, "Changing Video Speed")
return output_video_path
def _get_atempo_filter_string(speed):
"""Helper function to create a chained atempo filter string for FFMPEG."""
filters = []
# 'atempo' is limited to [0.5, 100.0]
if speed > 100.0:
while speed > 100.0:
filters.append("atempo=100.0")
speed /= 100.0
elif speed < 0.5:
while speed < 0.5:
filters.append("atempo=0.5")
speed /= 0.5
# Add the final filter for the remaining speed adjustment
if speed != 1.0: # Avoid adding atempo=1.0 which does nothing
filters.append(f"atempo={speed}")
return ",".join(filters) if filters else None
def process_slowmo_enhance_video(video_path, output_path, slowdown_factor, method, progress):
"""
Processes a single video for slow-motion and enhancement.
"""
if not video_path:
raise gr.Error("Missing video path for processing.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"slowmo_{os.path.basename(video_path)}_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
original_fps = get_video_fps(video_path)
final_fps = original_fps * slowdown_factor
has_audio = has_audio_stream(video_path)
try:
if method == "AI-Enhanced (High Quality)":
input_frames_dir = os.path.join(job_temp_dir, "input_frames")
processed_frames_dir = os.path.join(job_temp_dir, "processed_frames")
os.makedirs(input_frames_dir, exist_ok=True)
os.makedirs(processed_frames_dir, exist_ok=True)
load_enhance_ai_models()
progress(0.1, desc="Extracting frames...")
run_ffmpeg_command(["ffmpeg", "-i", video_path, os.path.join(input_frames_dir, "frame_%06d.png")])
input_frames = sorted([os.path.join(input_frames_dir, f) for f in os.listdir(input_frames_dir)])
if not input_frames:
raise gr.Error("Could not extract any frames from the video.")
progress(0.3, desc="AI Interpolating frames (This can be slow)...")
for i in progress.tqdm(range(len(input_frames) - 1), unit="frame pairs"):
frame0 = cv2.imread(input_frames[i])
frame1 = cv2.imread(input_frames[i+1])
shutil.copy(input_frames[i], os.path.join(processed_frames_dir, f"proc_{i:06d}_0.png"))
interpolated_frames = rife_model.process(frame0, frame1, count=slowdown_factor-1)
for j, int_frame in enumerate(interpolated_frames):
cv2.imwrite(os.path.join(processed_frames_dir, f"proc_{i:06d}_{j+1}.png"), int_frame)
shutil.copy(input_frames[-1], os.path.join(processed_frames_dir, f"proc_{len(input_frames)-1:06d}_0.png"))
progress(0.8, desc="Compiling final video...")
silent_video_path = os.path.join(job_temp_dir, "silent_video.mp4")
cmd = ["ffmpeg", "-framerate", str(original_fps), "-pattern_type", "glob", "-i", os.path.join(processed_frames_dir, "*.png"), "-c:v", "libx264", "-crf", "18", "-pix_fmt", "yuv420p", "-y", silent_video_path]
run_ffmpeg_command(cmd)
if has_audio:
progress(0.9, desc="Attaching slowed audio...")
atempo_filter_str = _get_atempo_filter_string(1.0 / slowdown_factor)
cmd_audio = ["ffmpeg", "-i", silent_video_path, "-i", video_path, "-filter:a", atempo_filter_str, "-c:v", "copy", "-map", "0:v:0", "-map", "1:a:0", "-y", output_path]
run_ffmpeg_command(cmd_audio)
else:
shutil.move(silent_video_path, output_path)
elif method == "Standard (Fast)":
progress(0.5, desc="Processing with FFMPEG filter...")
cmd = ["ffmpeg", "-y", "-i", video_path]
vf_filter_chain = f"minterpolate=fps={final_fps}:mi_mode=mci,setpts={float(slowdown_factor)}*PTS"
cmd.extend(["-vf", vf_filter_chain])
if has_audio:
atempo_filter_str = _get_atempo_filter_string(1.0 / slowdown_factor)
if atempo_filter_str:
cmd.extend(["-af", atempo_filter_str])
else:
cmd.append("-an")
cmd.extend(["-r", str(original_fps)])
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", "-crf", "18", output_path])
run_ffmpeg_command(cmd)
finally:
if os.path.exists(job_temp_dir):
shutil.rmtree(job_temp_dir)
def batch_slowmo_enhance_videos(videos, slowdown_factor_str, method, progress=gr.Progress(track_tqdm=True)):
if not videos:
raise gr.Error("Please upload at least one video.")
slowdown_factor = int(slowdown_factor_str.replace('x', ''))
if "AI-Enhanced" in method and not ENHANCE_AI_AVAILABLE:
raise gr.Error("AI-Enhanced method is not available. Please install 'rife-ncnn-vulkan-python' and restart the app.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"slowmo_batch_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for i, video_file in enumerate(videos):
progress(i / len(videos), desc=f"Processing video {i+1}/{len(videos)}: {os.path.basename(video_file.name)}")
base, _ = os.path.splitext(os.path.basename(video_file.name))
output_path = os.path.join(job_temp_dir, f"{base}_slowmo_{slowdown_factor}x.mp4")
process_slowmo_enhance_video(video_file.name, output_path, slowdown_factor, method, progress)
output_paths.append(output_path)
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No videos could be processed from the batch.")
zip_base_name = os.path.join(TEMP_DIR, f"slowmo_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths, zip_path
def change_audio_speed(audio_path, speed_multiplier):
if not audio_path:
raise gr.Error("Please upload an audio file.")
if speed_multiplier <= 0:
raise gr.Error("Speed multiplier must be a positive number.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
# Get original extension
_, extension = os.path.splitext(os.path.basename(audio_path))
if not extension: extension = ".mp3" # Fallback
output_audio_path = os.path.join(TEMP_DIR, f"audio_speed_{speed_multiplier}x_{timestamp}{extension}")
atempo_filter_str = _get_atempo_filter_string(speed_multiplier)
if not atempo_filter_str:
# If no speed change, just copy the file to avoid processing
gr.Info("No speed change applied (multiplier is 1.0).")
shutil.copy(audio_path, output_audio_path)
return output_audio_path
cmd = ["ffmpeg", "-i", audio_path, "-filter:a", atempo_filter_str, "-y", output_audio_path]
run_ffmpeg_command(cmd, "Changing audio speed...")
return output_audio_path
# ### --- NEW FEATURE FUNCTION --- ###
def chop_audio_on_silence(audio_path, silence_thresh, min_silence_len, progress=gr.Progress(track_tqdm=True)):
if not audio_path:
raise gr.Error("Please upload an audio file to chop.")
progress(0, desc="Loading audio file...")
try:
sound = AudioSegment.from_file(audio_path)
except Exception as e:
raise gr.Error(f"Could not read audio file. It may be corrupt or in an unsupported format. Details: {e}")
progress(0.2, desc="Detecting non-silent chunks...")
audio_chunks = split_on_silence(
sound,
min_silence_len=int(min_silence_len),
silence_thresh=int(silence_thresh),
keep_silence=200 # Keep a bit of silence at the start/end of each chunk
)
if not audio_chunks:
raise gr.Error("No audio chunks were found above the silence threshold. Try using a lower (more negative) threshold value.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"audio_chop_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for i, chunk in enumerate(progress.tqdm(audio_chunks, desc="Exporting chunks...")):
output_path = os.path.join(job_temp_dir, f"chunk_{i:04d}.mp3")
chunk.export(output_path, format="mp3")
output_paths.append(output_path)
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("Failed to export any audio chunks.")
zip_base_name = os.path.join(TEMP_DIR, f"audio_chop_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
# Return a preview gallery and the zip file
return output_paths, zip_path
def reverse_video(video_path, audio_option):
if not video_path: raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"reversed_video_{timestamp}.mp4")
filters = ["reverse"]
if audio_option == "Reverse Audio": filters.append("areverse")
cmd = ["ffmpeg", "-i", video_path, "-vf", filters[0]]
if len(filters) > 1: cmd.extend(["-af", filters[1]])
if audio_option == "Remove Audio": cmd.append("-an")
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path])
run_ffmpeg_command(cmd, "Reversing video...")
return output_video_path
def add_audio_to_video(video_path, audio_path):
if not video_path: raise gr.Error("Please upload a video.")
if not audio_path: raise gr.Error("Please upload an audio file.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"video_with_audio_{timestamp}.mp4")
cmd = ["ffmpeg", "-i", video_path, "-i", audio_path, "-c:v", "copy", "-c:a", "aac", "-shortest", "-y", output_video_path]
run_ffmpeg_command(cmd, "Adding Audio to Video")
return output_video_path
def extract_audio(video_path, audio_format="mp3"):
if not video_path: raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_audio_path = os.path.join(TEMP_DIR, f"extracted_audio_{timestamp}.{audio_format}")
cmd = ["ffmpeg", "-i", video_path, "-vn"] # -vn strips video
if audio_format == "mp3": cmd.extend(["-c:a", "libmp3lame", "-q:a", "2"]) # VBR quality
elif audio_format == "aac": cmd.extend(["-c:a", "aac", "-b:a", "192k"])
elif audio_format == "wav": cmd.extend(["-c:a", "pcm_s16le"])
cmd.extend(["-y", output_audio_path])
run_ffmpeg_command(cmd, "Extracting audio...")
return output_audio_path
def create_gif_from_video(video_path, start_time, end_time):
if not video_path: raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_gif_path, palette_path = os.path.join(TEMP_DIR, f"video_to_gif_{timestamp}.gif"), os.path.join(TEMP_DIR, f"palette_{timestamp}.png")
duration_filter = []
if start_time > 0 or end_time > 0:
if end_time > 0 and end_time <= start_time: raise gr.Error("End time must be after start time.")
if start_time > 0: duration_filter.extend(["-ss", str(start_time)])
if end_time > 0: duration_filter.extend(["-to", str(end_time)])
run_ffmpeg_command(["ffmpeg", "-i", video_path] + duration_filter + ["-vf", "fps=15,scale=480:-1:flags=lanczos,palettegen", "-y", palette_path])
run_ffmpeg_command(["ffmpeg", "-i", video_path] + duration_filter + ["-i", palette_path, "-filter_complex", "fps=15,scale=480:-1:flags=lanczos[x];[x][1:v]paletteuse", "-y", output_gif_path])
os.remove(palette_path)
return output_gif_path
def get_frame_at_time(video_path, time_in_seconds=0):
if not video_path: return None
try:
command = ['ffmpeg', '-ss', str(time_in_seconds), '-i', video_path, '-vframes', '1', '-f', 'image2pipe', '-c:v', 'png', '-']
pipe = subprocess.run(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, check=True)
return Image.open(io.BytesIO(pipe.stdout)).convert("RGB")
except Exception as e:
print(f"Error extracting frame for crop preview: {e}")
cap = cv2.VideoCapture(video_path); cap.set(cv2.CAP_PROP_POS_MSEC, time_in_seconds * 1000)
success, frame = cap.read(); cap.release()
if success: return Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
return None
def crop_video(video_path, x, y, w, h, do_resize, out_w, out_h):
if not video_path: raise gr.Error("Please upload a video first.")
w, h, x, y = int(w), int(h), int(x), int(y)
w -= w % 2; h -= h % 2
if w <= 0 or h <= 0: raise gr.Error("Crop dimensions must be positive.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"cropped_video_{timestamp}.mp4")
vf_filters = [f"crop={w}:{h}:{x}:{y}"]
if do_resize:
if out_w <= 0 or out_h <= 0: raise gr.Error("Resize dimensions must be positive.")
out_w, out_h = int(out_w), int(out_h)
out_w -= out_w % 2; out_h -= out_h % 2
vf_filters.append(f"scale={out_w}:{out_h}")
cmd = ["ffmpeg", "-i", video_path, "-vf", ",".join(vf_filters), "-c:a", "copy", "-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path]
run_ffmpeg_command(cmd, "Cropping video...")
return output_video_path
def trim_video(video_path, start_time, end_time):
if not video_path: raise gr.Error("Please upload a video first.")
if start_time < 0: start_time = 0
if end_time <= start_time: end_time = 0
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"trimmed_video_{timestamp}.mp4")
cmd = ["ffmpeg", "-i", video_path, "-ss", str(start_time)]
if end_time > 0: cmd.extend(["-to", str(end_time)])
cmd.extend(["-c:v", "libx264", "-c:a", "copy", "-pix_fmt", "yuv420p", "-y", output_video_path])
run_ffmpeg_command(cmd, "Trimming Video")
return output_video_path
def apply_video_watermark(video_path, text, position, opacity, size_scale, color):
if not video_path: raise gr.Error("Please upload a video first.")
if not text: raise gr.Error("Watermark text cannot be empty.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"watermarked_video_{timestamp}.mp4")
_ , video_h = get_video_dimensions(video_path)
if video_h == 0:
video_h = 720 # Fallback
escaped_text = text.replace("'", r"'\''").replace(":", r"\:").replace(",", r"\,")
pos_map = {"Top-Left": "x=20:y=20", "Top-Right": "x=w-tw-20:y=20", "Bottom-Left": "x=20:y=h-th-20", "Bottom-Right": "x=w-tw-20:y=h-th-20", "Center": "x=(w-tw)/2:y=(h-th)/2"}
font_opacity = opacity / 100.0
font_size = int(video_h / (50 - (size_scale * 3.5)))
drawtext_filter = (
f"drawtext="
f"text='{escaped_text}':"
f"{pos_map[position]}:"
f"fontsize={font_size}:"
f"fontcolor={color}@{font_opacity}"
)
cmd = [
"ffmpeg", "-i", video_path,
"-vf", drawtext_filter,
"-c:a", "copy",
"-c:v", "libx264",
"-pix_fmt", "yuv420p",
"-y", output_video_path
]
run_ffmpeg_command(cmd, "Applying text watermark...")
return output_video_path
def remove_video_background(video_path):
if not video_path: raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"bg_rem_job_{timestamp}"); input_frames_dir, output_frames_dir = os.path.join(job_temp_dir, "input_frames"), os.path.join(job_temp_dir, "output_frames")
os.makedirs(input_frames_dir, exist_ok=True); os.makedirs(output_frames_dir, exist_ok=True)
cap = cv2.VideoCapture(video_path); frame_count, fps = int(cap.get(cv2.CAP_PROP_FRAME_COUNT)), get_video_fps(video_path)
for i in range(frame_count):
success, frame = cap.read()
if not success: break
cv2.imwrite(os.path.join(input_frames_dir, f"frame_{i:05d}.png"), frame)
cap.release()
for filename in sorted(os.listdir(input_frames_dir)):
with Image.open(os.path.join(input_frames_dir, filename)) as img:
remove(img).save(os.path.join(output_frames_dir, filename))
output_video_path = os.path.join(TEMP_DIR, f"bg_removed_{timestamp}.webm")
cmd = ["ffmpeg", "-framerate", str(fps), "-i", os.path.join(output_frames_dir, "frame_%05d.png"), "-c:v", "libvpx-vp9", "-pix_fmt", "yuva420p", "-auto-alt-ref", "0", "-b:v", "1M", "-y", output_video_path]
run_ffmpeg_command(cmd, "Compiling transparent video...")
shutil.rmtree(job_temp_dir)
return output_video_path
def generate_ass_from_whisper(result):
"""Generates an ASS subtitle file content from a Whisper result object with word timestamps."""
ass_content = [
"[Script Info]",
"Title: Generated by Skriptz",
"ScriptType: v4.00+",
"WrapStyle: 0",
"PlayResX: 1920",
"PlayResY: 1080",
"\n[V4+ Styles]",
"Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding",
"Style: Default,Arial,55,&H00FFFFFF,&H0000FFFF,&H00000000,&H00000000,0,0,0,0,100,100,0,0,1,2,1,2,10,10,25,1",
"\n[Events]",
"Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text"
]
def format_time(s):
h, r = divmod(s, 3600)
m, s = divmod(r, 60)
cs = int((s - int(s)) * 100)
return f"{int(h)}:{int(m):02}:{int(s):02}.{cs:02}"
for segment in result['segments']:
start_time = format_time(segment['start'])
end_time = format_time(segment['end'])
karaoke_line = ""
for word_info in segment['words']:
word = word_info['word'].strip()
duration_cs = int((word_info['end'] - word_info['start']) * 100)
karaoke_line += f"{{\\k{duration_cs}}}{word} "
dialogue_line = f"Dialogue: 0,{start_time},{end_time},Default,,0,0,0,,{karaoke_line.strip()}"
ass_content.append(dialogue_line)
return "\n".join(ass_content)
def transcribe_media(media_path, model_name):
if media_path is None: raise gr.Error("Please upload a video or audio file first.")
model = load_whisper_model(model_name)
if model is None: raise gr.Error("Whisper model is not available.")
audio_path = media_path.name
base_name = os.path.splitext(os.path.basename(media_path.name))[0]
if audio_path.lower().endswith(('.mp4', '.mov', '.mkv', '.avi', '.webm')):
audio_path_temp = os.path.join(TEMP_DIR, f"{base_name}.mp3")
try:
run_ffmpeg_command(["ffmpeg", "-y", "-i", audio_path, "-q:a", "0", "-map", "a", audio_path_temp])
audio_path = audio_path_temp
except gr.Error as e:
if "does not contain any stream" in str(e): raise gr.Error("The uploaded video has no audio track.")
else: raise e
result = model.transcribe(audio_path, word_timestamps=True, verbose=False)
def format_ts(s, separator=','):
h, r = divmod(s, 3600); m, s = divmod(r, 60)
return f"{int(h):02}:{int(m):02}:{int(s):02}{separator}{int((s-int(s))*1000):03}"
srt_path = os.path.join(TEMP_DIR, f"{base_name}.srt")
vtt_path = os.path.join(TEMP_DIR, f"{base_name}.vtt")
ass_path = os.path.join(TEMP_DIR, f"{base_name}.ass")
with open(srt_path, "w", encoding="utf-8") as srt_f, open(vtt_path, "w", encoding="utf-8") as vtt_f:
vtt_f.write("WEBVTT\n\n")
for i, seg in enumerate(result["segments"]):
start, end, text = seg['start'], seg['end'], seg['text'].strip()
srt_f.write(f"{i + 1}\n{format_ts(start)} --> {format_ts(end)}\n{text}\n\n")
vtt_f.write(f"{format_ts(start, '.')} --> {format_ts(end, '.')}\n{text}\n\n")
ass_content = generate_ass_from_whisper(result)
with open(ass_path, "w", encoding="utf-8") as ass_f:
ass_f.write(ass_content)
return result["text"], [srt_path, vtt_path, ass_path]
def transcribe_and_prep_burn(media_file, model_name):
if not media_file: raise gr.Error("Please upload a file first.")
is_video = media_file.name.lower().endswith(('.mp4', '.mov', '.mkv', '.avi', '.webm'))
text, files = transcribe_media(media_file, model_name)
if is_video: return text, files, media_file.name, gr.update(visible=True)
else: return text, files, None, gr.update(visible=False)
def reformat_srt_for_word_wrap(original_srt_path, words_per_line):
if not original_srt_path or not os.path.exists(original_srt_path):
return None
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
reformatted_path = os.path.join(TEMP_DIR, f"reformatted_{timestamp}.srt")
with open(original_srt_path, 'r', encoding='utf-8') as f_in, \
open(reformatted_path, 'w', encoding='utf-8') as f_out:
content = f_in.read().strip().split('\n\n')
for block in content:
lines = block.split('\n')
if len(lines) < 3:
f_out.write(block + '\n\n')
continue
text_lines = ' '.join(lines[2:])
words = text_lines.split()
new_text_lines = []
current_line = []
for word in words:
current_line.append(word)
if len(current_line) >= words_per_line:
new_text_lines.append(' '.join(current_line))
current_line = []
if current_line: new_text_lines.append(' '.join(current_line))
reformatted_text = '\n'.join(new_text_lines)
f_out.write(f"{lines[0]}\n{lines[1]}\n{reformatted_text}\n\n")
return reformatted_path
def burn_block_subtitles(video_path, srt_file_obj, font_size_scale, font_color, words_per_line):
original_srt_path = srt_file_obj[0].name
reformatted_srt_path = None
try:
reformatted_srt_path = reformat_srt_for_word_wrap(original_srt_path, words_per_line)
if not reformatted_srt_path: raise gr.Error("Failed to reformat subtitle file.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"subtitled_video_{timestamp}.mp4")
_, video_h = get_video_dimensions(video_path)
if video_h == 0: video_h = 720
divisor = 32 - (font_size_scale * 2)
calculated_font_size = int(video_h / divisor)
color_bgr = font_color[5:7] + font_color[3:5] + font_color[1:3]
ffmpeg_color = f"&H00{color_bgr.upper()}"
escaped_srt_path = reformatted_srt_path.replace('\\', '/').replace(':', r'\\:')
vf_filter = f"subtitles='{escaped_srt_path}':force_style='Fontsize={calculated_font_size},PrimaryColour={ffmpeg_color},BorderStyle=1,Outline=1,Shadow=0.5,MarginV=25'"
cmd = ["ffmpeg", "-y", "-i", video_path, "-vf", vf_filter, "-c:a", "copy", "-c:v", "libx264", "-pix_fmt", "yuv420p", output_video_path]
run_ffmpeg_command(cmd, "Burning block subtitles into video...")
return output_video_path
finally:
if reformatted_srt_path and os.path.exists(reformatted_srt_path):
os.remove(reformatted_srt_path)
def burn_karaoke_subtitles(video_path, subtitle_files, font_size_scale, base_color, highlight_color):
ass_file_path = subtitle_files[2].name
temp_ass_path = None
try:
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
temp_ass_path = os.path.join(TEMP_DIR, f"style_applied_{timestamp}.ass")
output_video_path = os.path.join(TEMP_DIR, f"karaoke_video_{timestamp}.mp4")
_, video_h = get_video_dimensions(video_path)
if video_h == 0: video_h = 720
calculated_font_size = int((video_h / 20) * (font_size_scale / 5))
def format_ass_color(hex_color):
if hex_color.startswith('#'): hex_color = hex_color[1:]
r, g, b = hex_color[0:2], hex_color[2:4], hex_color[4:6]
return f"&H00{b.upper()}{g.upper()}{r.upper()}"
primary_color_ass = format_ass_color(highlight_color)
secondary_color_ass = format_ass_color(base_color)
with open(ass_file_path, 'r', encoding='utf-8') as f_in, open(temp_ass_path, 'w', encoding='utf-8') as f_out:
for line in f_in:
if line.startswith("Style:"):
parts = line.split(',')
parts[2] = str(calculated_font_size) # Fontsize
parts[3] = secondary_color_ass # PrimaryColour (Base text)
parts[4] = primary_color_ass # SecondaryColour (Karaoke fill)
if len(parts) > 17:
parts[16] = '0' # Outline width
parts[17] = '0' # Shadow width
f_out.write(','.join(parts))
else:
f_out.write(line)
escaped_ass_path = temp_ass_path.replace('\\', '/').replace(':', r'\\:')
vf_filter = f"subtitles='{escaped_ass_path}'"
cmd = ["ffmpeg", "-y", "-i", video_path, "-vf", vf_filter, "-c:a", "copy", "-c:v", "libx264", "-pix_fmt", "yuv420p", output_video_path]
run_ffmpeg_command(cmd, "Burning karaoke subtitles into video...")
return output_video_path
finally:
if temp_ass_path and os.path.exists(temp_ass_path):
os.remove(temp_ass_path)
def burn_subtitles_wrapper(video_path, subtitle_files, style, font_size_scale, block_font_color, block_words_per_line, kara_base_color, kara_highlight_color):
if not video_path or not subtitle_files: raise gr.Error("Missing video or subtitle files. Please transcribe first.")
if style == "Block":
return burn_block_subtitles(video_path, subtitle_files, font_size_scale, block_font_color, block_words_per_line)
elif style == "Karaoke":
return burn_karaoke_subtitles(video_path, subtitle_files, font_size_scale, kara_base_color, kara_highlight_color)
else:
raise gr.Error("Invalid subtitle style selected.")
def remove_background_single(input_path, output_path, **kwargs):
with Image.open(input_path) as img:
remove(img).save(output_path)
def remove_background_batch(files):
output_paths, zip_path, _ = batch_image_processor(files, remove_background_single, "bg_removed")
return output_paths, zip_path
def resize_convert_single_image(input_path, output_path, **kwargs):
output_format = kwargs.get('output_format', 'JPG')
quality = kwargs.get('quality', 95)
enable_resize = kwargs.get('enable_resize', False)
max_w = kwargs.get('max_w', 1024)
max_h = kwargs.get('max_h', 1024)
resize_mode = kwargs.get('resize_mode', "Fit (preserve aspect ratio)")
with Image.open(input_path) as img:
if output_format in ['JPG', 'WEBP'] and img.mode in ['RGBA', 'P', 'LA']:
img = img.convert("RGB")
if enable_resize:
if resize_mode == "Fit (preserve aspect ratio)":
img.thumbnail((max_w, max_h), Image.Resampling.LANCZOS)
else: # Stretch
img = img.resize((max_w, max_h), Image.Resampling.LANCZOS)
save_kwargs = {}
pil_format = 'JPEG' if output_format == 'JPG' else output_format
if pil_format in ['JPEG', 'WEBP']:
save_kwargs['quality'] = quality
img.save(output_path, pil_format, **save_kwargs)
def batch_resize_convert_images(files, output_format, quality, enable_resize, max_w, max_h, resize_mode):
if not files: raise gr.Error("Please upload at least one image.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_name = "resized_converted"
job_temp_dir = os.path.join(TEMP_DIR, f"{job_name}_{timestamp}"); os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
# Enable resizing if dimensions are provided, even if checkbox is somehow out of sync
if max_w > 0 and max_h > 0:
enable_resize = True
processing_kwargs = {
'output_format': output_format, 'quality': quality, 'enable_resize': enable_resize,
'max_w': max_w, 'max_h': max_h, 'resize_mode': resize_mode
}
for file_obj in files:
try:
base, _ = os.path.splitext(os.path.basename(file_obj.name))
output_filename = f"{base}.{output_format.lower()}"
output_path = os.path.join(job_temp_dir, output_filename)
resize_convert_single_image(file_obj.name, output_path, **processing_kwargs)
output_paths.append(output_path)
except Exception as e: print(f"Skipping file {file_obj.name} due to error: {e}"); continue
if not output_paths: shutil.rmtree(job_temp_dir); raise gr.Error("No images could be processed.")
zip_base_name = os.path.join(TEMP_DIR, f"{job_name}_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths[:100], zip_path
def apply_watermark_single(input_path, output_path, watermark_text, position, opacity):
with Image.open(input_path).convert("RGBA") as image:
if not watermark_text: raise ValueError("Watermark text cannot be empty.")
txt = Image.new("RGBA", image.size, (255, 255, 255, 0))
try: font = ImageFont.truetype("DejaVuSans.ttf", int(image.width / 20))
except IOError: font = ImageFont.load_default()
d = ImageDraw.Draw(txt); bbox = d.textbbox((0, 0), watermark_text, font=font); w, h = bbox[2]-bbox[0], bbox[3]-bbox[1]
pos_map = {"Top-Left":(10,10), "Top-Right":(image.width-w-10,10), "Bottom-Left":(10,image.height-h-10), "Bottom-Right":(image.width-w-10,image.height-h-10), "Center":((image.width-w)/2,(image.height-h)/2)}
d.text(pos_map[position], watermark_text, font=font, fill=(255, 255, 255, int(255 * (opacity / 100))))
Image.alpha_composite(image, txt).convert("RGB").save(output_path)
def apply_watermark_batch(files, watermark_text, position, opacity):
if not watermark_text: raise gr.Error("Please provide watermark text.")
processing_func = lambda input_path, output_path: apply_watermark_single(
input_path, output_path, watermark_text=watermark_text, position=position, opacity=opacity
)
output_paths, zip_path, _ = batch_image_processor(files, processing_func, "watermarked")
return output_paths, zip_path
# --- BATCH CONVERTER REPLACEMENT FUNCTIONS ---
def convert_compress_video(video_path, out_format, v_codec, crf_value, scale_option, a_codec, a_bitrate, output_dir=None, base_name=None):
if not video_path:
raise gr.Error("Please upload a video to convert.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
_output_dir = output_dir if output_dir else TEMP_DIR
_base_name = base_name if base_name else f"converted_{timestamp}"
output_filename = f"{_base_name}.{out_format.lower()}"
output_path = os.path.join(_output_dir, output_filename)
cmd = ["ffmpeg", "-i", video_path]
vf_filters = []
if scale_option != "Original":
w, h = get_video_dimensions(video_path)
if w > 0 and h > 0:
target_h = int(scale_option.replace('p', ''))
target_w = round(w * target_h / h / 2) * 2
vf_filters.append(f"scale={target_w}:{target_h}")
vf_filters.append("pad=ceil(iw/2)*2:ceil(ih/2)*2")
vf_filters.append("setsar=1")
if vf_filters:
cmd.extend(["-vf", ",".join(vf_filters)])
cmd.extend(["-c:v", v_codec])
if v_codec in ["libx264", "libx265"]:
cmd.extend(["-crf", str(crf_value)])
cmd.extend(["-pix_fmt", "yuv420p"])
if has_audio_stream(video_path):
if a_codec == "copy":
cmd.extend(["-c:a", "copy"])
else:
cmd.extend(["-c:a", a_codec, "-b:a", f"{a_bitrate}k"])
else:
cmd.append("-an")
if out_format.lower() in ["mp4", "mov"]:
cmd.extend(["-movflags", "+faststart"])
cmd.extend(["-y", output_path])
run_ffmpeg_command(cmd, f"Converting {os.path.basename(video_path)}.")
return output_path
def batch_convert_compress_videos(files, out_format, v_codec, crf_value, scale_option, a_codec, a_bitrate, progress=gr.Progress(track_tqdm=True)):
if not files:
raise gr.Error("Please upload at least one video to convert.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"batch_convert_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for video_file in progress.tqdm(files, desc="Converting videos"):
try:
base_name = os.path.splitext(os.path.basename(video_file.name))[0]
output_path = convert_compress_video(
video_path=video_file.name,
out_format=out_format,
v_codec=v_codec,
crf_value=crf_value,
scale_option=scale_option,
a_codec=a_codec,
a_bitrate=a_bitrate,
output_dir=job_temp_dir,
base_name=base_name
)
output_paths.append(output_path)
except Exception as e:
gr.Warning(f"Skipping file {os.path.basename(video_file.name)} due to an error: {e}")
continue
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No videos could be processed from the batch.")
zip_base_name = os.path.join(TEMP_DIR, f"video_convert_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths, zip_path
def convert_audio(media_path, out_format, a_bitrate, output_dir=None, base_name=None):
if not media_path:
raise gr.Error("Please provide a media file to convert.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
_output_dir = output_dir if output_dir else TEMP_DIR
_base_name = base_name if base_name else f"audio_converted_{timestamp}"
output_filename = f"{_base_name}.{out_format.lower()}"
output_path = os.path.join(_output_dir, output_filename)
cmd = ["ffmpeg", "-i", media_path, "-vn"]
if out_format == "mp3":
cmd.extend(["-c:a", "libmp3lame", "-b:a", f"{a_bitrate}k"])
elif out_format == "aac":
cmd.extend(["-c:a", "aac", "-b:a", f"{a_bitrate}k"])
elif out_format == "ogg":
cmd.extend(["-c:a", "libopus", "-b:a", f"{a_bitrate}k"])
elif out_format == "wav":
cmd.extend(["-c:a", "pcm_s16le"])
elif out_format == "flac":
cmd.extend(["-c:a", "flac"])
cmd.extend(["-y", output_path])
run_ffmpeg_command(cmd, f"Converting audio from {os.path.basename(media_path)}...")
return output_path
def batch_convert_audio(files, out_format, a_bitrate, progress=gr.Progress(track_tqdm=True)):
if not files:
raise gr.Error("Please upload at least one file to convert.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"batch_audio_convert_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
for media_file in progress.tqdm(files, desc="Converting audio"):
try:
is_video = get_file_type(media_file.name) == 'video'
if is_video and not has_audio_stream(media_file.name):
gr.Warning(f"Skipping video '{os.path.basename(media_file.name)}' as it has no audio track.")
continue
base_name = os.path.splitext(os.path.basename(media_file.name))[0]
output_path = convert_audio(
media_path=media_file.name,
out_format=out_format,
a_bitrate=a_bitrate,
output_dir=job_temp_dir,
base_name=base_name
)
output_paths.append(output_path)
except Exception as e:
gr.Warning(f"Skipping file {os.path.basename(media_file.name)} due to an error: {e}")
continue
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No files could be processed from the batch.")
zip_base_name = os.path.join(TEMP_DIR, f"audio_convert_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths[0], zip_path
# --- END BATCH CONVERTER REPLACEMENT FUNCTIONS ---
def apply_video_fade(video_path, fade_in_duration, fade_out_duration):
if not video_path: raise gr.Error("Please upload a video.")
video_duration = get_media_duration(video_path)
if fade_in_duration + fade_out_duration > video_duration: raise gr.Error("The sum of fade durations cannot be greater than the video duration.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"faded_video_{timestamp}.mp4")
fade_filters = []
if fade_in_duration > 0: fade_filters.append(f"fade=t=in:st=0:d={fade_in_duration}")
if fade_out_duration > 0: fade_out_start = video_duration - fade_out_duration; fade_filters.append(f"fade=t=out:st={fade_out_start}:d={fade_out_duration}")
if not fade_filters: gr.Info("No fade applied."); return video_path
cmd = ["ffmpeg", "-i", video_path, "-vf", ",".join(fade_filters), "-c:a", "copy", "-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path]
run_ffmpeg_command(cmd, "Applying video fade...")
return output_video_path
# --- ACCURATE Color Grading Functions ---
def preview_color_grading_ffmpeg(image_np, brightness, contrast, saturation, sharpness):
"""Applies color grading to a single frame using FFMPEG for an accurate preview."""
if image_np is None:
return None
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
input_path = os.path.join(TEMP_DIR, f"cg_preview_in_{timestamp}.png")
output_path = os.path.join(TEMP_DIR, f"cg_preview_out_{timestamp}.png")
try:
Image.fromarray(image_np).save(input_path)
eq_filters, other_filters = [], []
if brightness != 0.0: eq_filters.append(f"brightness={brightness}")
if contrast != 1.0: eq_filters.append(f"contrast={contrast}")
if saturation != 1.0: eq_filters.append(f"saturation={saturation}")
if sharpness > 0.0: other_filters.append(f"unsharp=5:5:{sharpness}")
vf_parts = []
if eq_filters: vf_parts.append("eq=" + ":".join(eq_filters))
if other_filters: vf_parts.extend(other_filters)
if not vf_parts:
return Image.fromarray(image_np)
vf_string = ",".join(vf_parts)
cmd = ["ffmpeg", "-i", input_path, "-vf", vf_string, "-y", output_path]
subprocess.run(cmd, capture_output=True, text=True, check=False)
if os.path.exists(output_path):
with Image.open(output_path) as img:
return img.copy()
else:
return Image.fromarray(image_np)
except Exception as e:
print(f"Error in FFMPEG preview: {e}")
return Image.fromarray(image_np)
finally:
if os.path.exists(input_path): os.remove(input_path)
if os.path.exists(output_path): os.remove(output_path)
def apply_color_grading(video_path, brightness, contrast, saturation, sharpness):
"""Applies color grading to a full video using FFMPEG."""
if not video_path:
raise gr.Error("Please upload a video first.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"graded_video_{timestamp}.mp4")
eq_filters, other_filters = [], []
if brightness != 0.0: eq_filters.append(f"brightness={brightness}")
if contrast != 1.0: eq_filters.append(f"contrast={contrast}")
if saturation != 1.0: eq_filters.append(f"saturation={saturation}")
if sharpness > 0.0: other_filters.append(f"unsharp=5:5:{sharpness}")
vf_parts = []
if eq_filters: vf_parts.append("eq=" + ":".join(eq_filters))
if other_filters: vf_parts.extend(other_filters)
if not vf_parts:
gr.Info("No adjustments made. Returning original video path.")
return video_path
vf_string = ",".join(vf_parts)
cmd = ["ffmpeg", "-i", video_path, "-vf", vf_string, "-c:a", "copy", "-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", output_video_path]
run_ffmpeg_command(cmd, "Applying Color Grading...")
return output_video_path
# --- END ACCURATE Color Grading Functions ---
def trim_and_fade_audio(audio_path, start_time, end_time, fade_in_duration, fade_out_duration):
if not audio_path: raise gr.Error("Please upload an audio file.")
audio_duration = get_media_duration(audio_path)
if start_time < 0: start_time = 0
if end_time <= 0 or end_time > audio_duration: end_time = audio_duration
if start_time >= end_time: raise gr.Error("Start time must be less than end time.")
trimmed_duration = end_time - start_time
if fade_in_duration + fade_out_duration > trimmed_duration: raise gr.Error("Sum of fade durations cannot be greater than the trimmed audio duration.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_audio_path = os.path.join(TEMP_DIR, f"edited_audio_{timestamp}.mp3")
af_filters = []
if fade_in_duration > 0: af_filters.append(f"afade=t=in:st=0:d={fade_in_duration}")
if fade_out_duration > 0: fade_out_start = trimmed_duration - fade_out_duration; af_filters.append(f"afade=t=out:st={fade_out_start}:d={fade_out_duration}")
cmd = ["ffmpeg", "-ss", str(start_time), "-to", str(end_time), "-i", audio_path]
if af_filters: cmd.extend(["-af", ",".join(af_filters)])
cmd.extend(["-y", output_audio_path])
run_ffmpeg_command(cmd, "Trimming and fading audio...")
return output_audio_path
# In app.py, replace the existing create_gradual_ramp_video function with this one.
def create_gradual_ramp_video(video_path, progress=gr.Progress(track_tqdm=True)):
"""
Creates a video with a gradual speed ramp: 1x -> 0.5x -> 1x.
The effect is applied over the entire duration of the video.
This uses a piecewise approximation with frame interpolation for a smooth result.
--- ROBUSTNESS ENHANCEMENT ---
A hybrid approach is used:
1. For videos < 2.0s: A simplified, robust method applies an *average* speed change across
the whole clip. This avoids errors from creating many tiny, unstable segments.
2. For videos >= 2.0s: The original advanced segmentation logic is used to create a
more detailed and noticeable ramp effect.
"""
if not video_path:
raise gr.Error("Please upload a video to process.")
progress(0, desc="Analyzing video properties...")
duration = get_media_duration(video_path)
if duration == 0:
raise gr.Error("Could not determine video duration. The file may be corrupt.")
fps = get_video_fps(video_path)
has_audio = has_audio_stream(video_path)
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"gradual_ramp_{timestamp}.mp4")
# --- THIS IS THE FIX: A DEDICATED PATH FOR SHORT VIDEOS ---
# For very short videos, the complex ramp is barely noticeable and prone to ffmpeg errors.
# We switch to a simpler, more stable method by applying an average speed change.
if duration < 2.0:
gr.Info("Video is short (< 2s). Applying a simplified, robust ramp effect.")
progress(0.2, desc="Applying simplified ramp for short video...")
# The integral of the speed curve from 1->0.5->1 gives a total duration multiplier of 1.5.
# So, the average speed is original_duration / new_duration = 1 / 1.5 = 2/3.
avg_speed = 2.0 / 3.0
# Calculate the target interpolated FPS to create new frames for the slowdown.
interpolated_fps = fps / avg_speed
filter_complex_parts = []
# Video filter: interpolate to the new framerate, then adjust timestamps to slow it down.
video_filter = f"[0:v]minterpolate=fps={interpolated_fps}:mi_mode=mci,setpts=PTS/{avg_speed}[vout]"
filter_complex_parts.append(video_filter)
# Audio filter: apply the same speed change to the audio.
if has_audio:
atempo_str = _get_atempo_filter_string(avg_speed)
audio_filter = f"[0:a]asetpts=PTS"
if atempo_str:
audio_filter += f",{atempo_str}"
audio_filter += f"[aout]"
filter_complex_parts.append(audio_filter)
filter_complex_str = ";".join(filter_complex_parts)
# Build the simplified ffmpeg command
cmd = ["ffmpeg", "-y", "-i", video_path, "-filter_complex", filter_complex_str, "-map", "[vout]"]
if has_audio:
cmd.extend(["-map", "[aout]"])
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", "-crf", "18", output_video_path])
progress(0.6, desc="Executing simplified FFMPEG command...")
run_ffmpeg_command(cmd, "Applying simplified speed ramp...")
return output_video_path
# --- Standard logic for videos >= 2.0 seconds ---
progress(0.1, desc="Planning detailed speed ramp...")
# Determine the number of segments to approximate the curve.
# More segments = smoother, but more complex command. Capped at 60 for performance.
target_segment_duration = 0.25 # Aim for segments of this length
num_segments = int(duration / target_segment_duration)
if num_segments % 2 != 0:
num_segments += 1 # Ensure even number of segments for symmetry
num_segments = max(10, min(num_segments, 60)) # Clamp between 10 and 60
min_speed = 0.5
half_segments = num_segments / 2.0
filter_complex_parts = []
video_outputs, audio_outputs = [], []
# Loop through each segment and build the corresponding filter chain
for i in progress.tqdm(range(num_segments), desc="Building FFMPEG filter command..."):
start_time = i * duration / num_segments
end_time = (i + 1) * duration / num_segments
# Parabolic speed calculation (y = ax^2 + c) for smooth ease-in/out
x = (i - half_segments + 0.5) / half_segments
speed = (1.0 - min_speed) * (x ** 2) + min_speed
speed = max(0.01, speed) # Prevent speed from being zero
# Video processing for this segment
interpolated_fps_seg = fps / speed
setpts_val_seg = 1.0 / speed
video_filter = (
f"[0:v]trim=start={start_time}:end={end_time},setpts=PTS-STARTPTS," # Cut the segment
f"minterpolate=fps={interpolated_fps_seg}:mi_mode=mci," # Interpolate frames for smoothness
f"setpts={setpts_val_seg}*PTS[v{i}]" # Adjust speed
)
filter_complex_parts.append(video_filter)
video_outputs.append(f"[v{i}]")
# Audio processing for this segment
if has_audio:
atempo_str_seg = _get_atempo_filter_string(speed)
audio_filter = f"[0:a]atrim=start={start_time}:end={end_time},asetpts=PTS-STARTPTS"
if atempo_str_seg:
audio_filter += f",{atempo_str_seg}"
audio_filter += f"[a{i}]"
filter_complex_parts.append(audio_filter)
audio_outputs.append(f"[a{i}]")
progress(0.5, desc="Finalizing filter command...")
# Concatenate all processed video and audio segments
concat_filter_v = f"{''.join(video_outputs)}concat=n={num_segments}:v=1:a=0[vout]"
filter_complex_parts.append(concat_filter_v)
if has_audio and audio_outputs:
concat_filter_a = f"{''.join(audio_outputs)}concat=n={num_segments}:v=0:a=1[aout]"
filter_complex_parts.append(concat_filter_a)
filter_complex_str = ";".join(filter_complex_parts)
# Build the final complex ffmpeg command
cmd = ["ffmpeg", "-y", "-i", video_path, "-filter_complex", filter_complex_str, "-map", "[vout]"]
if has_audio and audio_outputs:
cmd.extend(["-map", "[aout]"])
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", "-crf", "18", output_video_path])
progress(0.6, desc="Executing FFMPEG... This may take a while.")
run_ffmpeg_command(cmd, "Applying gradual speed ramp...")
return output_video_path
# --- FLUX API ---
FLUX_MODELS = {"FLUX.1-schnell (Fast)": "black-forest-labs/FLUX.1-schnell", "FLUX.1-dev (High Quality)": "black-forest-labs/FLUX.1-dev"}
def call_flux_api(prompt, model_choice, width, height, hf_token):
if not hf_token: raise gr.Error("Hugging Face User Access Token is required.")
try:
client = Client(FLUX_MODELS[model_choice], hf_token=hf_token)
return client.predict(prompt=prompt, seed=0, randomize_seed=True, width=width, height=height, num_inference_steps=8 if "dev" in model_choice else 4, api_name="/infer")[0]
except Exception as e: raise gr.Error(f"API call failed: {e}")
def get_image_as_base64(path):
try:
with open(path, "rb") as f: return f"data:image/png;base64,{base64.b64encode(f.read()).decode('utf-8')}"
except FileNotFoundError: return None
# --- Transfer Tab Functions (Simplified) ---
def filter_presets(query, all_presets):
if not query:
return gr.update(choices=sorted(list(all_presets.keys())))
filtered_keys = [key for key in all_presets.keys() if query.lower() in key.lower()]
return gr.update(choices=sorted(filtered_keys))
def save_preset(presets, name, url):
if not name or not name.strip():
gr.Warning("Preset name cannot be empty."); return presets, gr.update()
if not url or not url.strip():
gr.Warning("Target URL cannot be empty."); return presets, gr.update()
presets[name] = url
gr.Info(f"Preset '{name}' saved!")
return presets, gr.update(choices=sorted(list(presets.keys())))
def delete_preset(presets, name):
if name in presets:
del presets[name]
gr.Info(f"Preset '{name}' deleted!")
return presets, gr.update(choices=sorted(list(presets.keys())), value=None), ""
gr.Warning(f"Preset '{name}' not found.")
return presets, gr.update(), gr.update()
def load_preset(presets, name):
return presets.get(name, "")
# --- Join/Beat-Sync/Etc Video Feature Functions ---
def ping_pong_video(video_path, audio_option):
if not video_path: raise gr.Error("Please upload a video.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"pingpong_{timestamp}"); os.makedirs(job_temp_dir, exist_ok=True)
reversed_video_path = os.path.join(job_temp_dir, "reversed_temp.mp4")
cmd_reverse = ["ffmpeg", "-i", video_path, "-vf", "reverse"]
if audio_option == "Reverse Audio": cmd_reverse.extend(["-af", "areverse"])
else: cmd_reverse.append("-an")
cmd_reverse.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", "-y", reversed_video_path])
run_ffmpeg_command(cmd_reverse)
file_list_path = os.path.join(job_temp_dir, "files.txt")
with open(file_list_path, 'w', encoding='utf-8') as f:
f.write(f"file '{os.path.abspath(video_path)}'\n")
f.write(f"file '{os.path.abspath(reversed_video_path)}'\n")
output_video_path = os.path.join(TEMP_DIR, f"pingpong_video_{timestamp}.mp4")
cmd_join = ["ffmpeg", "-f", "concat", "-safe", "0", "-i", file_list_path, "-c", "copy", "-y", output_video_path]
if audio_option == "Original Audio Only":
cmd_join = ["ffmpeg", "-i", video_path, "-i", reversed_video_path, "-filter_complex", "[0:v][1:v]concat=n=2:v=1[v]", "-map", "[v]", "-map", "0:a?", "-c:a", "copy", "-y", output_video_path]
run_ffmpeg_command(cmd_join)
shutil.rmtree(job_temp_dir)
return output_video_path
# ### --- NEW FEATURE: VIDEO STABILIZATION --- ###
def stabilize_video(video_path, shakiness, smoothing):
"""
Stabilizes a video using a two-pass FFMPEG process.
Pass 1: Detects motion vectors.
Pass 2: Uses the motion vectors to smooth the video.
"""
if not video_path:
raise gr.Error("Please upload a video to stabilize.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
transforms_path = os.path.join(TEMP_DIR, f"transforms_{timestamp}.trf")
output_video_path = os.path.join(TEMP_DIR, f"stabilized_{timestamp}.mp4")
try:
# Pass 1: Detect shakiness
detect_cmd = [
"ffmpeg", "-i", video_path,
"-vf", f"vidstabdetect=shakiness={shakiness}:result={transforms_path}",
"-f", "null", "-"
]
run_ffmpeg_command(detect_cmd, "Analyzing video for stabilization (Pass 1/2)...")
# Pass 2: Apply stabilization
transform_cmd = [
"ffmpeg", "-i", video_path,
"-vf", f"vidstabtransform=input={transforms_path}:smoothing={smoothing}:optalgo=gauss",
"-c:a", "copy",
"-c:v", "libx264", "-pix_fmt", "yuv420p", "-y",
output_video_path
]
run_ffmpeg_command(transform_cmd, "Applying stabilization (Pass 2/2)...")
finally:
# Clean up the temporary transforms file
if os.path.exists(transforms_path):
os.remove(transforms_path)
return output_video_path
# ### --- NEW FEATURE: AUTO JUMP-CUT & WAVEFORM PREVIEW --- ###
def generate_waveform_preview(video_path):
"""Generates a PNG image of the audio waveform."""
if not video_path or not has_audio_stream(video_path):
return None # Return None if no video or no audio
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_image_path = os.path.join(TEMP_DIR, f"waveform_{timestamp}.png")
# FFMPEG command to generate a waveform picture
cmd = [
"ffmpeg", "-i", video_path,
"-filter_complex", "aformat=channel_layouts=mono,compand,showwavespic=s=1280x240:colors=#38bdf8",
"-frames:v", "1",
"-y", output_image_path
]
try:
# Use subprocess.run and check for errors, but don't raise gr.Error to avoid stopping the UI
result = subprocess.run(cmd, capture_output=True, text=True, check=True)
return output_image_path
except (subprocess.CalledProcessError, FileNotFoundError) as e:
print(f"--- WAVEFORM GENERATION ERROR ---\n{e}")
return None
def auto_jump_cut(video_path, silence_threshold, min_silence_duration, resolution_choice, custom_w, custom_h, progress=gr.Progress(track_tqdm=True)):
"""
Automatically removes silent parts from a video and stitches the remaining parts together.
"""
if not video_path:
raise gr.Error("Please upload a video to process.")
if not has_audio_stream(video_path):
raise gr.Error("The uploaded video has no audio track. Cannot detect silence.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"jumpcut_{timestamp}.mp4")
# --- 1. Detect Silence ---
progress(0.1, desc="Analyzing for silent sections...")
silence_cmd = [
"ffmpeg", "-i", video_path,
"-af", f"silencedetect=noise={silence_threshold}dB:d={min_silence_duration}",
"-f", "null", "-"
]
print(f"Running silence detection: {' '.join(silence_cmd)}")
result = subprocess.run(silence_cmd, capture_output=True, text=True, encoding='utf-8')
silence_starts = [float(x) for x in re.findall(r'silence_start: (\d+\.?\d*)', result.stderr)]
silence_ends = [float(x) for x in re.findall(r'silence_end: (\d+\.?\d*)', result.stderr)]
if not silence_starts:
gr.Info("No silence was detected with the current settings. Returning original video.")
return video_path
silences = list(zip(silence_starts, silence_ends))
# --- 2. Calculate Segments to Keep ---
progress(0.3, desc="Calculating video cuts...")
video_duration = get_media_duration(video_path)
keep_segments = []
last_silence_end = 0.0
for start, end in silences:
if start > last_silence_end:
keep_segments.append((last_silence_end, start))
last_silence_end = end
if last_silence_end < video_duration:
keep_segments.append((last_silence_end, video_duration))
if not keep_segments:
raise gr.Error("Failed to calculate any segments to keep. Try adjusting silence parameters.")
# --- 3. Build the FFMPEG Filter Complex Command ---
progress(0.5, desc="Building FFMPEG command...")
scale_pad_filter = ""
target_w, target_h = get_video_dimensions(video_path)
if resolution_choice != "Keep Original":
if resolution_choice == "1080p (1920x1080)":
target_w, target_h = 1920, 1080
elif resolution_choice == "Portrait (1080x1920)":
target_w, target_h = 1080, 1920
elif resolution_choice == "Custom":
target_w, target_h = int(custom_w), int(custom_h)
target_w -= target_w % 2
target_h -= target_h % 2
scale_pad_filter = f"scale={target_w}:{target_h}:force_original_aspect_ratio=decrease,pad={target_w}:{target_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
filter_complex_parts = []
video_outputs = []
audio_outputs = []
for i, (start, end) in enumerate(keep_segments):
filter_complex_parts.append(f"[0:v]trim=start={start}:end={end},setpts=PTS-STARTPTS[v{i}]")
filter_complex_parts.append(f"[0:a]atrim=start={start}:end={end},asetpts=PTS-STARTPTS[a{i}]")
if scale_pad_filter:
filter_complex_parts.append(f"[v{i}]{scale_pad_filter}[scaled_v{i}]")
video_outputs.append(f"[scaled_v{i}]")
else:
video_outputs.append(f"[v{i}]")
audio_outputs.append(f"[a{i}]")
filter_complex_parts.append(f"{''.join(video_outputs)}concat=n={len(keep_segments)}:v=1:a=0[vout]")
filter_complex_parts.append(f"{''.join(audio_outputs)}concat=n={len(keep_segments)}:v=0:a=1[aout]")
filter_complex_str = ";".join(filter_complex_parts)
# --- 4. Execute the Final Command ---
final_cmd = [
"ffmpeg", "-i", video_path,
"-filter_complex", filter_complex_str,
"-map", "[vout]", "-map", "[aout]",
"-c:v", "libx264", "-pix_fmt", "yuv420p",
"-c:a", "aac",
"-y", output_video_path
]
progress(0.7, desc="Generating final jump-cut video...")
run_ffmpeg_command(final_cmd, desc="Stitching video segments...")
return output_video_path
# ### --- NEW FEATURE: VIDEO SILENCE CHOPPER --- ###
def chop_video_on_silence(video_path, silence_threshold, min_silence_duration, resolution_choice, custom_w, custom_h, progress=gr.Progress(track_tqdm=True)):
"""
Splits a video into multiple clips, removing the silent parts.
"""
if not video_path:
raise gr.Error("Please upload a video to process.")
if not has_audio_stream(video_path):
raise gr.Error("The uploaded video has no audio track. Cannot detect silence.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"video_chop_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
output_paths = []
progress(0.1, desc="Analyzing for silent sections...")
silence_cmd = [
"ffmpeg", "-i", video_path,
"-af", f"silencedetect=noise={silence_threshold}dB:d={min_silence_duration}",
"-f", "null", "-"
]
result = subprocess.run(silence_cmd, capture_output=True, text=True, encoding='utf-8')
silence_starts = [float(x) for x in re.findall(r'silence_start: (\d+\.?\d*)', result.stderr)]
silence_ends = [float(x) for x in re.findall(r'silence_end: (\d+\.?\d*)', result.stderr)]
if not silence_starts:
shutil.rmtree(job_temp_dir)
raise gr.Error("No silence was detected with the current settings. Try adjusting the parameters.")
silences = list(zip(silence_starts, silence_ends))
progress(0.3, desc="Calculating video cuts...")
video_duration = get_media_duration(video_path)
keep_segments = []
last_silence_end = 0.0
for start, end in silences:
if start > last_silence_end:
keep_segments.append((last_silence_end, start))
last_silence_end = end
if last_silence_end < video_duration:
keep_segments.append((last_silence_end, video_duration))
if not keep_segments:
shutil.rmtree(job_temp_dir)
raise gr.Error("Failed to calculate any segments to keep.")
vf_filter = None
if resolution_choice != "Keep Original":
if resolution_choice == "1080p (1920x1080)":
target_w, target_h = 1920, 1080
elif resolution_choice == "Portrait (1080x1920)":
target_w, target_h = 1080, 1920
elif resolution_choice == "Custom":
target_w, target_h = int(custom_w), int(custom_h)
target_w -= target_w % 2
target_h -= target_h % 2
vf_filter = f"scale={target_w}:{target_h}:force_original_aspect_ratio=decrease,pad={target_w}:{target_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
for i, (start, end) in enumerate(progress.tqdm(keep_segments, desc="Exporting video clips...")):
output_clip_path = os.path.join(job_temp_dir, f"clip_{i:04d}.mp4")
duration = end - start
cmd = ["ffmpeg", "-y", "-ss", str(start), "-to", str(end), "-i", video_path]
if vf_filter:
# Re-encoding is necessary
cmd.extend(["-vf", vf_filter, "-c:v", "libx264", "-pix_fmt", "yuv420p", "-c:a", "aac"])
else:
# Can use fast stream copy
cmd.extend(["-c", "copy"])
cmd.append(output_clip_path)
try:
run_ffmpeg_command(cmd)
output_paths.append(output_clip_path)
except Exception as e:
gr.Warning(f"Skipping a clip due to an error: {e}")
continue
if not output_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No video clips could be exported.")
zip_base_name = os.path.join(TEMP_DIR, f"video_chop_archive_{timestamp}")
zip_path = shutil.make_archive(zip_base_name, 'zip', job_temp_dir)
return output_paths, zip_path
# --- STORYBOARD / ANIMATIC CREATOR FUNCTIONS ---
def get_file_type(file_path):
if not file_path: return "unknown"
image_exts = ['.png', '.jpg', '.jpeg', '.webp', '.bmp', '.gif']
video_exts = ['.mp4', '.mov', '.mkv', '.avi', '.webm']
ext = os.path.splitext(file_path.lower())[1]
if ext in image_exts: return "image"
if ext in video_exts: return "video"
return "unknown"
def add_assets_to_bin(files, current_assets):
if not files:
return current_assets, gr.update(value=[a['path'] for a in current_assets] if current_assets else None)
session_id = f"storyboard_session_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
asset_session_dir = os.path.join(TEMP_DIR, session_id)
os.makedirs(asset_session_dir, exist_ok=True)
updated_asset_list = list(current_assets)
for file_obj in files:
try:
file_type = get_file_type(file_obj.name)
if file_type == "unknown":
gr.Warning(f"Skipping unknown file type: {os.path.basename(file_obj.name)}")
continue
new_path = os.path.join(asset_session_dir, os.path.basename(file_obj.name))
shutil.copy(file_obj.name, new_path)
updated_asset_list.append({"path": new_path, "name": os.path.basename(new_path), "type": file_type})
except Exception as e:
gr.Warning(f"Error adding asset {os.path.basename(file_obj.name)}: {e}")
return updated_asset_list, gr.update(value=[a['path'] for a in updated_asset_list])
def handle_asset_selection(evt: gr.SelectData, assets_state, timeline_state):
if not evt.selected:
return timeline_state, None
selected_asset = assets_state[evt.index]
new_timeline = list(timeline_state)
item_to_add = {
"path": selected_asset['path'],
"name": selected_asset['name'],
"type": selected_asset['type'],
}
if selected_asset['type'] == 'image':
item_to_add.update({
"duration": 3.0,
"start_time": 0,
"original_duration": 0
})
else: # video
original_duration = get_media_duration(selected_asset['path'])
if original_duration <= 0:
gr.Warning(f"Could not read duration for '{selected_asset['name']}'. Defaulting to 3.0 seconds. The file may be corrupt or in an unsupported format.")
original_duration = 3.0
item_to_add.update({
"duration": round(original_duration, 2),
"start_time": 0.0,
"original_duration": round(original_duration, 2)
})
new_timeline.append(item_to_add)
gr.Info(f"Added '{selected_asset['name']}' to timeline.")
preview_frames = None
if selected_asset['type'] == 'video':
try:
preview_frames = extract_first_last_frame(selected_asset['path'])
except Exception as e:
print(f"Could not generate preview for {selected_asset['name']}: {e}")
return new_timeline, preview_frames
def add_all_assets_to_timeline(assets_state, timeline_state):
if not assets_state:
gr.Warning("Asset bin is empty.")
return timeline_state
new_timeline = list(timeline_state)
for asset in assets_state:
item_to_add = {
"path": asset['path'],
"name": asset['name'],
"type": asset['type'],
}
if asset['type'] == 'image':
item_to_add.update({
"duration": 3.0,
"start_time": 0,
"original_duration": 0
})
else: # video
original_duration = get_media_duration(asset['path'])
if original_duration <= 0:
gr.Warning(f"Could not read duration for '{asset['name']}'. Defaulting to 3.0 seconds.")
original_duration = 3.0
item_to_add.update({
"duration": round(original_duration, 2),
"start_time": 0.0,
"original_duration": round(original_duration, 2)
})
new_timeline.append(item_to_add)
gr.Info(f"Added {len(assets_state)} assets to the timeline.")
return new_timeline
def update_timeline_df(timeline_state):
if not timeline_state: return gr.update(value=None)
df_data = [[i + 1, item['name'], item['type'], item['duration']] for i, item in enumerate(timeline_state)]
return gr.update(value=df_data)
def handle_timeline_selection(timeline_state, evt: gr.SelectData):
if not evt.selected:
return -1, None, None, gr.update(interactive=False), gr.update(interactive=False), gr.update(interactive=False), gr.update(visible=False), 0, 0
index = evt.index[0]
if not (0 <= index < len(timeline_state)):
return -1, None, None, gr.update(interactive=False), gr.update(interactive=False), gr.update(interactive=False), gr.update(visible=False), 0, 0
selected_item = timeline_state[index]
preview_val = selected_item['path']
duration_val = selected_item['duration']
can_move_up = index > 0
can_move_down = index < len(timeline_state) - 1
if selected_item['type'] == 'video':
start_time = selected_item.get('start_time', 0.0)
end_time = start_time + selected_item['duration']
return (index, preview_val, duration_val,
gr.update(interactive=can_move_up), gr.update(interactive=can_move_down), gr.update(interactive=True),
gr.update(visible=True), round(start_time, 2), round(end_time, 2))
else: # Image
return (index, preview_val, duration_val,
gr.update(interactive=can_move_up), gr.update(interactive=can_move_down), gr.update(interactive=True),
gr.update(visible=False), 0, 0)
def apply_trim_and_update(timeline_state, selected_index, new_start, new_end):
if selected_index == -1 or not (0 <= selected_index < len(timeline_state)):
gr.Warning("No clip selected in timeline.")
return timeline_state, gr.update()
item_to_update = timeline_state[selected_index]
if item_to_update['type'] != 'video':
gr.Warning("Trimming is only available for video clips.")
return timeline_state, gr.update()
original_duration = item_to_update.get('original_duration', 0)
if not (0 <= new_start < new_end and new_end <= original_duration):
gr.Warning(f"Invalid trim times. Must be between 0 and {original_duration:.2f}s, and start must be before end.")
return timeline_state, gr.update()
new_duration = new_end - new_start
new_timeline = list(timeline_state)
new_timeline[selected_index]['start_time'] = round(new_start, 2)
new_timeline[selected_index]['duration'] = round(new_duration, 2)
gr.Info(f"Clip '{item_to_update['name']}' trimmed. New duration is {new_duration:.2f}s.")
return new_timeline, gr.update(value=round(new_duration, 2))
def update_clip_properties(timeline_state, selected_index, new_duration):
if selected_index == -1 or not (0 <= selected_index < len(timeline_state)):
gr.Warning("No clip selected in timeline.")
return timeline_state
if new_duration <= 0:
gr.Warning("Duration must be a positive number.")
return timeline_state
new_timeline = list(timeline_state)
item_to_update = new_timeline[selected_index]
if item_to_update['type'] == 'video':
start_time = item_to_update.get('start_time', 0.0)
original_duration = item_to_update.get('original_duration', 0.0)
max_possible_duration = original_duration - start_time
if new_duration > max_possible_duration:
gr.Warning(f"Duration cannot exceed available video length from start time ({max_possible_duration:.2f}s). Clamping value.")
new_duration = max_possible_duration
item_to_update['duration'] = round(new_duration, 2)
gr.Info(f"Updated duration for '{item_to_update['name']}'.")
return new_timeline
def handle_timeline_action(timeline_state, selected_index, action):
if selected_index == -1 or not (0 <= selected_index < len(timeline_state)):
gr.Warning("Please select a clip from the timeline first.")
return timeline_state, gr.update()
new_list = list(timeline_state)
new_index = selected_index
if action == "up" and selected_index > 0:
new_list.insert(selected_index - 1, new_list.pop(selected_index))
new_index = selected_index - 1
elif action == "down" and selected_index < len(new_list) - 1:
new_list.insert(selected_index + 1, new_list.pop(selected_index))
new_index = selected_index + 1
elif action == "remove":
new_list.pop(selected_index)
new_index = -1 # Deselect after removing
# Return the new list and tell the UI to select the new index
return new_list, gr.update(selected_index=new_index if new_index != -1 else None)
def set_resolution_from_first_asset(timeline_state):
if not timeline_state:
gr.Warning("Timeline is empty. Cannot determine resolution.")
return gr.update(), gr.update()
first_item = timeline_state[0]
path = first_item['path']
item_type = first_item['type']
w, h = 0, 0
if item_type == 'video':
w, h = get_video_dimensions(path)
elif item_type == 'image':
try:
with Image.open(path) as img:
w, h = img.size
except Exception as e:
print(f"Could not get image dimensions for {path}: {e}")
if w > 0 and h > 0:
gr.Info(f"Set resolution to {w}x{h} based on '{first_item['name']}'.")
return w, h
else:
gr.Warning(f"Could not get dimensions for the first asset: '{first_item['name']}'.")
return gr.update(), gr.update()
def create_animatic(timeline_data, audio_path, out_w, out_h, keep_original_audio):
if not timeline_data:
raise gr.Error("Timeline is empty. Please add assets to the timeline.")
out_w, out_h = int(out_w), int(out_h)
if out_w <= 0 or out_h <= 0:
raise gr.Error("Output width and height must be positive numbers.")
out_w -= out_w % 2
out_h -= out_h % 2
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"animatic_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
clip_paths = []
for i, item in enumerate(timeline_data):
item_path, item_type, item_duration = item['path'], item['type'], item['duration']
start_time = item.get('start_time', 0)
if item_duration <= 0:
gr.Warning(f"Skipping clip '{item['name']}' because its duration is zero.")
continue
output_clip_path = os.path.join(job_temp_dir, f"clip_{i:04d}.mp4")
cmd = ["ffmpeg", "-y"]
vf_base_scale = f"scale={out_w}:{out_h}:force_original_aspect_ratio=decrease,pad={out_w}:{out_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
if item_type == 'video':
video_has_audio = has_audio_stream(item_path)
if start_time > 0:
cmd.extend(["-ss", str(start_time)])
cmd.extend(["-t", str(item_duration), "-i", item_path])
vf_filters = [f"setpts=PTS-STARTPTS", vf_base_scale]
cmd.extend(["-vf", ",".join(vf_filters)])
if keep_original_audio:
if video_has_audio:
cmd.extend(["-af", "asetpts=PTS-STARTPTS"])
cmd.extend(["-c:a", "aac", "-ar", "44100"])
else:
cmd.extend(["-f", "lavfi", "-t", str(item_duration), "-i", "anullsrc=channel_layout=stereo:sample_rate=44100"])
cmd.extend(["-map", "0:v:0", "-map", "1:a:0"])
cmd.extend(["-c:a", "aac", "-ar", "44100"])
else:
cmd.append("-an")
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p"])
else: # item_type == 'image'
cmd.extend(["-loop", "1", "-i", item_path, "-t", str(item_duration)])
vf_filter_img = f"{vf_base_scale},format=yuv420p"
cmd.extend(["-vf", vf_filter_img])
if keep_original_audio:
cmd.extend(["-f", "lavfi", "-t", str(item_duration), "-i", "anullsrc=channel_layout=stereo:sample_rate=44100", "-shortest"])
cmd.extend(["-c:a", "aac", "-ar", "44100"])
else:
cmd.append("-an")
cmd.append(output_clip_path)
run_ffmpeg_command(cmd, f"Processing clip {i+1}/{len(timeline_data)}: {item['name']}")
clip_paths.append(output_clip_path)
if not clip_paths:
shutil.rmtree(job_temp_dir)
raise gr.Error("No valid clips were generated. Check clip durations and file integrity.")
combined_video_path = os.path.join(job_temp_dir, "combined_video.mp4")
if len(clip_paths) > 1:
cmd_concat = ["ffmpeg", "-y"]
video_inputs, audio_inputs = [], []
for i, path in enumerate(clip_paths):
cmd_concat.extend(["-i", path])
video_inputs.append(f"[{i}:v]")
if keep_original_audio: audio_inputs.append(f"[{i}:a]")
filter_complex_str = ""
if keep_original_audio:
video_concat_str = "".join(video_inputs) + f"concat=n={len(clip_paths)}:v=1:a=0[v_out];"
audio_concat_str = "".join(audio_inputs) + f"concat=n={len(clip_paths)}:v=0:a=1[a_out]"
filter_complex_str = video_concat_str + audio_concat_str
cmd_concat.extend(["-filter_complex", filter_complex_str, "-map", "[v_out]", "-map", "[a_out]"])
else:
video_concat_str = "".join(video_inputs) + f"concat=n={len(clip_paths)}:v=1:a=0[v_out]"
filter_complex_str = video_concat_str
cmd_concat.extend(["-filter_complex", filter_complex_str, "-map", "[v_out]"])
cmd_concat.append(combined_video_path)
run_ffmpeg_command(cmd_concat, "Joining and Finalizing Video (Robust Mode)...")
else:
if os.path.exists(clip_paths[0]):
shutil.copy(clip_paths[0], combined_video_path)
else:
shutil.rmtree(job_temp_dir)
raise gr.Error("The only clip in the timeline failed to process.")
final_output_path = os.path.join(TEMP_DIR, f"animatic_final_{timestamp}.mp4")
if not keep_original_audio and audio_path:
run_ffmpeg_command(["ffmpeg", "-y", "-i", combined_video_path, "-i", audio_path, "-c:v", "copy", "-c:a", "aac", "-shortest", final_output_path], "Muxing audio")
else:
shutil.move(combined_video_path, final_output_path)
shutil.rmtree(job_temp_dir)
return final_output_path
def detect_bpm(audio_path):
if not audio_path:
return "Please upload an audio track first."
try:
y, sr = librosa.load(audio_path)
tempo_val, _ = librosa.beat.beat_track(y=y, sr=sr)
if isinstance(tempo_val, np.ndarray):
tempo = tempo_val.item()
else:
tempo = float(tempo_val)
if tempo > 0:
return f"Detected BPM: {tempo:.2f}"
else:
return "Could not detect BPM."
except Exception as e:
print(f"--- BPM DETECTION ERROR ---\n{e}")
return "Error: Could not analyze audio file."
def update_new_bpm_display(original_bpm_text, speed_multiplier):
if not original_bpm_text or "Detected" not in original_bpm_text:
return "---"
try:
bpm_match = re.search(r"(\d+\.\d+)", original_bpm_text)
if bpm_match:
original_bpm = float(bpm_match.group(1))
new_bpm = original_bpm * speed_multiplier
return f"Estimated New BPM: {new_bpm:.2f}"
else:
return "---"
except (ValueError, TypeError):
return "---"
def create_rhythmic_animatic(timeline_data, audio_path, measure_choice, out_w, out_h):
if not timeline_data: raise gr.Error("Timeline is empty.")
if not audio_path: raise gr.Error("An audio track is required for rhythmic editing.")
try:
y, sr = librosa.load(audio_path)
tempo_val, _ = librosa.beat.beat_track(y=y, sr=sr)
if isinstance(tempo_val, np.ndarray):
tempo = tempo_val.item()
else:
tempo = float(tempo_val)
if not tempo or tempo <= 0:
raise gr.Error("Could not determine BPM from audio file.")
except Exception as e:
raise gr.Error(f"Audio analysis failed: {e}")
seconds_per_beat = 60.0 / tempo
seconds_per_measure = seconds_per_beat * 4.0
measure_multipliers = { "2 Measures": 2.0, "1 Measure": 1.0, "1/2 Measure": 0.5, "1/4 Measure (Beat)": 0.25 }
clip_duration = seconds_per_measure * measure_multipliers[measure_choice]
rhythmic_timeline = []
for item in timeline_data:
new_item = item.copy()
if new_item['type'] == 'video':
start_time = new_item.get('start_time', 0)
available_duration = new_item.get('original_duration', 0) - start_time
new_item['duration'] = min(clip_duration, available_duration)
else:
new_item['duration'] = clip_duration
rhythmic_timeline.append(new_item)
gr.Info(f"Re-timed {len(rhythmic_timeline)} clips to ~{clip_duration:.2f}s each based on {tempo:.2f} BPM.")
return create_animatic(rhythmic_timeline, audio_path, out_w, out_h, keep_original_audio=False)
# --- NEW CREATIVE FUNCTIONS ---
def _create_auto_trailer_impl(video_path, trailer_duration, clip_duration, analysis_method, transition_style, music_path, out_w, out_h, progress: Progress):
"""Internal implementation of the auto-trailer creator."""
if not video_path: raise gr.Error("Please upload a source video.")
source_duration = get_media_duration(video_path)
if source_duration < trailer_duration:
gr.Warning(f"Source video is only {source_duration:.1f}s long. The trailer duration will be capped at the source video length.")
trailer_duration = source_duration
if clip_duration > trailer_duration:
new_clip_duration = trailer_duration / 2 if trailer_duration > 2 else trailer_duration
gr.Warning(f"Clip duration ({clip_duration}s) is longer than the trailer duration ({trailer_duration:.1f}s). Adjusting clip duration to {new_clip_duration:.1f}s.")
clip_duration = new_clip_duration
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"trailer_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
try:
progress(0, desc="Analyzing video for high-motion scenes...")
cap = cv2.VideoCapture(video_path)
fps = cap.get(cv2.CAP_PROP_FPS)
if fps == 0: fps = 30 # fallback
chunk_duration_frames = int(clip_duration * fps)
video_total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
chunk_scores = []
prev_frame = None
frame_num = 0
while cap.isOpened():
ret, frame = cap.read()
if not ret: break
frame_skip = max(1, int(fps / 5)) # Analyze ~5 frames per second
if frame_num % frame_skip == 0:
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray, (21, 21), 0)
if prev_frame is not None:
frame_delta = cv2.absdiff(prev_frame, gray)
thresh = cv2.threshold(frame_delta, 25, 255, cv2.THRESH_BINARY)[1]
motion_score = np.sum(thresh)
chunk_index = frame_num // chunk_duration_frames
while len(chunk_scores) <= chunk_index:
chunk_scores.append({'start_time': (len(chunk_scores) * clip_duration), 'scores': []})
chunk_scores[chunk_index]['scores'].append(motion_score)
prev_frame = gray
frame_num += 1
if frame_num % 100 == 0:
progress(0.2 * (frame_num / video_total_frames), desc=f"Analyzing frame {frame_num}/{video_total_frames}...")
cap.release()
final_chunks = [{'start_time': chunk['start_time'], 'score': sum(chunk['scores']) / len(chunk['scores'])} for chunk in chunk_scores if chunk['scores']]
if not final_chunks: raise gr.Error("Could not analyze video for motion. Is the video very short or static?")
progress(0.2, desc="Selecting the best clips...")
num_clips_to_select = max(1, int(trailer_duration / clip_duration))
selected_clips_info = sorted(sorted(final_chunks, key=lambda x: x['score'], reverse=True)[:num_clips_to_select], key=lambda x: x['start_time'])
extracted_clips, out_w, out_h = [], int(out_w) - (int(out_w) % 2), int(out_h) - (int(out_h) % 2)
for i, clip_info in enumerate(selected_clips_info):
progress(0.2 + (0.5 * (i / len(selected_clips_info))), desc=f"Extracting clip {i+1}/{len(selected_clips_info)}...")
output_clip_path = os.path.join(job_temp_dir, f"clip_{i:03d}.mp4")
vf_filter = f"scale={out_w}:{out_h}:force_original_aspect_ratio=decrease,pad={out_w}:{out_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
cmd = ["ffmpeg", "-y", "-ss", str(clip_info['start_time']), "-i", video_path, "-t", str(clip_duration), "-vf", vf_filter, "-an", "-c:v", "libx264", "-pix_fmt", "yuv420p", output_clip_path]
run_ffmpeg_command(cmd)
extracted_clips.append(output_clip_path)
if not extracted_clips: raise gr.Error("Failed to extract any clips.")
progress(0.7, desc="Stitching clips together...")
final_silent_path = os.path.join(job_temp_dir, "final_silent.mp4")
if transition_style == "None" or len(extracted_clips) == 1:
if len(extracted_clips) == 1:
shutil.copy(extracted_clips[0], final_silent_path)
else:
file_list_path = os.path.join(job_temp_dir, "files.txt")
with open(file_list_path, 'w', encoding='utf-8') as f:
for path in extracted_clips: f.write(f"file '{os.path.abspath(path)}'\n")
run_ffmpeg_command(["ffmpeg", "-y", "-f", "concat", "-safe", "0", "-i", file_list_path, "-c", "copy", final_silent_path], "Concatenating clips...")
else:
transition_duration = 0.5
cmd = ["ffmpeg", "-y"]
filter_complex = []
running_duration = 0
for i, clip_path in enumerate(extracted_clips):
cmd.extend(["-i", clip_path])
for i in range(len(extracted_clips) - 1):
input1 = f"[{i}:v]" if i == 0 else f"[v{i-1}]"
input2 = f"[{i+1}:v]"
output = f"[v{i}]"
offset = max(0, running_duration + clip_duration - transition_duration)
filter_complex.append(f"{input1}{input2}xfade=transition={transition_style.lower()}:duration={transition_duration}:offset={offset}{output}")
running_duration += clip_duration - transition_duration
cmd.extend([
"-filter_complex", ";".join(filter_complex),
"-map", f"[v{len(extracted_clips)-2}]",
"-c:v", "libx264", "-pix_fmt", "yuv420p",
final_silent_path
])
run_ffmpeg_command(cmd, "Applying transitions...")
progress(0.95, desc="Adding background music...")
final_output_path = os.path.join(TEMP_DIR, f"trailer_final_{timestamp}.mp4")
if music_path:
run_ffmpeg_command(["ffmpeg", "-y", "-i", final_silent_path, "-i", music_path, "-c:v", "copy", "-c:a", "aac", "-map", "0:v:0", "-map", "1:a:0", "-shortest", final_output_path], "Muxing audio")
else:
shutil.move(final_silent_path, final_output_path)
return final_output_path
finally:
if os.path.exists(job_temp_dir): shutil.rmtree(job_temp_dir)
def auto_trailer_wrapper(video_path, trailer_duration, clip_duration, analysis_method, transition_style, music_path, out_w, out_h, progress=gr.Progress(track_tqdm=True)):
return _create_auto_trailer_impl(video_path, trailer_duration, clip_duration, analysis_method, transition_style, music_path, out_w, out_h, progress)
def generate_waveform_video(video_path, style, size, position, color):
if not video_path: raise gr.Error("Please upload a video first.")
if not has_audio_stream(video_path):
raise gr.Error("The uploaded video has no audio track. A waveform cannot be generated.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"waveform_video_{timestamp}.mp4")
pos_map = {
"Bottom": f"overlay=x=(W-w)/2:y=H-h-50",
"Center": f"overlay=x=(W-w)/2:y=(H-h)/2",
"Top": f"overlay=x=(W-w)/2:y=50"
}
safe_color = color.lstrip('#')
filter_complex = (
f"[0:a]showwaves=s={size}:mode={style}:colors={safe_color}:rate=25[wave];"
f"[0:v][wave]{pos_map[position]}"
)
cmd = [
"ffmpeg", "-i", video_path,
"-filter_complex", filter_complex,
"-c:a", "copy",
"-c:v", "libx264", "-pix_fmt", "yuv420p", "-y",
output_video_path
]
run_ffmpeg_command(cmd, "Generating Audio Waveform...")
return output_video_path
def create_pip_video(main_video, overlay_media, position, scale):
if not main_video: raise gr.Error("Please upload a main video.")
if not overlay_media: raise gr.Error("Please upload an overlay video or image.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_video_path = os.path.join(TEMP_DIR, f"pip_video_{timestamp}.mp4")
scale_filter = f"[1:v]scale=iw*{scale}:-1[scaled_overlay]"
pos_map = {
"Top-Left": "x=10:y=10",
"Top-Center": "x=(W-w)/2:y=10",
"Top-Right": "x=W-w-10:y=10",
"Center-Left": "x=10:y=(H-h)/2",
"Center": "x=(W-w)/2:y=(H-h)/2",
"Center-Right": "x=W-w-10:y=(H-h)/2",
"Bottom-Left": "x=10:y=H-h-10",
"Bottom-Center": "x=(W-w)/2:y=H-h-10",
"Bottom-Right": "x=W-w-10:y=H-h-10"
}
overlay_filter = f"[0:v][scaled_overlay]overlay={pos_map[position]}"
cmd = ["ffmpeg", "-i", main_video, "-i", overlay_media.name]
cmd.extend([
"-filter_complex", f"{scale_filter};{overlay_filter}",
"-map", "0:a?", "-c:a", "copy",
"-c:v", "libx264", "-pix_fmt", "yuv420p", "-y",
output_video_path
])
run_ffmpeg_command(cmd, "Creating Picture-in-Picture video...")
return output_video_path
def create_meme(image, text_input, position, font_choice, font_size_scale, text_color, outline_color):
if image is None: raise gr.Error("Please upload an image.")
parsed_text_color = parse_color(text_color)
parsed_outline_color = parse_color(outline_color)
img = Image.fromarray(image).convert("RGB")
draw = ImageDraw.Draw(img)
FONT_MAP = {
"Impact": "impact.ttf",
"Arial": "arial.ttf",
"Arial Black": "ariblk.ttf",
"Comic Sans MS": "comic.ttf",
"Courier New": "cour.ttf",
"Georgia": "georgia.ttf",
"Tahoma": "tahoma.ttf",
"Times New Roman": "times.ttf",
"Trebuchet MS": "trebuc.ttf",
"Verdana": "verdana.ttf"
}
font_path = FONT_MAP.get(font_choice, "impact.ttf")
try:
font_size = int(img.width / 10 * (font_size_scale / 5))
font = ImageFont.truetype(font_path, font_size)
except IOError:
gr.Warning(f"{font_choice} font ('{font_path}') not found. Trying Arial.")
try:
font_path = FONT_MAP["Arial"]
font = ImageFont.truetype(font_path, font_size)
except IOError:
gr.Warning("Arial font not found. Using default font.")
font = ImageFont.load_default()
def draw_text_with_outline(text, x, y):
# Outline
draw.text((x-2, y-2), text, font=font, fill=parsed_outline_color)
draw.text((x+2, y-2), text, font=font, fill=parsed_outline_color)
draw.text((x-2, y+2), text, font=font, fill=parsed_outline_color)
draw.text((x+2, y+2), text, font=font, fill=parsed_outline_color)
# Main Text
draw.text((x, y), text, font=font, fill=parsed_text_color)
if text_input:
bbox = draw.textbbox((0, 0), text_input.upper(), font=font)
text_width = bbox[2] - bbox[0]
text_height = bbox[3] - bbox[1]
x = (img.width - text_width) / 2
if position == "Top":
y = 10
elif position == "Bottom":
y = img.height - text_height - 10
else: # Center
y = (img.height - text_height) / 2
draw_text_with_outline(text_input.upper(), x, y)
return img
def stitch_images_smartly(img1_np, img2_np, output_size, bg_color_hex):
"""
Stitches two images together into a square.
- If input images are vertical (based on first image), they are placed side-by-side.
- If input images are horizontal, they are stacked vertically.
"""
if img1_np is None or img2_np is None:
raise gr.Error("Please upload two images.")
# Convert inputs to PIL Images
img1 = Image.fromarray(img1_np).convert("RGBA")
img2 = Image.fromarray(img2_np).convert("RGBA")
# Parse background color using the utility function
bg_color = parse_color(bg_color_hex)
# Create the final square canvas
final_image = Image.new("RGB", (output_size, output_size), bg_color)
# Determine orientation from the first image
w1, h1 = img1.size
is_vertical = h1 > w1
if is_vertical:
# --- Place two vertical images side-by-side ---
target_box_w = output_size // 2
target_box_h = output_size
# Process image 1: resize and paste centered in the left box
img1.thumbnail((target_box_w, target_box_h), Image.Resampling.LANCZOS)
x1_offset = (target_box_w - img1.width) // 2
y1_offset = (target_box_h - img1.height) // 2
final_image.paste(img1, (x1_offset, y1_offset), img1)
# Process image 2: resize and paste centered in the right box
img2.thumbnail((target_box_w, target_box_h), Image.Resampling.LANCZOS)
x2_offset = target_box_w + (target_box_w - img2.width) // 2
y2_offset = (target_box_h - img2.height) // 2
final_image.paste(img2, (x2_offset, y2_offset), img2)
else:
# --- Stack two horizontal images vertically ---
target_box_w = output_size
target_box_h = output_size // 2
# Process image 1: resize and paste centered in the top box
img1.thumbnail((target_box_w, target_box_h), Image.Resampling.LANCZOS)
x1_offset = (target_box_w - img1.width) // 2
y1_offset = (target_box_h - img1.height) // 2
final_image.paste(img1, (x1_offset, y1_offset), img1)
# Process image 2: resize and paste centered in the bottom box
img2.thumbnail((target_box_w, target_box_h), Image.Resampling.LANCZOS)
x2_offset = (target_box_w - img2.width) // 2
y2_offset = target_box_h + (target_box_h - img2.height) // 2
final_image.paste(img2, (x2_offset, y2_offset), img2)
return final_image
def merge_videos(videos):
if not videos or len(videos) < 2:
raise gr.Error("Please upload at least two videos to merge.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"merge_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
first_video_path = videos[0].name
w, h = get_video_dimensions(first_video_path)
fps = get_video_fps(first_video_path)
w -= w % 2
h -= h % 2
processed_clips = []
for i, video_file in enumerate(videos):
clip_path = os.path.join(job_temp_dir, f"clip_{i}.mp4")
cmd = [
"ffmpeg", "-i", video_file.name,
"-vf", f"scale={w}:{h},setsar=1", "-r", str(fps),
"-c:v", "libx264", "-pix_fmt", "yuv420p",
"-c:a", "aac", "-ar", "44100",
"-y", clip_path
]
run_ffmpeg_command(cmd)
processed_clips.append(clip_path)
file_list_path = os.path.join(job_temp_dir, "files.txt")
with open(file_list_path, 'w', encoding='utf-8') as f:
for path in processed_clips:
f.write(f"file '{os.path.abspath(path)}'\n")
output_video_path = os.path.join(TEMP_DIR, f"merged_video_{timestamp}.mp4")
cmd_merge = [
"ffmpeg", "-f", "concat", "-safe", "0", "-i", file_list_path,
"-c", "copy", "-y", output_video_path
]
run_ffmpeg_command(cmd_merge, "Merging videos...")
shutil.rmtree(job_temp_dir)
return output_video_path
# ### --- NEW: BATCH IMAGE CROPPER --- ###
def update_crop_preview(original_image, x, y, w, h):
if original_image is None:
return None
# Create a copy to draw on
preview_image = original_image.copy()
draw = ImageDraw.Draw(preview_image)
# Define the bounding box for the crop area
box = (x, y, x + w, y + h)
# Draw a rectangle outline
draw.rectangle(box, outline="#38bdf8", width=3)
return preview_image
def crop_single_image(input_path, output_path, **kwargs):
x = int(kwargs.get('x', 0))
y = int(kwargs.get('y', 0))
w = int(kwargs.get('w', 512))
h = int(kwargs.get('h', 512))
with Image.open(input_path) as img:
cropped_img = img.crop((x, y, x + w, y + h))
cropped_img.save(output_path)
def batch_crop_images(files, x, y, w, h):
if not files: raise gr.Error("Please upload at least one image.")
if w <= 0 or h <= 0: raise gr.Error("Width and Height must be positive.")
processing_kwargs = {'x': x, 'y': y, 'w': w, 'h': h}
output_paths, zip_path, _ = batch_image_processor(
files,
crop_single_image,
"cropped",
**processing_kwargs
)
return output_paths, zip_path
# ### --- NEW: COLLAGE MAKER --- ###
def create_collage(files, layout, width, height, bg_color_hex):
if not files:
raise gr.Error("Please upload images to create a collage.")
bg_color = parse_color(bg_color_hex)
images = [Image.open(file.name).convert("RGBA") for file in files]
n = len(images)
if layout == "Grid":
cols = int(math.ceil(math.sqrt(n)))
rows = int(math.ceil(n / cols))
elif layout == "Horizontal":
cols, rows = n, 1
else: # Vertical
cols, rows = 1, n
cell_w = width // cols
cell_h = height // rows
canvas = Image.new("RGB", (width, height), bg_color)
for i, img in enumerate(images):
row = i // cols
col = i % cols
img.thumbnail((cell_w, cell_h), Image.Resampling.LANCZOS)
paste_x = (col * cell_w) + (cell_w - img.width) // 2
paste_y = (row * cell_h) + (cell_h - img.height) // 2
canvas.paste(img, (paste_x, paste_y), img)
return canvas
# ### --- NEW: VIDEO GRID COMPILER --- ###
def update_audio_source_choices_for_grid(files):
if not files:
return gr.update(choices=["From Video 1", "None"], value="From Video 1")
choices = [f"From Video {i+1}" for i in range(len(files))]
choices.append("None")
return gr.update(choices=choices, value=choices[0])
def compile_video_grid(videos, layout, width, height, bg_color, audio_choice, music_path):
if not videos:
raise gr.Error("Please upload videos to compile.")
num_videos = len(videos)
layout_map = {
"2x1 (Side-by-Side)": 2,
"1x2 (Stacked)": 2,
"2x2 (Quad-View)": 4,
"4x4 (16-View)": 16,
"8x4 (32-View)": 32,
}
required_videos = layout_map.get(layout)
if num_videos != required_videos:
raise gr.Error(f"The '{layout}' layout requires exactly {required_videos} videos, but you uploaded {num_videos}.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
output_path = os.path.join(TEMP_DIR, f"grid_video_{timestamp}.mp4")
width, height = int(width) - (int(width) % 2), int(height) - (int(height) % 2)
cmd = ["ffmpeg", "-y"]
input_count = 0
for video_file in videos:
cmd.extend(["-i", video_file.name])
input_count += 1
if music_path:
cmd.extend(["-i", music_path])
filter_complex_parts = []
if layout in ["2x1 (Side-by-Side)", "1x2 (Stacked)", "2x2 (Quad-View)"]:
if layout == "2x1 (Side-by-Side)":
tile_w, tile_h = width // 2, height
filter_complex_parts.append(f"[0:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v0];[1:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v1];[v0][v1]hstack=inputs=2[vout]")
elif layout == "1x2 (Stacked)":
tile_w, tile_h = width, height // 2
filter_complex_parts.append(f"[0:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v0];[1:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v1];[v0][v1]vstack=inputs=2[vout]")
elif layout == "2x2 (Quad-View)":
tile_w, tile_h = width // 2, height // 2
for i in range(4):
filter_complex_parts.append(f"[{i}:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v{i}]")
filter_complex_parts.append("[v0][v1][v2][v3]xstack=inputs=4:layout=0_0|w0_0|0_h0|w0_h0[vout]")
elif layout in ["4x4 (16-View)", "8x4 (32-View)"]:
cols, rows = (4, 4) if layout == "4x4 (16-View)" else (8, 4)
tile_w, tile_h = width // cols, height // rows
# 1. Scale all inputs
for i in range(required_videos):
filter_complex_parts.append(f"[{i}:v]scale={tile_w}:{tile_h}:force_original_aspect_ratio=decrease,pad={tile_w}:{tile_h}:-1:-1:color={bg_color}[v{i}]")
# 2. Horizontally stack videos for each row
row_outputs = []
for r in range(rows):
start_index = r * cols
end_index = start_index + cols
row_inputs = "".join([f"[v{i}]" for i in range(start_index, end_index)])
row_output_label = f"[row{r}]"
filter_complex_parts.append(f"{row_inputs}hstack=inputs={cols}{row_output_label}")
row_outputs.append(row_output_label)
# 3. Vertically stack all the rows
final_vstack_inputs = "".join(row_outputs)
filter_complex_parts.append(f"{final_vstack_inputs}vstack=inputs={rows}[vout]")
cmd.extend(["-filter_complex", ";".join(filter_complex_parts)])
cmd.extend(["-map", "[vout]"])
# Audio mapping logic
if music_path:
cmd.extend(["-map", f"{input_count}:a?"]) # Map the external audio track
cmd.extend(["-c:a", "aac", "-shortest"])
elif audio_choice != "None":
try:
audio_idx_match = re.search(r'\d+', audio_choice)
if audio_idx_match:
audio_idx = int(audio_idx_match.group()) - 1
if 0 <= audio_idx < num_videos:
cmd.extend(["-map", f"{audio_idx}:a?"])
cmd.extend(["-c:a", "aac", "-shortest"])
except (AttributeError, IndexError):
raise gr.Error("Invalid audio source selected.")
cmd.extend(["-c:v", "libx264", "-pix_fmt", "yuv420p", output_path])
run_ffmpeg_command(cmd, "Compiling video grid...")
return output_path
def _create_automated_slideshow_impl(images, audio_path, kb_effect_style, transition_style, rhythm_choice, out_w, out_h, progress: Progress):
"""Internal implementation of the slideshow creator with progress tracking."""
if not images: raise gr.Error("Please upload at least one image.")
if not audio_path: raise gr.Error("Please upload an audio track for rhythmic editing.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"slideshow_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
try:
out_w, out_h = int(out_w) - (int(out_w) % 2), int(out_h) - (int(out_h) % 2)
output_res_str = f"{out_w}x{out_h}"
fps = 30
transition_duration = 0.5
progress(0, desc="Analyzing audio track...")
try:
y, sr = librosa.load(audio_path)
audio_duration = librosa.get_duration(y=y, sr=sr)
_, beat_frames = librosa.beat.beat_track(y=y, sr=sr, units='frames')
beat_times = librosa.frames_to_time(beat_frames, sr=sr)
except Exception as e:
raise gr.Error(f"Audio analysis failed: {e}")
beats_per_clip = {"1 Image per Beat": 1, "1 Image every 2 Beats": 2, "1 Image per Measure (4 Beats)": 4}[rhythm_choice]
clip_start_times = [0.0] + [beat_times[i] for i in range(beats_per_clip, len(beat_times), beats_per_clip)]
MAX_CLIPS = 200
if len(clip_start_times) > MAX_CLIPS:
gr.Warning(f"Audio beat detection resulted in {len(clip_start_times)} clips. Capping at {MAX_CLIPS} to ensure performance.")
clip_start_times = clip_start_times[:MAX_CLIPS]
num_clips = len(clip_start_times)
image_paths = [img.name for img in images]
looped_image_paths = [image_paths[i % len(image_paths)] for i in range(num_clips)]
kb_clips = []
total_steps = num_clips + 1
current_step = 0
for i in range(num_clips):
progress(current_step / total_steps, desc=f"Creating clip {i+1}/{num_clips}")
start_time = clip_start_times[i]
end_time = clip_start_times[i + 1] if i + 1 < len(clip_start_times) else audio_duration
clip_duration = end_time - start_time
if clip_duration <= transition_duration: continue
total_frames = int(clip_duration * fps)
if total_frames <= 0: continue
output_clip_path = os.path.join(job_temp_dir, f"kb_clip_{i:04d}.mp4")
with Image.open(looped_image_paths[i]) as img:
iw, ih = img.size
zoom_levels = {"Subtle": (1.1, 1.15), "Standard": (1.1, 1.25), "Dynamic": (1.2, 1.5)}
start_zoom = 1.0
end_zoom = random.uniform(*zoom_levels[kb_effect_style])
directions = ['top_left', 'top_right', 'bottom_left', 'bottom_right', 'center']
start_pos_name, end_pos_name = random.sample(directions, 2)
def get_xy(pos_name, zoom_val, img_w, img_h):
if pos_name == 'center': return (img_w/2 - (img_w/zoom_val)/2, img_h/2 - (img_h/zoom_val)/2)
if pos_name == 'top_left': return (0, 0)
if pos_name == 'top_right': return (img_w - img_w/zoom_val, 0)
if pos_name == 'bottom_left': return (0, img_h - img_h/zoom_val)
if pos_name == 'bottom_right': return (img_w - img_w/zoom_val, img_h - img_h/zoom_val)
return (0,0)
start_x, start_y = get_xy(start_pos_name, start_zoom, iw, ih)
end_x, end_y = get_xy(end_pos_name, end_zoom, iw, ih)
x_expr = f"{start_x}+({end_x}-({start_x}))*on/({total_frames}-1)"
y_expr = f"{start_y}+({end_y}-({start_y}))*on/({total_frames}-1)"
z_expr = f"if(lte(on,0),{start_zoom},{start_zoom}+({end_zoom}-{start_zoom})*on/({total_frames}-1))"
zoompan_filter = f"zoompan=z='{z_expr}':x='{x_expr}':y='{y_expr}':d={total_frames}:s={output_res_str}:fps={fps}"
cmd = ["ffmpeg", "-y", "-loop", "1", "-i", looped_image_paths[i], "-vf", zoompan_filter, "-t", str(clip_duration), "-c:v", "libx264", "-pix_fmt", "yuv420p", output_clip_path]
run_ffmpeg_command(cmd)
kb_clips.append({"path": output_clip_path, "duration": clip_duration})
current_step += 1
if not kb_clips: raise gr.Error("No clips were generated. The audio may be too short or the rhythm settings too fast.")
progress(current_step / total_steps, desc=f"Applying transitions...")
final_silent_path = os.path.join(job_temp_dir, "final_silent.mp4")
if len(kb_clips) == 1:
shutil.copy(kb_clips[0]['path'], final_silent_path)
else:
all_transitions = ["fade", "wipeleft", "wiperight", "wipeup", "wipedown", "slideleft", "slideright", "slideup", "slidedown", "dissolve"]
cmd = ["ffmpeg", "-y"]
filter_complex = []
running_duration = 0
for i, clip in enumerate(kb_clips):
cmd.extend(["-i", clip['path']])
for i in range(len(kb_clips) - 1):
input1 = f"[{i}:v]" if i == 0 else f"[v{i-1}]"
input2 = f"[{i+1}:v]"
output = f"[v{i}]"
transition = random.choice(all_transitions) if transition_style == "Random" else transition_style.lower()
offset = running_duration + kb_clips[i]['duration'] - transition_duration
filter_complex.append(f"{input1}{input2}xfade=transition={transition}:duration={transition_duration}:offset={offset}{output}")
running_duration += kb_clips[i]['duration'] - transition_duration
cmd.extend(["-filter_complex", ";".join(filter_complex), "-map", f"[v{len(kb_clips)-2}]", "-c:v", "libx264", "-pix_fmt", "yuv420p", final_silent_path])
run_ffmpeg_command(cmd)
progress(0.98, desc="Muxing final audio...")
final_output_path = os.path.join(TEMP_DIR, f"slideshow_final_{timestamp}.mp4")
run_ffmpeg_command(["ffmpeg", "-y", "-i", final_silent_path, "-i", audio_path, "-c:v", "copy", "-c:a", "aac", "-shortest", final_output_path], "Muxing audio")
return final_output_path
finally:
if os.path.exists(job_temp_dir):
shutil.rmtree(job_temp_dir)
def slideshow_wrapper(images, audio_path, kb_effect_style, transition_style, rhythm_choice, out_w, out_h, progress=gr.Progress(track_tqdm=True)):
return _create_automated_slideshow_impl(images, audio_path, kb_effect_style, transition_style, rhythm_choice, out_w, out_h, progress)
def _create_rhythmic_remix_impl(video_path, audio_path, cut_style, beat_sync, resolution_choice, custom_w, custom_h, progress: Progress):
"""Internal implementation of the auto-rhythmic video remixer."""
if not video_path: raise gr.Error("Please upload a source video.")
if not audio_path: raise gr.Error("Please upload an audio track.")
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
job_temp_dir = os.path.join(TEMP_DIR, f"remix_{timestamp}")
os.makedirs(job_temp_dir, exist_ok=True)
try:
vf_filter = None
if resolution_choice == "Match Source Video Dimensions":
out_w, out_h = get_video_dimensions(video_path)
if out_w == 0 or out_h == 0:
gr.Warning("Could not read source video dimensions. Defaulting to 1080p.")
out_w, out_h = 1920, 1080
vf_filter = f"scale={out_w}:{out_h}:force_original_aspect_ratio=decrease,pad={out_w}:{out_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
elif resolution_choice == "1080p (1920x1080)":
out_w, out_h = 1920, 1080
elif resolution_choice == "720p (1280x720)":
out_w, out_h = 1280, 720
elif resolution_choice == "Custom":
out_w, out_h = int(custom_w), int(custom_h)
if out_w <= 0 or out_h <= 0:
raise gr.Error("Custom width and height must be positive numbers.")
out_w, out_h = out_w - (out_w % 2), out_h - (out_h % 2)
if vf_filter is None and resolution_choice != "Match Source Video Dimensions":
vf_filter = f"scale={out_w}:{out_h}:force_original_aspect_ratio=decrease,pad={out_w}:{out_h}:(ow-iw)/2:(oh-ih)/2,setsar=1"
progress(0, desc="Analyzing audio track for beats...")
try:
y, sr = librosa.load(audio_path)
audio_duration = librosa.get_duration(y=y, sr=sr)
_, beat_frames = librosa.beat.beat_track(y=y, sr=sr, units='frames')
beat_times = librosa.frames_to_time(beat_frames, sr=sr)
if len(beat_times) < 2: raise ValueError("Not enough beats were detected in the audio.")
except Exception as e:
raise gr.Error(f"Audio analysis failed: {e}")
beats_per_clip = {"On the Beat": 1, "Every 2 Beats": 2, "Every Measure (4 beats)": 4}[beat_sync]
clip_definitions = []
clip_start_beat_indices = range(0, len(beat_times), beats_per_clip)
for i, beat_index in enumerate(clip_start_beat_indices):
start_beat_time = beat_times[beat_index]
if i + 1 < len(clip_start_beat_indices):
end_beat_time = beat_times[clip_start_beat_indices[i+1]]
else:
end_beat_time = audio_duration
duration = end_beat_time - start_beat_time
if duration > 0.1:
clip_definitions.append({'duration': duration})
if not clip_definitions:
raise gr.Error("Could not define any video clips based on the detected beats.")
progress(0.1, desc="Planning video cuts...")
source_duration = get_media_duration(video_path)
current_time_in_source = 0
for clip in clip_definitions:
if cut_style == "Sequential":
clip['source_start'] = current_time_in_source
current_time_in_source += clip['duration']
if current_time_in_source > source_duration:
gr.Warning("Source video is shorter than the music. Looping video from the beginning.")
current_time_in_source = 0
elif cut_style == "Random Shuffle":
max_start_time = source_duration - clip['duration']
clip['source_start'] = random.uniform(0, max_start_time) if max_start_time > 0 else 0
extracted_clip_paths = []
for i, clip in enumerate(clip_definitions):
progress(0.1 + (0.7 * (i / len(clip_definitions))), desc=f"Extracting clip {i+1}/{len(clip_definitions)}...")
output_clip_path = os.path.join(job_temp_dir, f"clip_{i:04d}.mp4")
cmd = [
"ffmpeg", "-y",
"-ss", str(clip['source_start']),
"-i", video_path,
"-t", str(clip['duration']),
"-an",
"-c:v", "libx264", "-pix_fmt", "yuv420p"
]
if vf_filter:
cmd.extend(["-vf", vf_filter])
cmd.append(output_clip_path)
run_ffmpeg_command(cmd)
extracted_clip_paths.append(output_clip_path)
progress(0.85, desc="Stitching clips together...")
file_list_path = os.path.join(job_temp_dir, "files.txt")
with open(file_list_path, 'w', encoding='utf-8') as f:
for path in extracted_clip_paths:
f.write(f"file '{os.path.abspath(path)}'\n")
silent_final_path = os.path.join(job_temp_dir, "final_silent.mp4")
run_ffmpeg_command(["ffmpeg", "-y", "-f", "concat", "-safe", "0", "-i", file_list_path, "-c", "copy", silent_final_path], "Concatenating clips...")
progress(0.95, desc="Adding music...")
final_output_path = os.path.join(TEMP_DIR, f"remix_final_{timestamp}.mp4")
run_ffmpeg_command(["ffmpeg", "-y", "-i", silent_final_path, "-i", audio_path, "-c:v", "copy", "-c:a", "aac", "-map", "0:v:0", "-map", "1:a:0", "-shortest", final_output_path], "Muxing audio")
return final_output_path
finally:
if os.path.exists(job_temp_dir):
shutil.rmtree(job_temp_dir)
def rhythmic_remix_wrapper(video_path, audio_path, cut_style, beat_sync, resolution_choice, custom_w, custom_h, progress=gr.Progress(track_tqdm=True)):
return _create_rhythmic_remix_impl(video_path, audio_path, cut_style, beat_sync, resolution_choice, custom_w, custom_h, progress)
# --- BLING --- CSS AND JS ---
bling_css = """
@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;700&display=swap');
:root {
--bling-main-font: 'Inter', sans-serif;
--bling-gradient-start: #0f172a;
--bling-gradient-mid: #1e293b;
--bling-gradient-end: #334155;
--bling-accent-color: #38bdf8; /* sky-400 */
}
body, .gradio-container {
font-family: var(--bling-main-font) !important;
background: var(--bling-gradient-start);
background: linear-gradient(135deg, var(--bling-gradient-start) 0%, var(--bling-gradient-mid) 50%, var(--bling-gradient-end) 100%);
background-size: 200% 200%;
animation: gradient-animation 15s ease infinite;
}
@keyframes gradient-animation {
0% { background-position: 0% 50%; }
50% { background-position: 100% 50%; }
100% { background-position: 0% 50%; }
}
/* Glassmorphism for containers */
.gradio-tabs, .gradio-accordion, .gradio-group {
background: rgba(255, 255, 255, 0.05) !important;
border: 1px solid rgba(255, 255, 255, 0.1) !important;
border-radius: 12px !important;
backdrop-filter: blur(10px) !important;
-webkit-backdrop-filter: blur(10px) !important;
box-shadow: 0 4px 30px rgba(0, 0, 0, 0.1) !important;
}
/* Button Bling */
.gradio-button {
transition: all 0.2s ease-in-out !important;
box-shadow: 0 2px 4px rgba(0,0,0,0.2) !important;
}
.gradio-button:hover {
transform: translateY(-2px);
box-shadow: 0 4px 8px rgba(0,0,0,0.3) !important;
filter: brightness(1.1);
}
/* Custom Info/Warning Boxes */
.gradio-info {
background: rgba(56, 189, 248, 0.1) !important; /* sky-400 with alpha */
color: #f0f9ff !important; /* sky-50 */
border-left: 4px solid var(--bling-accent-color) !important;
border-radius: 8px !important;
}
.gradio-warning {
background: rgba(251, 191, 36, 0.1) !important; /* amber-400 with alpha */
color: #fffbeb !important; /* amber-50 */
border-left: 4px solid #fbbf24 !important;
border-radius: 8px !important;
}
/* Custom Scrollbars */
::-webkit-scrollbar { width: 8px; }
::-webkit-scrollbar-track { background: rgba(255, 255, 255, 0.1); }
::-webkit-scrollbar-thumb { background-color: var(--bling-accent-color); border-radius: 4px; }
::-webkit-scrollbar-thumb:hover { background-color: #0ea5e9; } /* sky-500 */
#custom-footer {
text-align: center !important;
padding: 20px 0 5px 0 !important;
font-size: .9em;
color: #94a3b8; /* slate-400 */
}
/* Loading Overlay CSS */
#loading-overlay {
position: fixed;
top: 0;
left: 0;
width: 100vw;
height: 100vh;
background-color: rgba(15, 23, 42, 0.8);
z-index: 10000;
display: flex;
justify-content: center;
align-items: center;
flex-direction: column;
color: white;
font-size: 1.2em;
backdrop-filter: blur(5px);
-webkit-backdrop-filter: blur(5px);
opacity: 0;
visibility: hidden;
transition: opacity 0.3s ease, visibility 0.3s ease;
}
#loading-overlay.visible {
opacity: 1;
visibility: visible;
}
.spinner {
width: 60px;
height: 60px;
border: 5px solid rgba(255, 255, 255, 0.3);
border-top-color: var(--bling-accent-color);
border-radius: 50%;
animation: spin 1s linear infinite;
margin-bottom: 20px;
}
@keyframes spin {
to { transform: rotate(360deg); }
}
"""
bling_js = """
() => {
// --- JKL Video Control ---
let activeVideo = null;
document.addEventListener('mouseover', (e) => {
if (e.target.tagName === 'VIDEO') {
activeVideo = e.target;
}
});
document.addEventListener('keydown', (e) => {
const activeElement = document.activeElement;
if (activeElement && (activeElement.tagName === 'INPUT' || activeElement.tagName === 'TEXTAREA')) {
return;
}
if (!activeVideo) return;
const frameTime = 1 / 30;
let handled = false;
switch (e.key.toLowerCase()) {
case 'k': activeVideo.paused ? activeVideo.play() : activeVideo.pause(); handled = true; break;
case 'j': activeVideo.currentTime = Math.max(0, activeVideo.currentTime - frameTime); handled = true; break;
case 'l': activeVideo.currentTime += frameTime; handled = true; break;
}
if (handled) e.preventDefault();
});
// --- Loading Overlay ---
function show_overlay(message = 'Processing... Please wait.') {
let overlay = document.getElementById('loading-overlay');
if (!overlay) {
overlay = document.createElement('div');
overlay.id = 'loading-overlay';
overlay.innerHTML = `<div class="spinner"></div><p id="loading-message"></p>`;
document.body.appendChild(overlay);
}
document.getElementById('loading-message').textContent = message;
overlay.classList.add('visible');
}
function hide_overlay() {
const overlay = document.getElementById('loading-overlay');
if (overlay) {
overlay.classList.remove('visible');
}
}
// --- Confetti ---
function fire_confetti() {
const a=document.createElement("script");a.setAttribute("src","https://cdn.jsdelivr.net/npm/canvas-confetti@1.9.2/dist/confetti.browser.min.js"),document.head.appendChild(a),a.onload=()=>{var e=confetti.create(null,{resize:!0,useWorker:!0});e({particleCount:150,spread:90,origin:{y:.6}})}
}
// --- Audio Feedback with Synthesized Whistle ---
const skriptz_audio = {
context: null,
isInitialized: false,
};
async function init_audio() {
if (skriptz_audio.isInitialized) return;
try {
skriptz_audio.context = new (window.AudioContext || window.webkitAudioContext)();
if (skriptz_audio.context.state === 'suspended') {
await skriptz_audio.context.resume();
}
} catch (e) {
console.error('Failed to initialize Web Audio API:', e);
}
skriptz_audio.isInitialized = true;
}
async function play_finish_sound() {
if (!skriptz_audio.isInitialized) {
await init_audio();
}
const context = skriptz_audio.context;
if (!context) return;
if (context.state === 'suspended') {
await context.resume();
}
const now = context.currentTime;
const delay = 0.2;
const startTime = now + delay;
const oscillator = context.createOscillator();
const gainNode = context.createGain();
oscillator.connect(gainNode);
gainNode.connect(context.destination);
oscillator.type = 'sine';
const startFreq = 2000;
const endFreq = 1000;
oscillator.frequency.setValueAtTime(startFreq, startTime);
oscillator.frequency.exponentialRampToValueAtTime(endFreq, startTime + 0.15);
gainNode.gain.setValueAtTime(0, startTime);
gainNode.gain.linearRampToValueAtTime(0.4, startTime + 0.02);
gainNode.gain.linearRampToValueAtTime(0, startTime + 0.15);
oscillator.start(startTime);
oscillator.stop(startTime + 0.2);
}
// --- Dynamic Page Title ---
function update_title(tab_name) {
if (tab_name) {
const clean_name = tab_name.replace(/[\\u{1F600}-\\u{1F64F}\\u{1F300}-\\u{1F5FF}\\u{1F680}-\\u{1F6FF}\\u{1F700}-\\u{1F77F}\\u{1F780}-\\u{1F7FF}\\u{1F800}-\\u{1F8FF}\\u{1F900}-\\u{1F9FF}\\u{1FA00}-\\u{1FA6F}\\u{1FA70}-\\u{1FAFF}\\u{2600}-\\u{26FF}\\u{2700}-\\u{27BF}]/gu, '').trim();
document.title = `Skriptz - ${clean_name}`;
} else {
document.title = "Skriptz - Universal Tool";
}
}
// --- Copy to Clipboard ---
function copy_to_clipboard(text_id) {
const text_area = document.getElementById(text_id).querySelector('textarea');
if(text_area) {
text_area.select();
document.execCommand('copy');
const original_button = this.event.target;
const original_text = original_button.innerText;
original_button.innerText = 'Copied!';
setTimeout(() => { original_button.innerText = original_text; }, 2000);
}
}
// --- Storyboard Time Getter ---
function storyboard_get_time(){
const e=document.querySelector('#storyboard_clip_preview video');
return e?e.currentTime:0
}
// Make functions globally accessible for Gradio
window.skriptz_bling = {
show_overlay,
hide_overlay,
fire_confetti,
play_finish_sound,
update_title,
copy_to_clipboard,
storyboard_get_time
};
}
"""
with gr.Blocks(
title="Skriptz - Universal Tool",
css=bling_css,
js=bling_js
) as demo:
gr.HTML("""
<div id="loading-overlay">
<div class="spinner"></div>
<p id="loading-message">Processing... Please wait.</p>
</div>
<script src="https://cdn.jsdelivr.net/npm/canvas-confetti@1.9.2/dist/confetti.browser.min.js"></script>
""", visible=False)
logo_b64 = get_image_as_base64("logo.png")
if logo_b64: gr.HTML(f"""<div style="display: flex; justify-content: center; align-items: center; text-align: center; margin-bottom: 20px;"><a href="https://linktr.ee/skylinkd" target="_blank" rel="noopener noreferrer"><img src="{logo_b64}" alt="Skriptz Banner" style="max-width: 100%; max-height: 100px; height: auto;"></a></div>""")
else: gr.Markdown("# Skriptz Universal Tool")
gr.Markdown("<h3 style='text-align: center;'>Your one-stop shop for video and image processing</h3>")
storyboard_get_time_js = "() => { return window.skriptz_bling.storyboard_get_time(); }"
show_overlay_js = "() => { window.skriptz_bling.show_overlay('Working hard... this may take a moment!'); }"
hide_overlay_js = "() => { window.skriptz_bling.hide_overlay(); }"
fire_confetti_and_sound_js = "() => { window.skriptz_bling.fire_confetti(); window.skriptz_bling.play_finish_sound(); }"
copy_transcription_js = "() => { window.skriptz_bling.copy_to_clipboard('transcription_textbox'); }"
# --- UNIFIED HEIGHT FOR MEDIA COMPONENTS ---
UNIFIED_HEIGHT = 440
with gr.Tabs(elem_id="main_tabs") as main_tabs:
with gr.TabItem("🎬 Storyboard & Animatic", elem_id="storyboard_tab"):
gr.Markdown("## Create Video Animatics from Images and Clips")
gr.Info("1. **Build:** Upload assets, click to add to timeline. 2. **Time:** Set durations, trim videos, or use Rhythmic Editing. 3. **Generate:** Create your final video.")
assets_state = gr.State([])
timeline_state = gr.State([])
selected_timeline_index_state = gr.State(-1)
with gr.Row(equal_height=False):
with gr.Column(scale=2, min_width=400):
with gr.Group():
gr.Markdown("### 1. Asset Bin")
assets_upload_btn = gr.File(label="Upload Images & Video Clips", file_count="multiple", file_types=["image", "video"])
asset_gallery = gr.Gallery(label="Click an asset to add it to the timeline", columns=4, object_fit="contain", height=400)
asset_preview_gallery = gr.Gallery(label="Video Asset Preview (First & Last Frame)", columns=2, height=240, object_fit="contain", interactive=False)
with gr.Row():
add_all_to_timeline_btn = gr.Button("⬇️ Add All to Timeline")
clear_assets_btn = gr.Button("🗑️ Clear Asset Bin")
with gr.Column(scale=3, min_width=600):
with gr.Group():
gr.Markdown("### 2. Timeline & Generation")
timeline_df = gr.DataFrame(headers=["#", "Asset", "Type", "Duration (s)"], datatype=["number", "str", "str", "number"], interactive=False, row_count=(10, "fixed"))
with gr.Row():
timeline_up_btn = gr.Button("⬆️ Move Up", interactive=False)
timeline_down_btn = gr.Button("⬇️ Move Down", interactive=False)
timeline_remove_btn = gr.Button("🗑️ Remove", interactive=False)
clear_timeline_btn = gr.Button("💥 Clear Timeline")
gr.Markdown("### 3. Output Settings")
keep_audio_checkbox = gr.Checkbox(label="Keep Original Audio from Video Clips", value=False, info="If checked, the Project Audio Track below will be ignored.")
animatic_audio = gr.Audio(label="Project Audio Track (Narration/Music)", type="filepath")
with gr.Row():
animatic_out_w = gr.Number(label="Output Width", value=1920)
animatic_out_h = gr.Number(label="Output Height", value=1080)
match_first_asset_btn = gr.Button("📏 Match First Asset's Resolution")
generate_animatic_btn = gr.Button("🎬 Generate Manual Animatic", variant="secondary")
with gr.Accordion("🎵 Rhythmic Editing (Beat Sync)", open=False):
gr.Info("This will override manual durations and re-time all clips to match the music's rhythm.")
with gr.Row():
analyze_bpm_btn = gr.Button("🥁 Analyze BPM")
bpm_display = gr.Textbox(label="Audio BPM", interactive=False)
measure_dropdown = gr.Dropdown(
["2 Measures", "1 Measure", "1/2 Measure", "1/4 Measure (Beat)"],
value="1 Measure", label="Cut Duration per Clip"
)
generate_rhythmic_btn = gr.Button("🎶 Generate Rhythmic Animatic", variant="primary")
animatic_output_video = gr.Video(label="Final Video Output", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
with gr.Column(scale=2, min_width=300):
with gr.Group():
gr.Markdown("### 4. Clip Properties")
gr.Info("Select a clip in the timeline table to edit it.")
clip_preview = gr.Video(label="Selected Clip Preview", interactive=True, elem_id="storyboard_clip_preview", height=UNIFIED_HEIGHT)
clip_duration_input = gr.Number(label="Set Duration (seconds)", interactive=True, precision=2)
update_clip_btn = gr.Button("🔄 Update Clip Duration")
with gr.Group(visible=False) as trim_group:
gr.Markdown("#### Video Trimming")
gr.Info("Use player (K=Play, J/L=Frame Step) to find a frame, then use buttons below.")
with gr.Row():
clip_start_time_input = gr.Number(label="Start Time (s)", precision=2)
clip_end_time_input = gr.Number(label="End Time (s)", precision=2)
with gr.Row():
set_clip_start_btn = gr.Button("Set START")
set_clip_end_btn = gr.Button("Set END")
apply_trim_btn = gr.Button("✂️ Apply Trim")
assets_upload_btn.upload(add_assets_to_bin, [assets_upload_btn, assets_state], [assets_state, asset_gallery])
asset_gallery.select(handle_asset_selection, [assets_state, timeline_state], [timeline_state, asset_preview_gallery])
add_all_to_timeline_btn.click(add_all_assets_to_timeline, [assets_state, timeline_state], timeline_state)
timeline_state.change(update_timeline_df, timeline_state, timeline_df)
timeline_state.change(lambda: (-1, None, None, gr.update(interactive=False), gr.update(interactive=False), gr.update(interactive=False), gr.update(visible=False), 0, 0),
outputs=[selected_timeline_index_state, clip_preview, clip_duration_input, timeline_up_btn, timeline_down_btn, timeline_remove_btn, trim_group, clip_start_time_input, clip_end_time_input])
timeline_df.select(
handle_timeline_selection,
timeline_state,
[selected_timeline_index_state, clip_preview, clip_duration_input, timeline_up_btn, timeline_down_btn, timeline_remove_btn, trim_group, clip_start_time_input, clip_end_time_input]
)
update_clip_btn.click(update_clip_properties, [timeline_state, selected_timeline_index_state, clip_duration_input], timeline_state)
apply_trim_btn.click(apply_trim_and_update, [timeline_state, selected_timeline_index_state, clip_start_time_input, clip_end_time_input], [timeline_state, clip_duration_input])
set_clip_start_btn.click(fn=None, js=storyboard_get_time_js, outputs=clip_start_time_input)
set_clip_end_btn.click(fn=None, js=storyboard_get_time_js, outputs=clip_end_time_input)
timeline_up_btn.click(handle_timeline_action, [timeline_state, selected_timeline_index_state, gr.State("up")], [timeline_state, timeline_df])
timeline_down_btn.click(handle_timeline_action, [timeline_state, selected_timeline_index_state, gr.State("down")], [timeline_state, timeline_df])
timeline_remove_btn.click(handle_timeline_action, [timeline_state, selected_timeline_index_state, gr.State("remove")], [timeline_state, timeline_df])
clear_assets_btn.click(lambda: ([], gr.update(value=None), None), outputs=[assets_state, asset_gallery, asset_preview_gallery])
clear_timeline_btn.click(lambda: [], outputs=[timeline_state])
keep_audio_checkbox.change(fn=lambda x: gr.update(interactive=not x), inputs=keep_audio_checkbox, outputs=animatic_audio)
generate_animatic_btn.click(fn=create_animatic, inputs=[timeline_state, animatic_audio, animatic_out_w, animatic_out_h, keep_audio_checkbox], outputs=animatic_output_video).then(fn=None, js=fire_confetti_and_sound_js)
analyze_bpm_btn.click(detect_bpm, animatic_audio, bpm_display)
generate_rhythmic_btn.click(fn=create_rhythmic_animatic, inputs=[timeline_state, animatic_audio, measure_dropdown, animatic_out_w, animatic_out_h], outputs=animatic_output_video).then(fn=None, js=fire_confetti_and_sound_js)
match_first_asset_btn.click(set_resolution_from_first_asset, timeline_state, [animatic_out_w, animatic_out_h])
with gr.TabItem("🎨 Creative Suite", elem_id="creative_tab"):
with gr.Tabs():
with gr.TabItem("🎬 Automated Slideshow"):
gr.Markdown("## Automated Rhythmic Slideshow Creator")
gr.Info("Turn a collection of images and a music track into a dynamic video with Ken Burns effects and transitions synced to the beat.")
with gr.Row():
with gr.Column(scale=2):
slideshow_input_images = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
slideshow_audio = gr.Audio(label="Upload Music Track", type="filepath")
with gr.Accordion("⚙️ Style & Timing", open=True):
slideshow_kb_effect = gr.Dropdown(["Subtle", "Standard", "Dynamic"], value="Standard", label="Ken Burns Effect Intensity")
slideshow_transition = gr.Dropdown(["Random", "Fade", "WipeLeft", "WipeRight", "Dissolve", "SlideLeft", "SlideRight"], value="Random", label="Transition Style")
slideshow_rhythm = gr.Dropdown(["1 Image per Beat", "1 Image every 2 Beats", "1 Image per Measure (4 Beats)"], value="1 Image every 2 Beats", label="Image Display Rhythm")
with gr.Row():
slideshow_out_w = gr.Number(label="Output Width", value=1920)
slideshow_out_h = gr.Number(label="Output Height", value=1080)
slideshow_generate_btn = gr.Button("🚀 Generate Slideshow", variant="primary")
with gr.Column(scale=3):
slideshow_output_video = gr.Video(label="Generated Slideshow Video", show_download_button=True, height=UNIFIED_HEIGHT)
slideshow_generate_btn.click(
fn=slideshow_wrapper,
inputs=[slideshow_input_images, slideshow_audio, slideshow_kb_effect, slideshow_transition, slideshow_rhythm, slideshow_out_w, slideshow_out_h],
outputs=slideshow_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎵 Auto Music Video"):
gr.Markdown("## Automatic Rhythmic Video Remixer")
gr.Info("Automatically cuts a source video to the beat of a music track, creating a dynamic music video.")
with gr.Row():
with gr.Column(scale=2):
remix_input_video = gr.Video(label="Upload Source Video", height=UNIFIED_HEIGHT)
remix_audio = gr.Audio(label="Upload Music Track", type="filepath")
with gr.Accordion("⚙️ Remix Settings", open=True):
remix_cut_style = gr.Radio(
["Sequential", "Random Shuffle"],
value="Random Shuffle",
label="Video Cut Style",
info="Sequential: Cuts the video from start to finish. Random Shuffle: Picks random moments from the source video."
)
remix_beat_sync = gr.Dropdown(
["On the Beat", "Every 2 Beats", "Every Measure (4 beats)"],
value="Every 2 Beats",
label="Cutting Frequency"
)
remix_resolution = gr.Radio(
["Match Source Video Dimensions", "1080p (1920x1080)", "720p (1280x720)", "Custom"],
value="1080p (1920x1080)",
label="Output Resolution"
)
with gr.Row(visible=False) as custom_res_row:
remix_custom_w = gr.Number(label="Custom Width", value=1920)
remix_custom_h = gr.Number(label="Custom Height", value=1080)
remix_generate_btn = gr.Button("🎤 Generate Music Video", variant="primary")
with gr.Column(scale=3):
remix_output_video = gr.Video(label="Generated Music Video", show_download_button=True, height=UNIFIED_HEIGHT)
remix_resolution.change(
fn=lambda choice: gr.update(visible=(choice == "Custom")),
inputs=remix_resolution,
outputs=custom_res_row
)
remix_generate_btn.click(
fn=rhythmic_remix_wrapper,
inputs=[remix_input_video, remix_audio, remix_cut_style, remix_beat_sync, remix_resolution, remix_custom_w, remix_custom_h],
outputs=remix_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("⚡ Gradual Speed Ramp"):
gr.Markdown("## Create a Smooth, Gradual Speed Ramp Effect")
gr.Info("Applies a smooth 'bullet-time' like effect to the entire video, slowing down to 50% speed in the middle and then ramping back up to normal speed. This uses frame interpolation for a high-quality result.")
with gr.Row():
with gr.Column(scale=2):
gradual_ramp_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
gradual_ramp_btn = gr.Button("⌛ Apply Gradual Ramp", variant="primary")
with gr.Column(scale=3):
gradual_ramp_output_video = gr.Video(label="Ramped Video Output", show_download_button=True, height=UNIFIED_HEIGHT)
gradual_ramp_btn.click(
fn=create_gradual_ramp_video,
inputs=[gradual_ramp_input_video],
outputs=gradual_ramp_output_video,
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎞️ Auto-Trailer"):
gr.Markdown("## Automatic Trailer Creator")
gr.Info("Upload a long video, and this tool will find the most action-packed moments to create a short, dynamic trailer.")
with gr.Row():
with gr.Column(scale=2):
trailer_input_video = gr.Video(label="Upload Source Video", height=UNIFIED_HEIGHT)
trailer_music = gr.Audio(label="Add Background Music (Optional)", type="filepath")
with gr.Accordion("⚙️ Trailer Settings", open=True):
trailer_total_duration = gr.Slider(10, 120, 30, step=5, label="Total Trailer Length (s)")
trailer_clip_duration = gr.Slider(1.0, 5.0, 2.0, step=0.5, label="Duration of Each Clip (s)")
trailer_analysis_method = gr.Dropdown(["Motion"], value="Motion", label="Scene Analysis Method", info="Currently only motion detection is supported.")
trailer_transition = gr.Dropdown(["None", "Fade", "WipeLeft", "WipeRight", "Dissolve", "SlideLeft", "SlideRight"], value="Fade", label="Transition Style")
with gr.Row():
trailer_out_w = gr.Number(label="Output Width", value=1920)
trailer_out_h = gr.Number(label="Output Height", value=1080)
trailer_generate_btn = gr.Button("🚀 Generate Trailer", variant="primary")
with gr.Column(scale=3):
trailer_output_video = gr.Video(label="Generated Trailer Video", show_download_button=True, height=UNIFIED_HEIGHT)
trailer_generate_btn.click(
fn=auto_trailer_wrapper,
inputs=[
trailer_input_video, trailer_total_duration, trailer_clip_duration,
trailer_analysis_method, trailer_transition, trailer_music,
trailer_out_w, trailer_out_h
],
outputs=trailer_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎵 Audio Waveform"):
gr.Markdown("## Generate & Overlay Audio Waveforms")
gr.Info("Upload a video with an audio track to generate a dynamic waveform visualization.")
with gr.Row():
with gr.Column():
waveform_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
with gr.Row():
waveform_style = gr.Dropdown(["line", "p2p", "point"], value="line", label="Waveform Style")
waveform_size = gr.Textbox(value="800x200", label="Size (WxH)")
with gr.Row():
waveform_position = gr.Dropdown(["Bottom", "Center", "Top"], value="Bottom", label="Position")
waveform_color = gr.ColorPicker(value="#38bdf8", label="Color")
waveform_btn = gr.Button("🎶 Generate Waveform Video", variant="primary")
with gr.Column():
waveform_output_video = gr.Video(label="Output Video with Waveform", show_download_button=True, height=UNIFIED_HEIGHT)
waveform_btn.click(
fn=generate_waveform_video,
inputs=[waveform_input_video, waveform_style, waveform_size, waveform_position, waveform_color],
outputs=waveform_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🖼️ Picture-in-Picture"):
gr.Markdown("## Create Picture-in-Picture (PiP) & Image Overlays")
gr.Info("Upload a main video and an overlay (video or image). The overlay will be placed on top of the main video.")
with gr.Row():
with gr.Column():
pip_main_video = gr.Video(label="Main Video (Background)", height=UNIFIED_HEIGHT)
pip_overlay_media = gr.File(label="Overlay Media (Video or Image)", file_types=["video", "image"])
with gr.Row():
pip_position = gr.Dropdown(
["Top-Left", "Top-Center", "Top-Right", "Center-Left", "Center", "Center-Right", "Bottom-Left", "Bottom-Center", "Bottom-Right"],
value="Bottom-Right", label="Position"
)
pip_scale = gr.Slider(0.01, 0.75, 0.25, step=0.01, label="Overlay Size")
pip_btn = gr.Button("✨ Create PiP Video", variant="primary")
with gr.Column():
pip_output_video = gr.Video(label="Output PiP Video", show_download_button=True, height=UNIFIED_HEIGHT)
pip_btn.click(
fn=create_pip_video,
inputs=[pip_main_video, pip_overlay_media, pip_position, pip_scale],
outputs=pip_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("😂 Meme Creator"):
gr.Markdown("## Classic Meme & Text Overlay Creator")
gr.Info("Upload an image and add text. Font availability depends on your operating system.")
with gr.Row():
with gr.Column():
meme_input_image = gr.Image(type="numpy", label="Input Image", height=UNIFIED_HEIGHT)
meme_text = gr.Textbox(label="Text", placeholder="Your witty text here...")
meme_position = gr.Radio(["Top", "Center", "Bottom"], value="Top", label="Text Position")
meme_font = gr.Dropdown(
["Impact", "Arial", "Arial Black", "Comic Sans MS", "Courier New", "Georgia", "Tahoma", "Times New Roman", "Trebuchet MS", "Verdana"],
value="Impact",
label="Font"
)
with gr.Row():
meme_text_color = gr.ColorPicker(value="#FFFFFF", label="Text Color")
meme_outline_color = gr.ColorPicker(value="#000000", label="Outline Color")
meme_font_size = gr.Slider(1, 10, 5, step=1, label="Relative Font Size")
meme_btn = gr.Button("😂 Generate Meme", variant="primary")
with gr.Column():
meme_output_image = gr.Image(label="Output Image", show_download_button=True, height=UNIFIED_HEIGHT)
meme_btn.click(
fn=create_meme,
inputs=[meme_input_image, meme_text, meme_position, meme_font, meme_font_size, meme_text_color, meme_outline_color],
outputs=meme_output_image
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🤖 FLUX.1 API"):
gr.Markdown("### Generate an image using `FLUX.1` models via Gradio Client.")
gr.Info("Requires a Hugging Face User Access Token.")
with gr.Row():
with gr.Column():
hf_token_input = gr.Textbox(label="HF Token", type="password", placeholder="Enter hf_... token")
flux_model_dropdown = gr.Dropdown(list(FLUX_MODELS.keys()), value="FLUX.1-schnell (Fast)", label="Select FLUX Model")
prompt_input_flux = gr.Textbox(label="Prompt", lines=3, placeholder="A cinematic photo...")
with gr.Row():
flux_width_slider = gr.Slider(256, 2048, 1024, step=64, label="Width")
flux_height_slider = gr.Slider(256, 2048, 1024, step=64, label="Height")
flux_btn = gr.Button("☄️ Generate Image", variant="primary")
with gr.Column():
output_image_flux = gr.Image(label="Generated Image", show_download_button=True, height=UNIFIED_HEIGHT)
flux_btn.click(call_flux_api, [prompt_input_flux, flux_model_dropdown, flux_width_slider, flux_height_slider, hf_token_input], output_image_flux).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🖼️ Image Utilities", elem_id="image_tab"):
with gr.Tabs():
with gr.TabItem("✨ Manipulate"):
gr.Markdown("### Simple Image Manipulation")
gr.Info("Apply a single transformation like inverting colors, flipping, or rotating.")
with gr.Row():
with gr.Column():
manip_input_image = gr.Image(type="numpy", label="Input Image", height=UNIFIED_HEIGHT)
manip_operation_radio = gr.Radio(
["Invert Colors", "Flip Horizontal", "Flip Vertical", "Rotate 90° Right", "Rotate 90° Left"],
label="Select Operation", value="Invert Colors"
)
manip_apply_btn = gr.Button("🎨 Apply Manipulation", variant="primary")
with gr.Column():
manip_output_image = gr.Image(label="Output Image", show_download_button=True, height=UNIFIED_HEIGHT)
manip_apply_btn.click(fn=manipulate_image, inputs=[manip_input_image, manip_operation_radio], outputs=manip_output_image).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("✂️ Batch Cropper"):
gr.Markdown("### Crop a batch of images to the same dimensions.")
gr.Info("Upload images, and the first image will be used as a preview. Adjust the sliders to see the crop area, then process the whole batch.")
crop_original_preview_state = gr.State()
with gr.Row():
with gr.Column(scale=1):
crop_input_images = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
gr.Markdown("#### Crop Box Settings")
crop_box_w = gr.Slider(label="Width", minimum=64, maximum=4096, step=8, value=512)
crop_box_h = gr.Slider(label="Height", minimum=64, maximum=4096, step=8, value=512)
crop_box_x = gr.Slider(label="X Offset", minimum=0, maximum=4096, step=8, value=0)
crop_box_y = gr.Slider(label="Y Offset", minimum=0, maximum=4096, step=8, value=0)
crop_btn = gr.Button("✂️ Crop All Images", variant="primary")
with gr.Column(scale=2):
crop_preview_image_display = gr.Image(label="Crop Preview (on first image)", type="pil", interactive=False, height=UNIFIED_HEIGHT)
crop_output_gallery = gr.Gallery(label="Cropped Images Preview", columns=4, object_fit="contain", height="auto")
crop_output_zip = gr.File(label="Download All as .zip", interactive=False)
def setup_crop_preview(files):
if not files:
return None, None, gr.update(), gr.update(), gr.update(), gr.update()
first_image_path = files[0].name
try:
img = Image.open(first_image_path).convert("RGB")
w, h = img.size
return img, img, gr.update(maximum=w, value=min(512, w)), gr.update(maximum=h, value=min(512, h)), gr.update(maximum=w), gr.update(maximum=h)
except Exception as e:
gr.Warning(f"Could not load preview image: {e}")
return None, None, gr.update(), gr.update(), gr.update(), gr.update()
crop_input_images.upload(
fn=setup_crop_preview,
inputs=[crop_input_images],
outputs=[crop_preview_image_display, crop_original_preview_state, crop_box_w, crop_box_h, crop_box_x, crop_box_y]
)
crop_sliders = [crop_box_x, crop_box_y, crop_box_w, crop_box_h]
for slider in crop_sliders:
slider.release(
fn=update_crop_preview,
inputs=[crop_original_preview_state] + crop_sliders,
outputs=crop_preview_image_display
)
crop_btn.click(
fn=batch_crop_images,
inputs=[crop_input_images] + crop_sliders,
outputs=[crop_output_gallery, crop_output_zip]
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🧩 Collage Maker"):
gr.Markdown("### Create a collage from multiple images.")
with gr.Row():
with gr.Column():
collage_input_images = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
collage_layout = gr.Radio(["Grid", "Horizontal", "Vertical"], value="Grid", label="Layout")
collage_w = gr.Slider(label="Collage Width", minimum=256, maximum=4096, value=1920, step=64)
collage_h = gr.Slider(label="Collage Height", minimum=256, maximum=4096, value=1080, step=64)
collage_bg_color = gr.ColorPicker(value="#000000", label="Background Color")
collage_btn = gr.Button("🎨 Create Collage", variant="primary")
with gr.Column():
collage_output_image = gr.Image(label="Output Collage", show_download_button=True, height=UNIFIED_HEIGHT)
collage_btn.click(
fn=create_collage,
inputs=[collage_input_images, collage_layout, collage_w, collage_h, collage_bg_color],
outputs=collage_output_image
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("📸 Duo-Stitcher"):
gr.Markdown("### Smart Image Stitcher")
gr.Info("Upload two vertical images to join them side-by-side, or two horizontal images to stack them. The result is always a square.")
with gr.Row():
with gr.Column():
stitch_img1 = gr.Image(type="numpy", label="Image 1 (Left/Top)", height=UNIFIED_HEIGHT)
stitch_img2 = gr.Image(type="numpy", label="Image 2 (Right/Bottom)", height=UNIFIED_HEIGHT)
with gr.Row():
stitch_size = gr.Slider(512, 4096, 1024, step=128, label="Output Size (pixels)")
stitch_bg_color = gr.ColorPicker(value="#000000", label="Background Color")
stitch_btn = gr.Button("🧩 Stitch Images", variant="primary")
with gr.Column():
stitch_output_image = gr.Image(label="Stitched Output Image", show_download_button=True, height=UNIFIED_HEIGHT)
stitch_btn.click(
fn=stitch_images_smartly,
inputs=[stitch_img1, stitch_img2, stitch_size, stitch_bg_color],
outputs=stitch_output_image
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("📹 Image to Video"):
gr.Markdown("### Create a short, looping video from a single static image.")
with gr.Row():
with gr.Column():
input_image_i2v = gr.Image(type="numpy", label="Input Image", height=UNIFIED_HEIGHT)
duration_slider_i2v = gr.Slider(1, 30, 5, step=0.1, label="Duration (s)")
input_audio_i2v = gr.Audio(label="Add Music (Optional)", type="filepath")
compile_i2v_btn = gr.Button("🎬 Create Looping Video", variant="primary")
with gr.Column():
output_video_i2v = gr.Video(label="Output Looping Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
compile_i2v_btn.click(image_to_looping_video, [input_image_i2v, duration_slider_i2v, input_audio_i2v], output_video_i2v).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🔎 Zoom Video"):
gr.Markdown("### Create a 'Ken Burns' style zoom/pan video from an image.")
gr.Info("Upload one or more images. The output will be a gallery of videos, or a single combined video if you check the box.")
with gr.Row():
with gr.Column():
i2zv_input_images = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
i2zv_duration = gr.Slider(1, 30, 5, step=0.5, label="Video Duration (s) per Image")
i2zv_zoom_ratio = gr.Slider(1.0, 2.0, 1.25, step=0.05, label="Zoom Ratio")
i2zv_zoom_dir = gr.Dropdown(
["Center", "Top", "Bottom", "Left", "Right", "Top-Left", "Top-Right", "Bottom-Left", "Bottom-Right"],
value="Center", label="Zoom Direction"
)
i2zv_combine = gr.Checkbox(label="Combine all videos into one", value=False)
i2zv_audio = gr.Audio(label="Add Music (Optional)", type="filepath")
i2zv_btn = gr.Button("✨ Create Zoom Video(s)", variant="primary")
with gr.Column():
i2zv_output_gallery = gr.Gallery(label="Output Video Previews", columns=2, object_fit="contain", visible=True)
i2zv_output_video = gr.Video(label="Combined Output Video", interactive=True, visible=False, show_download_button=True, height=UNIFIED_HEIGHT)
i2zv_output_zip = gr.File(label="Download All as .zip", interactive=False)
i2zv_combine.change(
fn=lambda x: [gr.update(visible=not x), gr.update(visible=x)],
inputs=i2zv_combine,
outputs=[i2zv_output_gallery, i2zv_output_video]
)
i2zv_btn.click(
fn=create_zoom_videos,
inputs=[i2zv_input_images, i2zv_duration, i2zv_zoom_ratio, i2zv_zoom_dir, i2zv_combine, i2zv_audio],
outputs=[i2zv_output_gallery, i2zv_output_video, i2zv_output_zip]
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("✂️ Batch BG Remover"):
gr.Markdown("### Remove the background from a batch of images.")
with gr.Row():
with gr.Column():
input_images_bg = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
remove_bg_btn = gr.Button("🧼 Remove Backgrounds", variant="primary")
with gr.Column():
output_gallery_bg = gr.Gallery(label="Images with Transparent Background", columns=4, object_fit="contain", height="auto")
output_zip_bg = gr.File(label="Download All as .zip", interactive=False)
remove_bg_btn.click(remove_background_batch, input_images_bg, [output_gallery_bg, output_zip_bg]).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🖋️ Batch Watermarker"):
gr.Markdown("### Apply a text watermark to a batch of images.")
with gr.Row():
with gr.Column():
input_images_wm = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
watermark_text = gr.Textbox(label="Watermark Text", placeholder="(c) My Awesome Project")
watermark_pos = gr.Radio(["Top-Left", "Top-Right", "Bottom-Left", "Bottom-Right", "Center"], value="Bottom-Right", label="Position")
watermark_opacity = gr.Slider(0, 100, 50, step=1, label="Opacity (%)")
watermark_btn = gr.Button("✍️ Apply Watermarks", variant="primary")
with gr.Column():
output_gallery_wm = gr.Gallery(label="Watermarked Images", columns=4, object_fit="contain", height="auto")
output_zip_wm = gr.File(label="Download All as .zip", interactive=False)
watermark_btn.click(apply_watermark_batch, [input_images_wm, watermark_text, watermark_pos, watermark_opacity], [output_gallery_wm, output_zip_wm]).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("📏 Batch Resizer & Converter"):
gr.Markdown("### Convert, resize, and compress a batch of images.")
gr.Info("Choose a preset for quick resizing or select 'Custom' to enter your own dimensions.")
with gr.Row():
with gr.Column():
brc_input_images = gr.File(label="Upload Images", file_count="multiple", file_types=["image"])
with gr.Accordion("⚙️ Output Settings", open=True):
brc_preset = gr.Dropdown(
label="Resolution Presets",
choices=[
"1024x1024 (Square)", "768x768 (Square)", "720x720 (Square)",
"1280x720 (Landscape 16:9)", "720x1280 (Portrait 9:16)",
"768x512 (Landscape 3:2)", "512x768 (Portrait 2:3)",
"Custom"
],
value="1024x1024 (Square)"
)
with gr.Row(visible=False) as brc_custom_size_row:
brc_max_w = gr.Number(label="Width", value=1024)
brc_max_h = gr.Number(label="Height", value=1024)
brc_resize_mode = gr.Radio(["Fit (preserve aspect ratio)", "Stretch to Fit"], value="Fit (preserve aspect ratio)", label="Resize Mode")
brc_format = gr.Dropdown(["JPG", "PNG", "WEBP"], value="JPG", label="Output Format")
brc_quality = gr.Slider(1, 100, 90, step=1, label="JPG/WEBP Quality", interactive=True)
brc_btn = gr.Button("🚀 Process Images", variant="primary")
with gr.Column():
brc_output_gallery = gr.Gallery(label="Processed Images Preview", columns=4, object_fit="contain", height="auto")
brc_output_zip = gr.File(label="Download All as .zip", interactive=False)
def update_resizer_from_preset(preset_str):
if preset_str == "Custom":
return gr.update(visible=True), gr.update(), gr.update()
match = re.search(r'(\d+)\s*x\s*(\d+)', preset_str)
if match:
w, h = int(match.group(1)), int(match.group(2))
return gr.update(visible=False), w, h
# Fallback to a default if parsing fails for some reason
return gr.update(visible=False), 1024, 1024
brc_preset.change(
fn=update_resizer_from_preset,
inputs=brc_preset,
outputs=[brc_custom_size_row, brc_max_w, brc_max_h]
)
# Initialize the values on load
demo.load(
fn=update_resizer_from_preset,
inputs=brc_preset,
outputs=[brc_custom_size_row, brc_max_w, brc_max_h]
)
brc_format.change(lambda f: gr.update(visible=f in ["JPG", "WEBP"]), brc_format, brc_quality)
brc_btn.click(
# The 'enable_resize' parameter is now implicitly True
lambda files, out_f, qual, w, h, mode: batch_resize_convert_images(files, out_f, qual, True, w, h, mode),
[brc_input_images, brc_format, brc_quality, brc_max_w, brc_max_h, brc_resize_mode],
[brc_output_gallery, brc_output_zip]
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎥 Video Utilities", elem_id="video_tab"):
gr.Markdown("## A collection of useful video tools.")
with gr.Tabs():
with gr.TabItem("🎞️ Frame Tools"):
with gr.Tabs():
with gr.TabItem("First & Last"):
gr.Markdown("### Extract the very first and very last frames of a video.")
with gr.Row():
with gr.Column():
input_video_fl = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
extract_fl_btn = gr.Button("🎬 Extract Frames", variant="primary")
with gr.Column():
output_gallery_fl = gr.Gallery(label="Output Frames (First, Last)", columns=2, object_fit="contain", height="auto")
extract_fl_btn.click(fn=extract_first_last_frame, inputs=input_video_fl, outputs=output_gallery_fl).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Batch First & Last"):
gr.Markdown("### Extract the first and last frames from multiple videos.")
with gr.Row():
with gr.Column():
batch_fl_input_videos = gr.File(label="Upload Videos", file_count="multiple", file_types=["video"])
batch_fl_process_btn = gr.Button("🎬 Extract All Frames", variant="primary")
with gr.Column():
batch_fl_output_gallery = gr.Gallery(label="Extracted Frames Preview", columns=6, object_fit="contain", height="auto")
batch_fl_output_zip = gr.File(label="Download All Frames (.zip)", interactive=False)
batch_fl_process_btn.click(
fn=batch_extract_first_last_frames,
inputs=batch_fl_input_videos,
outputs=[batch_fl_output_gallery, batch_fl_output_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Video to Frames"):
gr.Markdown("### Extract all individual frames from a video file.")
with gr.Row():
with gr.Column():
input_video_v2f = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
v2f_fps_display = gr.Textbox(label="Detected FPS", interactive=False, value="N/A")
with gr.Accordion("⚙️ Advanced Options", open=False):
v2f_skip_rate = gr.Slider(1, 30, 1, step=1, label="Extract Every Nth Frame")
v2f_rotation = gr.Dropdown(["None", "90 Degrees Clockwise", "90 Degrees Counter-Clockwise", "180 Degrees"], value="None", label="Rotation")
v2f_format = gr.Radio(["PNG", "JPG"], value="PNG", label="Output Format")
v2f_jpg_quality = gr.Slider(1, 100, 95, step=1, label="JPG Quality", interactive=False)
v2f_resize = gr.Checkbox(label="Resize all extracted frames", value=False)
with gr.Row():
v2f_width = gr.Number(label="Output Width", value=1024, interactive=False)
v2f_height = gr.Number(label="Output Height", value=576, interactive=False)
extract_v2f_btn = gr.Button("🪚 Extract All Frames", variant="primary")
with gr.Column():
output_gallery_v2f = gr.Gallery(label="Extracted Frames Preview (max 100 shown)", columns=8, object_fit="contain", height="auto")
output_zip_v2f = gr.File(label="Download All Frames (.zip)", interactive=False)
input_video_v2f.upload(lambda v: f"{get_video_fps(v):.2f} FPS", input_video_v2f, v2f_fps_display)
v2f_resize.change(lambda x: [gr.update(interactive=x), gr.update(interactive=x)], v2f_resize, [v2f_width, v2f_height])
v2f_format.change(lambda x: gr.update(interactive=(x=="JPG")), v2f_format, v2f_jpg_quality)
extract_v2f_btn.click(video_to_frames_extractor, [input_video_v2f, v2f_skip_rate, v2f_rotation, v2f_resize, v2f_width, v2f_height, v2f_format, v2f_jpg_quality], [output_gallery_v2f, output_zip_v2f]).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Frames to Video"):
gr.Markdown("### Compile a sequence of image frames into a video file.")
with gr.Row():
with gr.Column():
input_frames_f2v = gr.File(label="Upload Frames", file_count="multiple", file_types=["image"])
fps_slider_f2v = gr.Slider(1, 60, 24, step=1, label="FPS")
with gr.Accordion("⚙️ Advanced Options", open=False):
f2v_rotation = gr.Dropdown(["None", "90 Degrees Clockwise", "90 Degrees Counter-Clockwise", "180 Degrees"], value="None", label="Rotation")
f2v_resize = gr.Checkbox(label="Resize all frames", value=False)
with gr.Row():
f2v_width = gr.Number(label="Output Width", value=1024, interactive=False)
f2v_height = gr.Number(label="Output Height", value=576, interactive=False)
compile_f2v_btn = gr.Button("📽️ Create Video", variant="primary")
with gr.Column():
output_video_f2v = gr.Video(label="Compiled Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
f2v_resize.change(lambda x: [gr.update(interactive=x), gr.update(interactive=x)], f2v_resize, [f2v_width, f2v_height])
compile_f2v_btn.click(create_video_from_frames, [input_frames_f2v, fps_slider_f2v, f2v_rotation, f2v_resize, f2v_width, f2v_height], output_video_f2v).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("✂️ Editing"):
with gr.Tabs():
with gr.TabItem("🎞️ Slo-Mo & Enhance"):
gr.Markdown("## AI Slow-Motion")
gr.Info(
"Create super smooth slow-motion videos. The 'AI-Enhanced' method uses frame interpolation for the best quality."
)
if not ENHANCE_AI_AVAILABLE:
gr.Warning(
"AI models not found. The 'AI-Enhanced' option is disabled. Please install 'rife-ncnn-vulkan-python' to enable it."
)
with gr.Row():
with gr.Column():
slowmo_input_videos = gr.File(label="Upload Videos", file_count="multiple", file_types=["video"])
with gr.Accordion("⚙️ Settings", open=True):
slowmo_factor = gr.Dropdown(["2x", "4x", "8x"], value="2x", label="Slowdown Factor")
method_choices = ["Standard (Fast)"]
if ENHANCE_AI_AVAILABLE:
method_choices.append("AI-Enhanced (High Quality)")
slowmo_method = gr.Radio(
method_choices,
value=method_choices[-1] if ENHANCE_AI_AVAILABLE else method_choices[0],
label="Interpolation Method"
)
slowmo_btn = gr.Button("🚀 Process Videos", variant="primary")
with gr.Column():
slowmo_output_gallery = gr.Gallery(label="Output Video Previews", columns=1, object_fit="contain")
slowmo_output_zip = gr.File(label="Download All as .zip", interactive=False)
slowmo_btn.click(
fn=batch_slowmo_enhance_videos,
inputs=[slowmo_input_videos, slowmo_factor, slowmo_method],
outputs=[slowmo_output_gallery, slowmo_output_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🔪 Auto Jump-Cut"):
gr.Markdown("### Automatic Silence Remover (Jump-Cutter)")
gr.Info("1. Upload a video. A waveform will appear. 2. Adjust the 'Silence Threshold' slider so it's above the thin 'noise floor' line but below your voice peaks. 3. Process the video!")
with gr.Row():
with gr.Column(scale=2):
jumpcut_input_video = gr.Video(label="Input Video (with audio)", height=UNIFIED_HEIGHT)
jc_waveform_preview = gr.Image(label="Audio Waveform Preview", interactive=False)
with gr.Accordion("⚙️ Cut Settings", open=True):
jumpcut_threshold = gr.Slider(minimum=-60, maximum=-20, value=-30, step=1, label="Silence Threshold (dB)", info="Anything quieter than this is 'silence'. Lower values are more strict.")
jumpcut_duration = gr.Slider(minimum=0.1, maximum=2.0, value=0.5, step=0.1, label="Minimum Silence Duration (s)", info="Silences shorter than this will be ignored.")
with gr.Accordion("📐 Output Resolution", open=False):
jc_resolution = gr.Radio(
["Keep Original", "1080p (1920x1080)", "Portrait (1080x1920)", "Custom"],
value="Keep Original", label="Output Resolution"
)
with gr.Row(visible=False) as jc_custom_res_row:
jc_custom_w = gr.Number(label="Custom Width", value=1920)
jc_custom_h = gr.Number(label="Custom Height", value=1080)
jumpcut_btn = gr.Button("🔪 Perform Jump-Cut", variant="primary")
with gr.Column(scale=3):
jumpcut_output_video = gr.Video(label="Edited Video Output", show_download_button=True, height=UNIFIED_HEIGHT)
jumpcut_input_video.upload(fn=generate_waveform_preview, inputs=jumpcut_input_video, outputs=jc_waveform_preview)
jc_resolution.change(
fn=lambda choice: gr.update(visible=(choice == "Custom")),
inputs=jc_resolution,
outputs=jc_custom_res_row
)
jumpcut_btn.click(
fn=auto_jump_cut,
inputs=[jumpcut_input_video, jumpcut_threshold, jumpcut_duration, jc_resolution, jc_custom_w, jc_custom_h],
outputs=jumpcut_output_video,
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
# NEW FEATURE: Video Silence Chopper
with gr.TabItem("✂️ Video Silence Chopper"):
gr.Markdown("### Automatic Video Silence Chopper")
gr.Info("Splits a video into multiple clips, removing the silent parts. Ideal for isolating spoken phrases from interviews or lectures.")
with gr.Row():
with gr.Column(scale=2):
video_chopper_input = gr.Video(label="Input Video (with audio)", height=UNIFIED_HEIGHT)
with gr.Accordion("⚙️ Silence Settings", open=True):
video_chopper_threshold = gr.Slider(minimum=-60, maximum=-20, value=-30, step=1, label="Silence Threshold (dB)")
video_chopper_duration = gr.Slider(minimum=0.1, maximum=2.0, value=0.5, step=0.1, label="Minimum Silence Duration (s)")
with gr.Accordion("📐 Output Resolution", open=False):
video_chopper_resolution = gr.Radio(["Keep Original", "1080p (1920x1080)", "Portrait (1080x1920)", "Custom"], value="Keep Original", label="Output Resolution")
with gr.Row(visible=False) as vc_custom_res_row:
vc_custom_w = gr.Number(label="Custom Width", value=1920)
vc_custom_h = gr.Number(label="Custom Height", value=1080)
video_chopper_btn = gr.Button("✂️ Chop Video into Clips", variant="primary")
with gr.Column(scale=3):
video_chopper_gallery = gr.Gallery(label="Chopped Video Clips (Preview)", columns=2, object_fit="contain", height="auto")
video_chopper_zip = gr.File(label="Download All Clips as .zip", interactive=False)
video_chopper_resolution.change(
fn=lambda choice: gr.update(visible=(choice == "Custom")),
inputs=video_chopper_resolution,
outputs=vc_custom_res_row
)
video_chopper_btn.click(
fn=chop_video_on_silence,
inputs=[video_chopper_input, video_chopper_threshold, video_chopper_duration, video_chopper_resolution, vc_custom_w, vc_custom_h],
outputs=[video_chopper_gallery, video_chopper_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("▦ Video Grid"):
gr.Markdown("### Video Grid Compiler")
gr.Info("Combine multiple videos into a single-screen layout. Videos will be scaled to fit within their grid tile.")
with gr.Row():
with gr.Column():
grid_input_videos = gr.File(label="Upload Videos", file_count="multiple", file_types=["video"])
grid_layout = gr.Dropdown(
["2x1 (Side-by-Side)", "1x2 (Stacked)", "2x2 (Quad-View)", "4x4 (16-View)", "8x4 (32-View)"],
value="2x1 (Side-by-Side)",
label="Grid Layout"
)
grid_w = gr.Slider(label="Output Width", minimum=256, maximum=4096, value=1920, step=64)
grid_h = gr.Slider(label="Output Height", minimum=256, maximum=4096, value=1080, step=64)
grid_bg_color = gr.ColorPicker(value="#000000", label="Background Color")
with gr.Accordion("Audio Settings", open=True):
grid_audio_choice = gr.Dropdown(
choices=["From Video 1", "None"],
value="From Video 1",
label="Video Audio Source",
info="Select which video's audio to use, or None."
)
grid_music_track = gr.Audio(label="Add External Music (Overrides above)", type="filepath")
grid_btn = gr.Button("▦ Compile Grid", variant="primary")
with gr.Column():
grid_output_video = gr.Video(label="Grid Video Output", show_download_button=True, height=UNIFIED_HEIGHT)
grid_input_videos.upload(update_audio_source_choices_for_grid, grid_input_videos, grid_audio_choice)
grid_btn.click(
fn=compile_video_grid,
inputs=[grid_input_videos, grid_layout, grid_w, grid_h, grid_bg_color, grid_audio_choice, grid_music_track],
outputs=grid_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Manipulate"):
gr.Markdown("### Simple Video Manipulation")
gr.Info("Apply a single transformation like inverting colors, flipping, or rotating to every frame of a video.")
with gr.Row():
with gr.Column():
vmanip_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
vmanip_operation_radio = gr.Radio(
["Invert Colors", "Flip Horizontal", "Flip Vertical", "Rotate 90° Right", "Rotate 90° Left"],
label="Select Operation", value="Invert Colors"
)
vmanip_apply_btn = gr.Button("✨ Apply Manipulation", variant="primary")
with gr.Column():
vmanip_output_video = gr.Video(label="Output Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
vmanip_apply_btn.click(fn=manipulate_video, inputs=[vmanip_input_video, vmanip_operation_radio], outputs=vmanip_output_video).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Ping-Pong"):
gr.Markdown("### Create a forward-then-reverse video loop (Boomerang).")
with gr.Row():
with gr.Column():
input_video_pingpong = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
audio_option_pingpong = gr.Radio(["Remove Audio", "Original Audio Only", "Reverse Audio"], value="Remove Audio", label="Audio Handling")
pingpong_btn = gr.Button("🏓 Create Ping-Pong Video", variant="primary")
with gr.Column():
output_video_pingpong = gr.Video(label="Ping-Pong Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
pingpong_btn.click(fn=ping_pong_video, inputs=[input_video_pingpong, audio_option_pingpong], outputs=output_video_pingpong).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Reverse"):
gr.Markdown("### Reverse a video clip.")
with gr.Row():
with gr.Column():
input_video_reverse = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
audio_option_reverse = gr.Radio(["Remove Audio", "Reverse Audio"], value="Remove Audio", label="Audio Handling")
reverse_btn = gr.Button("🔄 Reverse Video", variant="primary")
with gr.Column():
output_video_reverse = gr.Video(label="Reversed Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
reverse_btn.click(fn=reverse_video, inputs=[input_video_reverse, audio_option_reverse], outputs=output_video_reverse).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Merger"):
gr.Markdown("### Simple Video Merger")
gr.Info("Upload two or more video clips to join them together in sequence. All clips will be conformed to the resolution and framerate of the first video.")
with gr.Row():
with gr.Column():
merger_input_videos = gr.File(label="Upload Videos (2 or more)", file_count="multiple", file_types=["video"])
merger_btn = gr.Button("🔗 Merge Videos", variant="primary")
with gr.Column():
merger_output_video = gr.Video(label="Merged Video", show_download_button=True, height=UNIFIED_HEIGHT)
merger_btn.click(
fn=merge_videos,
inputs=merger_input_videos,
outputs=merger_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Stabilizer"):
gr.Markdown("### Smooth Out Shaky Video Footage")
gr.Info("This tool uses a two-pass process to analyze and stabilize shaky videos. Higher 'Shakiness' values are for very unstable footage. Higher 'Smoothing' creates a more fluid, gliding look but can introduce slight cropping/warping.")
with gr.Row():
with gr.Column():
stab_input_video = gr.Video(label="Input Shaky Video", height=UNIFIED_HEIGHT)
stab_shakiness = gr.Slider(minimum=1, maximum=10, value=5, step=1, label="Shakiness Level", info="How shaky is the source video? (1=low, 10=high)")
stab_smoothing = gr.Slider(minimum=1, maximum=50, value=10, step=1, label="Smoothing Amount", info="How much smoothing to apply. (More = smoother but might feel 'floaty')")
stab_btn = gr.Button("✨ Stabilize Video", variant="primary")
with gr.Column():
stab_output_video = gr.Video(label="Stabilized Output Video", show_download_button=True, height=UNIFIED_HEIGHT)
stab_btn.click(
fn=stabilize_video,
inputs=[stab_input_video, stab_shakiness, stab_smoothing],
outputs=stab_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Trimmer"):
gr.Markdown("### Visually trim a video. Use the player to find a frame, then set it as the start or end point.")
gr.Info("Keyboard hotkeys enabled: K = Play/Pause, J = Back 1 Frame, L = Forward 1 Frame (hover mouse over video)")
with gr.Row():
with gr.Column(scale=2):
input_video_trim = gr.Video(label="Input Video", elem_id="video-trim-input", height=UNIFIED_HEIGHT)
with gr.Row():
set_start_btn = gr.Button("Set Current Frame as START")
set_end_btn = gr.Button("Set Current Frame as END")
trim_btn = gr.Button("✂️ Trim Video", variant="primary")
with gr.Column(scale=1):
gr.Markdown("#### Trim Points")
start_frame_img = gr.Image(label="Start Frame", interactive=False)
trim_start_time_display = gr.Textbox(label="Start Time (s)", interactive=False)
end_frame_img = gr.Image(label="End Frame", interactive=False)
trim_end_time_display = gr.Textbox(label="End Time (s)", interactive=False)
trim_start_time = gr.Number(value=0, visible=False)
trim_end_time = gr.Number(value=0, visible=False)
with gr.Row():
output_video_trim = gr.Video(label="Trimmed Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
get_current_time_js = """()=>{const e=document.querySelector("#video-trim-input video");return e?e.currentTime:0}"""
def get_frame_from_time_wrapper(v,t): return get_frame_at_time(v,t), f"{t:.3f}"
set_start_btn.click(fn=None, js=get_current_time_js, outputs=[trim_start_time])
set_end_btn.click(fn=None, js=get_current_time_js, outputs=[trim_end_time])
trim_start_time.change(fn=get_frame_from_time_wrapper, inputs=[input_video_trim, trim_start_time], outputs=[start_frame_img, trim_start_time_display])
trim_end_time.change(fn=get_frame_from_time_wrapper, inputs=[input_video_trim, trim_end_time], outputs=[end_frame_img, trim_end_time_display])
trim_btn.click(fn=trim_video, inputs=[input_video_trim, trim_start_time, trim_end_time], outputs=output_video_trim).then(fn=None, js=fire_confetti_and_sound_js)
input_video_trim.clear(fn=lambda: (None, "0.00", None, "0.00", 0, 0), outputs=[start_frame_img, trim_start_time_display, end_frame_img, trim_end_time_display, trim_start_time, trim_end_time])
with gr.TabItem("Crop & Resize"):
gr.Markdown("### Visually crop a video.")
gr.Info("Upload a video to see a preview frame. Adjust the sliders to define the crop area, then process.")
video_crop_original_preview_state = gr.State()
with gr.Row():
with gr.Column(scale=1):
video_crop_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
gr.Markdown("#### Crop Box Settings")
video_crop_w = gr.Slider(label="Width", minimum=64, maximum=4096, step=8, value=1280)
video_crop_h = gr.Slider(label="Height", minimum=64, maximum=4096, step=8, value=720)
video_crop_x = gr.Slider(label="X Offset", minimum=0, maximum=4096, step=8, value=0)
video_crop_y = gr.Slider(label="Y Offset", minimum=0, maximum=4096, step=8, value=0)
with gr.Accordion("Optional: Resize after cropping", open=False):
video_crop_do_resize = gr.Checkbox(label="Enable Resizing", value=False)
video_crop_resize_w = gr.Number(label="Output Width", value=1024, interactive=False)
video_crop_resize_h = gr.Number(label="Output Height", value=576, interactive=False)
video_crop_btn = gr.Button("📐 Crop Video", variant="primary")
with gr.Column(scale=2):
video_crop_preview_image = gr.Image(label="Crop Preview", type="pil", interactive=False, height=UNIFIED_HEIGHT)
video_crop_output_video = gr.Video(label="Cropped Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
def setup_video_crop_preview(video_path):
if not video_path:
return None, None, gr.update(), gr.update(), gr.update(), gr.update()
try:
img = get_frame_at_time(video_path, 0)
w, h = img.size
return img, img, gr.update(maximum=w, value=w), gr.update(maximum=h, value=h), gr.update(maximum=w), gr.update(maximum=h)
except Exception as e:
gr.Warning(f"Could not load preview frame: {e}")
return None, None, gr.update(), gr.update(), gr.update(), gr.update()
video_crop_input_video.upload(
fn=setup_video_crop_preview,
inputs=video_crop_input_video,
outputs=[video_crop_preview_image, video_crop_original_preview_state, video_crop_w, video_crop_h, video_crop_x, video_crop_y]
)
video_crop_sliders = [video_crop_x, video_crop_y, video_crop_w, video_crop_h]
for slider in video_crop_sliders:
slider.release(
fn=update_crop_preview,
inputs=[video_crop_original_preview_state] + video_crop_sliders,
outputs=video_crop_preview_image
)
video_crop_do_resize.change(lambda x: [gr.update(interactive=x), gr.update(interactive=x)], inputs=video_crop_do_resize, outputs=[video_crop_resize_w, video_crop_resize_h])
video_crop_btn.click(
fn=crop_video,
inputs=[video_crop_input_video, video_crop_x, video_crop_y, video_crop_w, video_crop_h, video_crop_do_resize, video_crop_resize_w, video_crop_resize_h],
outputs=video_crop_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Change Speed"):
gr.Markdown("### Create slow-motion or fast-forward videos.")
with gr.Row():
with gr.Column():
input_video_speed = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
speed_multiplier = gr.Slider(0.1, 10.0, 1.0, step=0.1, label="Speed Multiplier")
speed_btn = gr.Button("🏃 Change Speed", variant="primary")
with gr.Column():
output_video_speed = gr.Video(label="Modified Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
speed_btn.click(fn=change_video_speed, inputs=[input_video_speed, speed_multiplier], outputs=output_video_speed).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎨 Effects & Overlays"):
with gr.Tabs():
with gr.TabItem("Fader"):
gr.Markdown("### Apply Fade-In and/or Fade-Out to a Video")
with gr.Row():
with gr.Column():
fade_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
with gr.Row():
fade_in_slider = gr.Slider(0.0, 10.0, 1.0, step=0.1, label="Fade-In Duration (s)")
fade_out_slider = gr.Slider(0.0, 10.0, 1.0, step=0.1, label="Fade-Out Duration (s)")
fade_video_btn = gr.Button("✨ Apply Fade", variant="primary")
with gr.Column():
fade_output_video = gr.Video(label="Faded Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
fade_video_btn.click(apply_video_fade, [fade_input_video, fade_in_slider, fade_out_slider], fade_output_video).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🎨 Color Grading"):
gr.Markdown("### Adjust Video Color, Contrast, and Sharpness")
gr.Info("Upload a video, load a preview frame, adjust the sliders for a live preview, and then apply the changes to the full video. The preview is 100% accurate to the final render.")
with gr.Row(equal_height=False):
with gr.Column(scale=1):
color_grade_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
load_preview_btn = gr.Button("🖼️ Load/Reset Preview Frame")
gr.Markdown("#### Adjustments")
cg_brightness = gr.Slider(minimum=-0.5, maximum=0.5, value=0.0, step=0.05, label="Brightness")
cg_contrast = gr.Slider(minimum=0.5, maximum=2.0, value=1.0, step=0.05, label="Contrast")
cg_saturation = gr.Slider(minimum=0.0, maximum=3.0, value=1.0, step=0.1, label="Saturation")
cg_sharpness = gr.Slider(minimum=0.0, maximum=1.5, value=0.0, step=0.1, label="Sharpness")
apply_grading_btn = gr.Button("🎬 Apply Grading to Full Video", variant="primary")
with gr.Column(scale=2):
with gr.Row():
cg_before_preview = gr.Image(label="Before", type="numpy", interactive=False)
cg_after_preview = gr.Image(label="After (Accurate Preview)", type="pil", interactive=False)
color_grade_output_video = gr.Video(label="Graded Video Output", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
grading_inputs = [cg_before_preview, cg_brightness, cg_contrast, cg_saturation, cg_sharpness]
cg_brightness.release(preview_color_grading_ffmpeg, inputs=grading_inputs, outputs=cg_after_preview)
cg_contrast.release(preview_color_grading_ffmpeg, inputs=grading_inputs, outputs=cg_after_preview)
cg_saturation.release(preview_color_grading_ffmpeg, inputs=grading_inputs, outputs=cg_after_preview)
cg_sharpness.release(preview_color_grading_ffmpeg, inputs=grading_inputs, outputs=cg_after_preview)
load_preview_btn.click(
fn=get_frame_at_time,
inputs=color_grade_input_video,
outputs=cg_before_preview
).then(
fn=preview_color_grading_ffmpeg,
inputs=grading_inputs,
outputs=cg_after_preview
)
color_grade_input_video.upload(
fn=get_frame_at_time,
inputs=color_grade_input_video,
outputs=cg_before_preview
).then(
fn=preview_color_grading_ffmpeg,
inputs=grading_inputs,
outputs=cg_after_preview
)
apply_grading_btn.click(
fn=apply_color_grading,
inputs=[color_grade_input_video, cg_brightness, cg_contrast, cg_saturation, cg_sharpness],
outputs=color_grade_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("BG Remover"):
gr.Markdown("## Video Background Remover")
gr.Warning("This is a very slow process. A short video can take several minutes. Output is a .webm file.")
with gr.Row():
with gr.Column():
vbg_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
vbg_btn = gr.Button("✂️ Remove Video Background", variant="primary")
with gr.Column():
vbg_output_video = gr.Video(label="Output Video with Transparency (.webm)", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
vbg_btn.click(remove_video_background, vbg_input_video, vbg_output_video).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Watermark"):
gr.Markdown("### Apply a text watermark to a video.")
with gr.Row():
with gr.Column():
wm_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
wm_text = gr.Textbox(label="Watermark Text", placeholder="(c) My Video 2025")
wm_pos = gr.Radio(["Top-Left", "Top-Right", "Bottom-Left", "Bottom-Right", "Center"], value="Bottom-Right", label="Position")
wm_opacity = gr.Slider(0, 100, 70, step=1, label="Opacity (%)")
with gr.Accordion("Advanced Options", open=False):
wm_size = gr.Slider(1, 10, 5, step=1, label="Relative Font Size")
wm_color = gr.ColorPicker(value="#FFFFFF", label="Font Color")
wm_btn = gr.Button("🖋️ Apply Watermark", variant="primary")
with gr.Column():
wm_output_video = gr.Video(label="Watermarked Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
wm_btn.click(apply_video_watermark, [wm_input_video, wm_text, wm_pos, wm_opacity, wm_size, wm_color], wm_output_video).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Create GIF"):
gr.Markdown("### Convert a video clip into a high-quality animated GIF.")
with gr.Row():
with gr.Column():
input_video_gif = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
with gr.Row():
gif_start_time = gr.Number(value=0, label="Start Time (s)")
gif_end_time = gr.Number(value=0, label="End Time (s)", info="Set to 0 for full duration")
gif_btn = gr.Button("🖼️ Create GIF", variant="primary")
with gr.Column():
output_gif = gr.Image(label="Output GIF", show_download_button=True, height=UNIFIED_HEIGHT)
gif_btn.click(create_gif_from_video, [input_video_gif, gif_start_time, gif_end_time], output_gif).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🔊 Audio & Transcription"):
with gr.Tabs():
with gr.TabItem("Add Audio"):
gr.Markdown("### Combine a silent video with an audio file.")
with gr.Row():
with gr.Column():
input_video_audio = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
input_audio = gr.Audio(type="filepath", label="Input Audio")
add_audio_btn = gr.Button("🎶 Add Audio", variant="primary")
with gr.Column():
output_video_audio = gr.Video(label="Final Video with Audio", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
add_audio_btn.click(add_audio_to_video, [input_video_audio, input_audio], output_video_audio).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Extract Audio"):
gr.Markdown("### Strip the audio track from a video file.")
with gr.Row():
with gr.Column():
extract_audio_input_video = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
extract_audio_format = gr.Dropdown(["mp3", "wav", "aac"], value="mp3", label="Output Audio Format")
extract_audio_btn = gr.Button("🎵 Extract Audio", variant="primary")
with gr.Column():
extract_audio_output = gr.Audio(label="Extracted Audio", type="filepath")
extract_audio_btn.click(extract_audio, [extract_audio_input_video, extract_audio_format], extract_audio_output).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Silence Chopper"):
gr.Markdown("### Automatic Silence Chopper")
gr.Info("Automatically split an audio file into multiple smaller files, removing the silent parts.")
with gr.Row():
with gr.Column():
chopper_input_audio = gr.Audio(type="filepath", label="Input Audio File")
with gr.Accordion("⚙️ Silence Settings", open=True):
chopper_thresh = gr.Slider(minimum=-70, maximum=-20, value=-40, step=1, label="Silence Threshold (dBFS)", info="Anything quieter than this is considered silence. Lower numbers are stricter.")
chopper_min_len = gr.Slider(minimum=100, maximum=2000, value=500, step=50, label="Minimum Silence Length (ms)", info="Silences shorter than this will be ignored.")
chopper_btn = gr.Button("✂️ Chop Audio", variant="primary")
with gr.Column():
chopper_output_gallery = gr.Gallery(label="Chopped Audio Files (Preview)", columns=2, object_fit="contain", height="auto", allow_preview=False)
chopper_output_zip = gr.File(label="Download All Chunks as .zip", interactive=False)
chopper_btn.click(
fn=chop_audio_on_silence,
inputs=[chopper_input_audio, chopper_thresh, chopper_min_len],
outputs=[chopper_output_gallery, chopper_output_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Audio Editor"):
gr.Markdown("### Trim and Apply Fades to an Audio File")
gr.Info("Set start/end times to trim the clip, then apply optional fades.")
with gr.Row():
with gr.Column():
audio_trim_input = gr.Audio(type="filepath", label="Input Audio")
with gr.Row():
audio_start_time = gr.Number(label="Start Time (s)", value=0)
audio_end_time = gr.Number(label="End Time (s)", info="Set to 0 for full duration")
with gr.Row():
audio_fade_in = gr.Slider(0.0, 10.0, 0.5, step=0.1, label="Fade-In Duration (s)")
audio_fade_out = gr.Slider(0.0, 10.0, 1.0, step=0.1, label="Fade-Out Duration (s)")
audio_trim_fade_btn = gr.Button("✂️ Process Audio", variant="primary")
with gr.Column():
audio_trim_output = gr.Audio(label="Processed Audio", type="filepath")
audio_trim_fade_btn.click(trim_and_fade_audio, [audio_trim_input, audio_start_time, audio_end_time, audio_fade_in, audio_fade_out], audio_trim_output).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🥁 BPM & Speed Tool"):
gr.Markdown("### Analyze Audio BPM and Adjust Speed")
gr.Info("Upload an audio file to find its BPM. Then, use the slider to create a speed-adjusted version.")
with gr.Row():
with gr.Column(scale=2):
bpm_input_audio = gr.Audio(type="filepath", label="Input Audio")
with gr.Row():
bpm_detect_btn = gr.Button("🥁 Detect BPM")
bpm_display_box = gr.Textbox(label="Detected BPM", interactive=False)
gr.Markdown("#### Speed Adjustment")
with gr.Row():
bpm_speed_slider = gr.Slider(minimum=0.5, maximum=2.0, value=1.0, step=0.05, label="Speed Multiplier")
bpm_new_display = gr.Textbox(label="New BPM Estimate", interactive=False, value="---")
bpm_apply_speed_btn = gr.Button("🏃 Apply Speed Change", variant="primary")
with gr.Column(scale=3):
bpm_output_audio = gr.Audio(label="Processed Audio Output", type="filepath")
bpm_input_audio.upload(lambda: ("", "---"), outputs=[bpm_display_box, bpm_new_display])
bpm_detect_btn.click(
fn=detect_bpm,
inputs=bpm_input_audio,
outputs=bpm_display_box
).then(
fn=update_new_bpm_display,
inputs=[bpm_display_box, bpm_speed_slider],
outputs=bpm_new_display
).then(fn=None, js=fire_confetti_and_sound_js)
bpm_speed_slider.release(
fn=update_new_bpm_display,
inputs=[bpm_display_box, bpm_speed_slider],
outputs=bpm_new_display
)
bpm_apply_speed_btn.click(
fn=change_audio_speed,
inputs=[bpm_input_audio, bpm_speed_slider],
outputs=bpm_output_audio
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Transcription", visible=(whisper is not None)):
gr.Markdown("## Transcribe Speech and Burn Subtitles")
gr.Info("Uses OpenAI's Whisper model with word-level timestamps. First run will download model files.")
transcribed_video_path_state = gr.State(None)
with gr.Row():
with gr.Column():
transcribe_input = gr.File(label="Upload Video or Audio File", file_types=["video", "audio"])
transcribe_model = gr.Dropdown(["tiny", "base", "small", "medium", "large"], value="base", label="Whisper Model Size")
transcribe_btn = gr.Button("🎙️ Transcribe", variant="primary")
with gr.Column():
with gr.Row():
transcribe_text = gr.Textbox(label="Transcription Result", lines=10, interactive=True, elem_id="transcription_textbox")
copy_transcription_btn = gr.Button("📋 Copy")
transcribe_files = gr.File(label="Download Subtitle Files (.srt, .vtt, .ass)", file_count="multiple", interactive=False)
with gr.Accordion("🔥 Burn Subtitles onto Video", open=True, visible=False) as burn_accordion:
gr.Markdown("Set styling and burn the generated subtitles into the video.")
with gr.Row():
burn_style = gr.Radio(["Block", "Karaoke"], value="Block", label="Subtitle Style")
burn_font_size = gr.Slider(1, 10, 5, step=1, label="Relative Font Size")
with gr.Row(visible=True) as block_style_row:
burn_words_per_line = gr.Slider(1, 20, 7, step=1, label="Max Words Per Line", info="Splits long subtitle lines for better readability.")
burn_block_font_color = gr.ColorPicker(value="#FFFFFF", label="Font Color")
with gr.Row(visible=False) as karaoke_style_row:
burn_karaoke_base_color = gr.ColorPicker(value="#FFFFFF", label="Base Color")
burn_karaoke_highlight_color = gr.ColorPicker(value="#FFFF00", label="Highlight Color")
burn_btn = gr.Button("🔥 Burn Subtitles", variant="primary")
burn_output_video = gr.Video(label="Video with Burned-in Subtitles", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
def toggle_subtitle_styles(style_choice):
return gr.update(visible=(style_choice == "Block")), gr.update(visible=(style_choice == "Karaoke"))
burn_style.change(toggle_subtitle_styles, burn_style, [block_style_row, karaoke_style_row])
copy_transcription_btn.click(fn=None, js=copy_transcription_js)
transcribe_btn.click(
fn=transcribe_and_prep_burn,
inputs=[transcribe_input, transcribe_model],
outputs=[transcribe_text, transcribe_files, transcribed_video_path_state, burn_accordion]
).then(fn=None, js=fire_confetti_and_sound_js)
burn_btn.click(
fn=burn_subtitles_wrapper,
inputs=[transcribed_video_path_state, transcribe_files, burn_style, burn_font_size, burn_block_font_color, burn_words_per_line, burn_karaoke_base_color, burn_karaoke_highlight_color],
outputs=burn_output_video
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🧠 ControlNet", elem_id="controlnet_tab"):
gr.Markdown("## ControlNet Preprocessing")
with gr.Tabs():
with gr.TabItem("Process a Video"):
gr.Markdown("### Convert a Video into a ControlNet-Ready Map")
with gr.Row():
with gr.Column():
input_video_cn = gr.Video(label="Input Video", height=UNIFIED_HEIGHT)
detector_dropdown_cn = gr.Dropdown(choices=list(DETECTOR_CONFIG.keys()), value="Canny", label="Choose Detector")
process_btn_cn = gr.Button("✨ Process Video", variant="primary")
with gr.Column():
output_video_cn = gr.Video(label="Output ControlNet Video", interactive=True, show_download_button=True, height=UNIFIED_HEIGHT)
process_btn_cn.click(fn=process_video_with_detector, inputs=[input_video_cn, detector_dropdown_cn], outputs=output_video_cn).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Process Batch Images"):
gr.Markdown("### Generate ControlNet Maps from one or more images.")
with gr.Row():
with gr.Column():
input_images_cn = gr.File(label="Upload Images or Folder", file_count="multiple", file_types=["image"])
detector_dropdown_img = gr.Dropdown(choices=list(DETECTOR_CONFIG.keys()), value="Canny", label="Choose Detector")
process_btn_img = gr.Button("✨ Process Images", variant="primary")
with gr.Column():
output_gallery_cn = gr.Gallery(label="Output ControlNet Images", columns=4, object_fit="contain", height="auto")
output_zip_cn = gr.File(label="Download All as .zip", interactive=False)
process_btn_img.click(fn=process_batch_images_with_detector, inputs=[input_images_cn, detector_dropdown_img], outputs=[output_gallery_cn, output_zip_cn]).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🗜️ Converter", elem_id="converter_tab"):
gr.Markdown("## Universal Media Converter & Compressor")
gr.Info("Convert your video or audio files to different formats, change codecs, and adjust quality to save space.")
with gr.Tabs():
with gr.TabItem("Batch Video Converter"):
with gr.Row():
with gr.Column():
conv_input_videos = gr.File(label="Upload Videos", file_count="multiple", file_types=["video"])
conv_btn = gr.Button("⚙️ Convert & Compress Videos", variant="primary")
conv_output_gallery = gr.Gallery(label="Converted Videos Preview", columns=2, object_fit="contain")
conv_output_zip = gr.File(label="Download All as .zip", interactive=False)
with gr.Column():
gr.Markdown("#### Output Settings")
with gr.Row():
conv_format = gr.Dropdown(["mp4", "mkv", "webm", "mov"], value="mp4", label="Output Format")
conv_vcodec = gr.Dropdown(["libx264", "libx265", "vp9"], value="libx264", label="Video Codec")
conv_crf = gr.Slider(minimum=18, maximum=30, value=23, step=1, label="Quality (CRF)", info="Lower = higher quality/size, Higher = lower quality/size. 23 is a good default.")
conv_scale = gr.Dropdown(["Original", "1080p", "720p", "480p"], value="Original", label="Downscale Resolution (optional)")
gr.Markdown("##### Audio Settings")
with gr.Row():
conv_acodec = gr.Dropdown(["copy", "aac", "opus"], value="copy", label="Audio Codec", info="'copy' is fastest and preserves quality.")
conv_abitrate = gr.Dropdown([96, 128, 192, 256, 320], value=192, label="Audio Bitrate (kbps)", interactive=False)
conv_acodec.change(lambda x: gr.update(interactive=(x != "copy")), conv_acodec, conv_abitrate)
conv_btn.click(
fn=batch_convert_compress_videos,
inputs=[conv_input_videos, conv_format, conv_vcodec, conv_crf, conv_scale, conv_acodec, conv_abitrate],
outputs=[conv_output_gallery, conv_output_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("Batch Audio Converter / Extractor"):
gr.Markdown("### Batch Audio Converter & Extractor")
gr.Info("Upload multiple audio OR video files. The audio will be extracted from videos and all files will be converted to your chosen format.")
with gr.Row():
with gr.Column(scale=2):
audio_conv_inputs = gr.File(label="Upload Audio or Video Files", file_count="multiple", file_types=["audio", "video"])
with gr.Row():
audio_conv_format = gr.Dropdown(["mp3", "wav", "aac", "flac", "ogg"], value="mp3", label="Output Format")
audio_conv_bitrate = gr.Dropdown([96, 128, 192, 256, 320], value=192, label="Bitrate (kbps)", info="Higher = better quality. Not used for WAV/FLAC.")
audio_conv_btn = gr.Button("🎵 Convert All Files", variant="primary")
with gr.Column(scale=3):
audio_conv_output_player = gr.Audio(label="Preview First Converted File", type="filepath")
audio_conv_output_zip = gr.File(label="Download All as .zip", interactive=False)
audio_conv_format.change(
fn=lambda fmt: gr.update(interactive=(fmt not in ["wav", "flac"])),
inputs=audio_conv_format,
outputs=audio_conv_bitrate
)
audio_conv_btn.click(
fn=batch_convert_audio,
inputs=[audio_conv_inputs, audio_conv_format, audio_conv_bitrate],
outputs=[audio_conv_output_player, audio_conv_output_zip],
show_progress="full"
).then(fn=None, js=fire_confetti_and_sound_js)
with gr.TabItem("🔗 Transfer", elem_id="transfer_tab"):
gr.Markdown("## Image & Link Transfer Utility")
gr.Info("Drop images below, manage URL presets, and open the target application in a new tab.")
link_presets = gr.State(DEFAULT_LINK_PRESETS.copy())
with gr.Row():
with gr.Column(scale=1):
transfer_gallery = gr.Gallery(label="Drop Images Here", height=300, columns=3, object_fit="contain")
with gr.Column(scale=2):
gr.Markdown("### Link Preset Management")
target_url = gr.Textbox(label="Target URL", value="https://huggingface.co/spaces/bep40/FramePack_rotate_landscape", interactive=True, elem_id="transfer_target_url")
search_bar = gr.Textbox(label="Search Presets", placeholder="Type to filter...", interactive=True)
with gr.Row():
preset_dropdown = gr.Dropdown(
label="Load Link Preset",
choices=sorted(list(DEFAULT_LINK_PRESETS.keys())),
interactive=True
)
delete_preset_btn = gr.Button("🗑️ Delete", variant="stop")
with gr.Accordion("Create a new preset", open=False):
with gr.Row():
new_preset_name = gr.Textbox(label="New Preset Name", placeholder="e.g., My Favorite App")
save_preset_btn = gr.Button("💾 Save")
open_link_btn = gr.Button("🚀 Open in New Tab", variant="primary")
search_bar.input(fn=filter_presets, inputs=[search_bar, link_presets], outputs=[preset_dropdown])
preset_dropdown.change(fn=load_preset, inputs=[link_presets, preset_dropdown], outputs=[target_url])
save_preset_btn.click(
fn=save_preset, inputs=[link_presets, new_preset_name, target_url], outputs=[link_presets, preset_dropdown]
).then(lambda: ("", ""), outputs=[new_preset_name, search_bar])
delete_confirm_js = """(name) => { if (!name) { alert('Please select a preset to delete.'); return false; } return confirm(`Are you sure you want to delete the preset: '` + name + `'?`); }"""
delete_preset_btn.click(fn=None, js=delete_confirm_js, inputs=[preset_dropdown]).then(
fn=delete_preset, inputs=[link_presets, preset_dropdown], outputs=[link_presets, preset_dropdown, target_url]
).then(lambda: "", outputs=[search_bar])
open_link_btn.click(fn=None, js="()=>{const url=document.getElementById('transfer_target_url').querySelector('textarea').value;if(url){window.open(url,'_blank')}else{alert('Target URL is empty.')}}")
main_tabs.select(fn=None, inputs=main_tabs, js="(tab) => { window.skriptz_bling.update_title(tab); }")
gr.HTML('<a href="https://linktr.ee/skylinkd" target="_blank" style="color: #94a3b8; text-decoration: none;">skylinkd production 2025 (c)</a>', elem_id="custom-footer")
if __name__ == "__main__":
if os.path.exists(TEMP_DIR):
try: shutil.rmtree(TEMP_DIR)
except OSError as e: print(f"Error removing temp directory {TEMP_DIR}: {e}")
os.makedirs(TEMP_DIR, exist_ok=True)
if whisper:
load_whisper_model("base") # Pre-load the default model on startup
demo.launch(inbrowser=True) |