JSON-LD parsing, processing, serialization
Clone
HTTPS:
darcs clone https://vervis.peers.community/repos/kv5zE
SSH:
darcs clone USERNAME@vervis.peers.community:kv5zE
Tags
TODO
Berenice.hs
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 | {- This file is part of berenice.
-
- Written in 2018, 2019 by fr33domlover <fr33domlover@riseup.net>.
-
- ♡ Copying is an act of love. Please copy, reuse and share.
-
- The author(s) have dedicated all copyright and related and neighboring
- rights to this software to the public domain worldwide. This software is
- distributed without any warranty.
-
- You should have received a copy of the CC0 Public Domain Dedication along
- with this software. If not, see
- <http://creativecommons.org/publicdomain/zero/1.0/>.
-}
{-# LANGUAGE DeriveFoldable #-}
{-# LANGUAGE DeriveFunctor #-}
{-# LANGUAGE DeriveGeneric #-}
{-# LANGUAGE DeriveTraversable #-}
{-# LANGUAGE OverloadedStrings #-}
{-# LANGUAGE TemplateHaskell #-}
module Data.Berenice
(
)
where
import Control.Applicative ((<|>))
import Control.Monad ((>=>), when, guard)
import Control.Monad.Trans.Class (lift)
import Control.Monad.Trans.Except
import Control.Monad.Trans.State.Strict
import Data.Bifunctor (bimap, first, second)
import Data.ByteString (ByteString)
import Data.Either (partitionEithers)
import Data.Foldable (fold, foldl', foldlM, foldrM, asum, minimumBy)
import Data.Function (on)
import Data.Functor ((<&>))
import Data.Hashable
import Data.HashMap.Strict (HashMap)
import Data.Int (Int64)
import Data.List (sortBy, sortOn)
import Data.List.NonEmpty (NonEmpty (..), (<|), nonEmpty)
import Data.Maybe (fromMaybe, isJust, isNothing, mapMaybe, catMaybes)
import Data.Scientific
import Data.Vector (Vector)
import Data.Text (Text)
import Data.Text.Encoding (encodeUtf8, decodeUtf8)
import Data.Traversable (for)
import GHC.Generics (Generic)
import Network.IRI hiding (Ref)
import Numeric (showEFloat)
import Text.Read (readMaybe)
import qualified Data.ByteString as B (length)
import qualified Data.ByteString.Char8 as BC (pack)
import qualified Data.HashMap.Strict as M
import qualified Data.HashSet as S
import qualified Data.List.NonEmpty as NE
import qualified Data.Text as T
import qualified Data.Vector as V
import qualified Network.IRI as U (Ref)
import Data.Berenice.TH
-- Tuple utils
(&&&) :: (a -> b) -> (a -> c) -> a -> (b, c)
(f &&& g) x = (f x, g x)
-- Maybe utils
partitionMaybes :: [(Maybe a, b)] -> ([(a, b)], [b])
partitionMaybes = foldr f ([], [])
where
f (Nothing, y) (ps, ys) = (ps , y : ys)
f (Just x , y) (ps, ys) = ((x, y) : ps, ys)
partitionMaybesNE
:: NonEmpty (Maybe a, b)
-> Either (NonEmpty (a, b), [b]) ([(a, b)], NonEmpty b)
partitionMaybesNE ((mx, y) :| zs) =
let (ps, ys) = partitionMaybes zs
in case mx of
Nothing -> Right (ps , y :| ys)
Just x -> Left ((x, y) :| ps, ys)
-- Foldable utils
-- | Like 'foldl'', except it checks each intermediate result (including the
-- initial value passed) with the given predicate. If the predicate gives
-- False, then that intermediate result is returned and the rest of the
-- structure isn't observed, allowing for early termination.
--
-- >>> foldlWhile' (< 200) (+) 0 $ repeat 31
-- 217
foldlWhile' :: Foldable f => (b -> Bool) -> (b -> a -> b) -> b -> f a -> b
foldlWhile' p f i xs = foldr f' id xs i
where
f' x k z =
if p z
then k $! f z x
else z
-- | Like 'foldlWhile'', except folding stops when the predicate returns 'True'
-- rather than False. In other words, @foldlUntil' p@ is the same as
-- @foldlWhile' $ not . p@. Fold while checking each intermediate result
-- (including the initial value passed) with a predicate, stopping the folding
-- if it returns True. In other words, return the first intermediate result for
-- which the predicate holds (or the result of the full fold, if the predicate
-- returns 'False' all the way).
--
-- >>> foldlUntil' (> 106) (+) 18 $ repeat 19
-- 113
foldlUntil' :: Foldable f => (b -> Bool) -> (b -> a -> b) -> b -> f a -> b
foldlUntil' p = foldlWhile' $ not . p
-- NonEmpty utils
extract :: (a -> b) -> (a -> c) -> NonEmpty a -> (b, NonEmpty c)
extract f g (head :| tail) = (f head, g head :| map g tail)
groupAllExtract :: Ord b => (a -> b) -> (a -> c) -> [a] -> [(b, NonEmpty c)]
groupAllExtract f g = map (extract f g) . NE.groupAllWith f
groupAllExtract1
:: Ord b => (a -> b) -> (a -> c) -> NonEmpty a -> NonEmpty (b, NonEmpty c)
groupAllExtract1 f g = NE.map (extract f g) . NE.groupAllWith1 f
groupAllExtractDefault
:: Ord b => b -> (a -> Maybe b) -> (a -> c) -> [a] -> [(b, NonEmpty c)]
groupAllExtractDefault d mf g l =
let (justs, nothings) = partitionMaybes $ map (mf &&& g) l
groups = groupAllExtract fst snd justs
in case nonEmpty nothings of
Nothing -> groups
Just nothings' ->
case update fst (second (<> nothings')) d groups of
Just groups' -> groups'
Nothing -> (d, nothings') : groups
where
update :: Ord b => (a -> b) -> (a -> a) -> b -> [a] -> Maybe [a]
update p f x l =
let (lt, eqgt) = span ((< x) . p) l
in case eqgt of
[] -> Nothing
y:ys ->
if x == p y
then Just $ lt ++ f y : ys
else Nothing
groupAllExtractDefault1
:: Ord b
=> b
-> (a -> Maybe b)
-> (a -> c)
-> NonEmpty a
-> NonEmpty (b, NonEmpty c)
groupAllExtractDefault1 d mf g l =
case partitionMaybesNE $ NE.map (mf &&& g) l of
Left (justs, nothings) ->
let groups = groupAllExtract1 fst snd justs
in case nonEmpty nothings of
Nothing -> groups
Just nothings' -> insertNothings nothings' groups
Right (justs, nothings) ->
case nonEmpty justs of
Nothing -> (d, nothings) :| []
Just justs' ->
let groups = groupAllExtract1 fst snd justs'
in insertNothings nothings groups
where
update
:: Ord b
=> (a -> b) -> (a -> a) -> b -> NonEmpty a -> Maybe (NonEmpty a)
update p f x l =
let (lt, eqgt) = NE.span ((< x) . p) l
in case eqgt of
[] -> Nothing
y:ys ->
if x == p y
then Just $ case nonEmpty lt of
Nothing -> f y :| ys
Just lt' -> lt' <> (f y :| ys)
else Nothing
insertNothings nothings groups =
case update fst (second (<> nothings)) d groups of
Just groups' -> groups'
Nothing -> (d, nothings) <| groups
-- HashMap utils
lookupAndDelete k m =
case M.lookup k m of
Nothing -> (Nothing, m)
Just v -> (Just v , M.delete k m)
recordMap3
:: (Ord b1, Hashable b1, Ord b2, Hashable b2)
=> (a -> b1)
-> (a -> b2)
-> (a -> b3)
-> [a]
-> HashMap b1 (HashMap b2 (Vector b3))
recordMap3 field1 field2 field3 =
M.fromList .
map (second $ ne2hm . NE.map (second ne2v) . groupAllExtract1 fst snd) .
groupAllExtract field1 (\ x -> (field2 x, field3 x))
where
ne2hm = M.fromList . NE.toList
ne2v = V.fromList . NE.toList
-- We're going to go through the process of parsing JSON-LD and converting to
-- RDF triples. We start with the basic data model that JSON-LD uses, which is
-- the one of JSON:
-- | A container that holds either a single value or an array or zero or more
-- values. It preserves the distinction between a single value and an array.
data Multi a = One a | Many (Vector a)
deriving (Eq, Functor, Foldable, Traversable)
data Scalar
= ScalarString Text
| ScalarNumber Scientific
| ScalarBool Bool
type Dictionary = HashMap Text (Multi Value)
data Value
= ValueScalar Scalar
| ValueDictionary Dictionary
| ValueNull
-- Once we parse a JSON document into that structure, in the next step we'll do
-- the following:
--
-- * Parse the contexts, from 'Value' to a dedicated context type
-- * Parse the properties, identifying keywords and URIs and so on
-- * Parse string scalars, identifying keywords and URIs and so on
--
-- Here's the new structure we'll produce. First, some building blocks:
data KeywordC
= KeywordContainer
| KeywordBase
| KeywordVocab
deriving (Eq, Ord, Generic)
instance Hashable KeywordC
data KeywordNC
= KeywordId
| KeywordValue
| KeywordLanguage
| KeywordType
| KeywordList
| KeywordSet
| KeywordReverse
| KeywordIndex
| KeywordGraph
deriving (Eq, Ord, Generic)
instance Hashable KeywordNC
data Keyword
= KeywordContext
| KeywordC KeywordC
| KeywordNC KeywordNC
-- | A string that syntactically can be interpretted as a absolute URI, or a
-- compact URI, or as both.
data AbsoluteOrCompact
= JustAbsolute (URI Gen)
-- ^ The string parses as a valid absolute URI, but it can't be used as a
-- compact URI. For example if the colon is followed by two slashes, the
-- string may be considered as as an absolute URI and not allowed to be
-- considered as a compact one.
| JustCompact CompactURI
-- ^ The string fails to parse as a valid absolute URI, but it has a colon
-- and can be used as a compact URI. In the JSON-LD spec, strings can be
-- considered as absolute URIs without being valid URIs, so if a string
-- with a colon doesn't have a term definition for its prefix, the error to
-- give may be "invalid absolute URI". Because in JSON-LD it's blindly
-- treated as an absolute URI, while here we discard it on this basis.
| AbsoluteOrCompact (URI Gen) CompactURI
-- ^ The string is syntactically both a valid absolute URI and a valid
-- compact URI.
deriving (Eq, Ord, Generic)
instance Hashable AbsoluteOrCompact where
hashWithSalt s (JustAbsolute u) =
s `hashWithSalt` (0 :: Int) `hashWithSalt` u
hashWithSalt s (JustCompact c) =
s `hashWithSalt` (1 :: Int) `hashWithSalt` c
hashWithSalt s (AbsoluteOrCompact _u c) =
s `hashWithSalt` (1 :: Int) `hashWithSalt` c
-- | A reference that possibly points to some other string value, such as a URI
-- or a keyword. Ref expansion (or IRI expansion, as the JSON-LD spec calls it)
-- can produce the value that's being referred to.
data Ref
= RefURI AbsoluteOrCompact
| RefBlank RelToken
| RefTerm RelNoAuth
deriving (Eq, Ord, Generic)
instance Hashable Ref
data ObjectKey
= ObjectKeyKeyword KeywordNC
| ObjectKeyOther Ref
data Weird
= WKeywordContext
| WKeywordC KeywordC
| WKeywordU Text
| WEmpty
| WBlank Text
| WTerm Text
deriving Eq
data Token
= TWeird Weird
| TKeywordNC KeywordNC
| TRef Ref
deriving Eq
data TScalar
= TScalarString Text Token
| TScalarNumber Scientific
| TScalarBool Bool
deriving Eq
-- Now, here's the typed parsed context structure:
type Reverse = Ref
-- TODO
-- Does it really make sense for Id to allow Keyword? Maybe just KeywordNC? Or
-- KeywordNC+KeywordC but excluding KeywordContext?
data Id
= IdKeyword Keyword
| IdRef Ref
| IdNull
data Type
= TypeURI AbsoluteOrCompact
| TypeTerm RelNoAuth
| TypeId
| TypeVocab
| TypeNull
data Container
= ContainerNull
| ContainerSet
| ContainerList
| ContainerLanguage
| ContainerIndex
data ContainerReverse
= ContainerReverseNull
| ContainerReverseSet
| ContainerReverseIndex
data ExpandedTermDefinitionId = ExpandedTermDefinitionId
{ expandedId :: Maybe Id
, expandedContainer :: Maybe Container
}
data ExpandedTermDefinitionReverse = ExpandedTermDefinitionReverse
{ expandedReverse :: Reverse
, expandedContainerReverse :: Maybe ContainerReverse
}
data ExpandedTermDefinition = ExpandedTermDefinition
{ expandedIdOrReverse ::
Either ExpandedTermDefinitionId ExpandedTermDefinitionReverse
, expandedType :: Maybe Type
, expandedLanguage :: Maybe ContextLanguage -- just reusing bc same spec
}
-- Reusing Id simply because same possibilities
data ContextValue
= ContextValueId Id
| ContextValueExpanded ExpandedTermDefinition
data ContextLanguage
= ContextLanguageTag Text
| ContextLanguageNull
data ContextBase
= ContextBaseAbsolute (URI Gen)
| ContextBaseRelative (URI Rel)
| ContextBaseNull
data ContextVocab
= ContextVocabAbsolute (URI Gen)
| ContextVocabBlank RelToken -- the stuff after the _:
| ContextVocabNull
data Context = Context
{ contextLanguage :: Maybe ContextLanguage
, contextBase :: Maybe ContextBase
, contextVocab :: Maybe ContextVocab
, contextValues :: HashMap Ref ContextValue
}
data LocalContextItem
= LocalContextObject Context
| LocalContextString (URI U.Ref)
| LocalContextNull
-- And finally the complete new structure:
data UnknownItem
= UnknownItemScalar TScalar
| UnknownItemNull
| UnknownItemObject UnknownObject
data UnknownObject = UnknownObject
{ unknownContext :: Maybe (Multi LocalContextItem)
, unknownMap :: HashMap Text (ObjectKey, Multi UnknownItem)
}
-- Now let's see how we convert 'Value' to 'UnknownItem'. And a dictionary into
-- an 'UnknownObject'. First, let's define some error types for the functions
-- to use. Note that these error types are also for later steps and may be
-- including some structures defined later below.
data SpecErrorCode
-- | LoadingDocumentFailed
= ListOfLists
| InvalidIndexValue -- berenice-aeson, when @index field in a node object is not a JSON string
| ConflictingIndexes
| InvalidIdValue -- berenice-aeson, when @id field in a node object is not a JSON string
-- | InvalidLocalContext
-- | MultipleContextLinkHeaders
-- | LoadingRemoteContextFailed
| InvalidRemoteContext
| RecursiveContextInclusion
| InvalidBaseURI
| InvalidVocabMapping
| InvalidDefaultLanguage -- berenice-aeson should use this if it parses a @language that isn't an aeson String
| KeywordRedefinition -- berenice-aeson should raise if parsing a Ref in the @context finds it to be a keyword
| InvalidTermDefinition -- berenice-aeson should raise if parsing a ContextValue fails, in particular it's neither JSON string nor object
| InvalidReverseProperty -- berenice-aeson, when expanded term def contains both @id and @reverse. Or otherwise not by the rules? :) Also if @reverse is present and @container isn't one of the values of 'ContainerReverse'
| InvalidURIMapping -- berenice-aeson, when parsing expanded term definition and @id or @reverse isn't a JSON string, or @reverse parsing fails
| CyclicURIMapping Ref
| InvalidKeywordAlias -- berenice-aeson, when parsing context value or expanded term def @id and the value is @context
| InvalidTypeMapping -- berenice-aeson, when parsing expanded term def and @type maps to a non-string. It's also used here below
| InvalidLanguageMapping
| CollidingKeywords
| InvalidContainerMapping -- used below + berenice aeson when parsing @container in @context (at least when without @reverse) fails
| InvalidTypeValue -- berenice-aeson, when parsing node object field, key is @type and value isn't JSON string or array of strings
| InvalidValueObject
| InvalidValueObjectValue
| InvalidLanguageTaggedString
| InvalidLanguageTaggedValue
| InvalidTypedValue
-- | InvalidSetOrListObject
| InvalidLanguageMapValue
| CompactionToListOfLists
| InvalidReversePropertyMap
| InvalidReverseValue -- berenice-aeson, if in a node object @reverse is mapped to something that isn't a JSON object
| InvalidReversePropertyValue
data TermDefInvalidType
= ITBlank RelToken
| ITKeywordNC KeywordNC
| ITWeird Weird
data Weird'
= WKeywordU' Text
| WEmpty'
| WBlank' Text
data BereniceError
= SpecError SpecErrorCode String
| InvalidObjectKey Weird
| ContextInvalidURI Text
| ContextInvalidScalar (Either Scientific Bool)
| ContextInvalidKey Weird'
| TermDefInvalidMember
| TermDefInvalidId Weird'
| TermDefInvalidReverse (Either KeywordNC Weird)
| TermDefInvalidType TermDefInvalidType
| InvalidTaggedKey KeywordC
| ValueInvalidType (Either KeywordNC Weird)
| ListInvalidMember
| NodeInvalidId (Either KeywordNC Weird)
| NodeInvalidType (Either KeywordNC Weird)
| NodeInvalidReverse (Either KeywordNC Weird)
| NodeInvalidGraph
| ENodeInvalidId (Either KeywordNC Weird)
| ENodeInvalidType (Either KeywordC KeywordNC)
| NoTermDefForCompactRefPrefix CompactURI
| TermOrRelativeNotExpanded RelNoAuth
| InvalidConcatenatedURI (URI Gen) RelNoAuth
| InvalidConcatenatedBlank RelToken RelNoAuth
| InvalidExpandedURI (URI Gen) CompactURI
| KeywordAliasUsedAsPrefix CompactURI
| NodeObjectIgnoredField NodeObject KeywordNC NodeValue
| GraphIsntNodeObject NodeObject Ref NodeValue
| ExpandValueSetIdToNonString ActiveContext Ref TScalar
| BlankProperty Identifier RelToken
| LangStringNoLanguage (Either TypedLiteralNB TypedLiteralString)
| NonLangStringHasLanguage (Either TypedLiteralNB TypedLiteralString) Text
| SuspiciouslyBigInteger Scientific
| OutOfRangeOfDouble Scientific
-- below are errors for the rdf->jsonld direction:
| InvalidXsdBoolean Text
| InvalidXsdInteger Text
| InvalidXsdDouble Text
| InvalidRdfType Identifier PlainLiteral
-- Finally, the functions for transforming our document from terms of 'Value'
-- into terms of 'UnknownItem':
uri2ac :: URI Gen -> AbsoluteOrCompact
uri2ac u =
case parseCompactFromURI u of
Nothing -> JustAbsolute u
Just c -> AbsoluteOrCompact u c
compact2ac :: CompactURI -> AbsoluteOrCompact
compact2ac c =
case parseURIFromCompact c of
Nothing -> JustCompact c
Just u -> AbsoluteOrCompact u c
parseToken :: Text -> Token
parseToken t = case T.uncons t of
Nothing -> TWeird WEmpty
Just ('@', r) ->
case parseKeyword r of
Just kw -> case kw of
KeywordContext -> TWeird WKeywordContext
KeywordC kwc -> TWeird $ WKeywordC kwc
KeywordNC kwnc -> TKeywordNC kwnc
Nothing -> TWeird $ WKeywordU r
_ -> case T.stripPrefix "_:" t of
Just b ->
case parseRelToken $ encodeUtf8 b of
Just rt -> TRef $ RefBlank rt
Nothing -> TWeird $ WBlank b
Nothing ->
case parseAC t of
Just ac -> TRef $ RefURI ac
Nothing ->
case parseRelNoAuth $ encodeUtf8 t of
Nothing -> TWeird $ WTerm t
Just r -> TRef $ RefTerm r
where
parseKeyword r = case r of
"context" -> Just KeywordContext
"id" -> Just $ KeywordNC KeywordId
"value" -> Just $ KeywordNC KeywordValue
"language" -> Just $ KeywordNC KeywordLanguage
"type" -> Just $ KeywordNC KeywordType
"container" -> Just $ KeywordC KeywordContainer
"list" -> Just $ KeywordNC KeywordList
"set" -> Just $ KeywordNC KeywordSet
"reverse" -> Just $ KeywordNC KeywordReverse
"index" -> Just $ KeywordNC KeywordIndex
"base" -> Just $ KeywordC KeywordBase
"vocab" -> Just $ KeywordC KeywordVocab
"graph" -> Just $ KeywordNC KeywordGraph
_ -> Nothing
parseAC t =
case parseURI b of
Nothing -> JustCompact <$> parseCompactURI b
Just u -> Just $ uri2ac u
where
b = encodeUtf8 t
parseContextItem :: Value -> Either BereniceError LocalContextItem
parseContextItem (ValueScalar s) =
case s of
ScalarString t ->
case parseURIReference $ encodeUtf8 t of
Just u -> Right $ LocalContextString u
Nothing -> Left $ ContextInvalidURI t
ScalarNumber n -> Left $ ContextInvalidScalar $ Left n
ScalarBool b -> Left $ ContextInvalidScalar $ Right b
parseContextItem ValueNull = Right LocalContextNull
parseContextItem (ValueDictionary m) =
let (ml, m1) = lookupAndDelete "@language" m
(mb, m2) = lookupAndDelete "@base" m1
(mv, m3) = lookupAndDelete "@vocab" m2
in fmap LocalContextObject $
Context
<$> traverse parseLanguage ml
<*> traverse parseBase mb
<*> traverse parseVocab mv
<*> fmap M.fromList (traverse parseNonKW $ M.toList m3)
where
w2w' :: Weird -> Either (Maybe KeywordC) Weird'
w2w' WKeywordContext = Left Nothing
w2w' (WKeywordC kw) = Left $ Just kw
w2w' (WKeywordU t) = Right $ WKeywordU' t
w2w' WEmpty = Right WEmpty'
w2w' (WBlank rt) = Right $ WBlank' rt
parseLanguage v = case v of
One (ValueScalar (ScalarString t)) ->
Right $ ContextLanguageTag t
One ValueNull -> Right ContextLanguageNull
_ -> Left $ SpecError InvalidDefaultLanguage ""
parseBase v = case v of
One (ValueScalar (ScalarString t)) ->
let b = encodeUtf8 t
in case parseURI b of
Just u -> Right $ ContextBaseAbsolute u
Nothing -> case parseRelativeReference b of
Just u -> Right $ ContextBaseRelative u
Nothing -> Left $ SpecError InvalidBaseURI ""
One ValueNull -> Right ContextBaseNull
_ -> Left $ SpecError InvalidBaseURI ""
parseVocab v = case v of
One (ValueScalar (ScalarString t)) -> case parseToken t of
TRef (RefURI (JustAbsolute u)) ->
Right $ ContextVocabAbsolute u
TRef (RefURI (AbsoluteOrCompact u _)) ->
Right $ ContextVocabAbsolute u
TRef (RefBlank b) -> Right $ ContextVocabBlank b
_ -> Left $ SpecError InvalidVocabMapping ""
One ValueNull -> Right ContextVocabNull
_ -> Left $ SpecError InvalidVocabMapping ""
parseId' t = case parseToken t of
TWeird w ->
case w2w' w of
Left mkw ->
Right $ IdKeyword $
case mkw of
Nothing -> KeywordContext
Just kw -> KeywordC kw
Right w' -> Left $ TermDefInvalidId w'
TKeywordNC kw -> Right $ IdKeyword $ KeywordNC kw
TRef ck -> Right $ IdRef ck
parseId v = case v of
One (ValueScalar (ScalarString t)) -> parseId' t
One ValueNull -> Right IdNull
_ -> Left $ SpecError InvalidURIMapping ""
parseContainer v = case v of
One (ValueScalar (ScalarString t)) -> case t of
"@set" -> Right ContainerSet
"@list" -> Right ContainerList
"@language" -> Right ContainerLanguage
"@index" -> Right ContainerIndex
_ -> Left $ SpecError InvalidContainerMapping ""
One ValueNull -> Right ContainerNull
_ -> Left $ SpecError InvalidContainerMapping ""
parseContainerReverse v = case v of
One (ValueScalar (ScalarString t)) -> case t of
"@set" -> Right ContainerReverseSet
"@index" -> Right ContainerReverseIndex
_ -> Left $ SpecError InvalidReverseProperty ""
One ValueNull -> Right ContainerReverseNull
_ -> Left $ SpecError InvalidReverseProperty ""
parseReverse v = case v of
One (ValueScalar (ScalarString t)) -> case parseToken t of
TWeird w -> Left $ TermDefInvalidReverse $ Right w
TKeywordNC kw -> Left $ TermDefInvalidReverse $ Left kw
TRef ck -> Right ck
_ -> Left $ SpecError InvalidURIMapping ""
parseIdReverse (Just _) (Just _) _ =
Left $ SpecError InvalidReverseProperty ""
parseIdReverse Nothing (Just r) mc = fmap Right $
ExpandedTermDefinitionReverse
<$> parseReverse r
<*> traverse parseContainerReverse mc
parseIdReverse mi Nothing mc = fmap Left $
ExpandedTermDefinitionId
<$> traverse parseId mi
<*> traverse parseContainer mc
parseType v = case v of
One (ValueScalar (ScalarString t)) ->
case parseToken t of
TWeird w ->
case w of
WKeywordC KeywordVocab -> Right TypeVocab
_ -> Left $ TermDefInvalidType $ ITWeird w
TKeywordNC kw ->
case kw of
KeywordId -> Right TypeId
_ -> Left $ TermDefInvalidType $ ITKeywordNC kw
TRef ck ->
case ck of
RefURI ac -> Right $ TypeURI ac
RefBlank b ->
Left $ TermDefInvalidType $ ITBlank b
RefTerm r -> Right $ TypeTerm r
One ValueNull -> Right TypeNull
_ -> Left $ SpecError InvalidTypeMapping ""
parseLang v = case v of
One (ValueScalar (ScalarString t)) -> Right $ ContextLanguageTag t
One ValueNull -> Right ContextLanguageNull
_ -> Left $ SpecError InvalidLanguageMapping ""
parseTermDef m =
let (mt, m1) = lookupAndDelete "@type" m
(mr, m2) = lookupAndDelete "@reverse" m1
(mi, m3) = lookupAndDelete "@id" m2
(mc, m4) = lookupAndDelete "@container" m3
(ml, m5) = lookupAndDelete "@language" m4
in if M.null m5
then ExpandedTermDefinition
<$> parseIdReverse mi mr mc
<*> traverse parseType mt
<*> traverse parseLang ml
else Left TermDefInvalidMember
parseNonKW (k, vs) =
(,)
<$> (case parseToken k of
TWeird w ->
case w2w' w of
Left _ -> Left $ SpecError KeywordRedefinition ""
Right w' -> Left $ ContextInvalidKey w'
TKeywordNC _ -> Left $ SpecError KeywordRedefinition ""
TRef ck -> Right ck
)
<*> (case vs of
One (ValueScalar (ScalarString t)) ->
ContextValueId <$> parseId' t
One ValueNull -> Right $ ContextValueId IdNull
One (ValueDictionary m) ->
ContextValueExpanded <$> parseTermDef m
_ -> Left $ SpecError InvalidTermDefinition ""
)
parseDocument :: Dictionary -> Either BereniceError UnknownObject
parseDocument = parseDict
where
parseObjectKey t =
case parseToken t of
TWeird w -> Left $ InvalidObjectKey w
TKeywordNC kw -> Right $ ObjectKeyKeyword kw
TRef ck -> Right $ ObjectKeyOther ck
parseDict m =
let (mc, m') = lookupAndDelete "@context" m
in UnknownObject
<$> traverse (traverse parseContextItem) mc
<*> M.traverseWithKey
(\ p mv ->
(,) <$> parseObjectKey p <*> traverse parseValue mv
)
m'
parseValue ValueNull = Right UnknownItemNull
parseValue (ValueScalar s) =
Right $ UnknownItemScalar $
case s of
ScalarString t -> TScalarString t $ parseToken t
ScalarNumber n -> TScalarNumber n
ScalarBool b -> TScalarBool b
parseValue (ValueDictionary m) = UnknownItemObject <$> parseDict m
-- Once we're done with the *parsing* step, producing an 'UnknownObject'
-- representing our JSON-LD document, we do the *tagging* step. In the tagging
-- step we do some of the steps required for expanding the document:
--
-- * Process contexts, attaching an active context to each object
-- * Running ref expansion on properties i.e. object keys
-- Here are some helper types we'll need, such as the ones describing an active
-- context:
data TermDefinitionType
= TermDefinitionTypeId
| TermDefinitionTypeVocab
| TermDefinitionTypeAbsolute (URI Gen)
data TermDefinition = TermDefinition
{ termTarget :: IdentKw
, termReverse :: Bool
, termTypeOrLang :: Maybe (Either TermDefinitionType ContextLanguage)
, termContainer :: Maybe Container
}
data ActiveContext = ActiveContext
{ activeBaseURI :: Maybe (URI Gen)
, activeVocabulary :: Maybe Identifier
, activeLanguage :: Maybe Text
, activeTerms :: HashMap Ref TermDefinition
}
data Identifier
= IdentURI (URI Gen)
| IdentBlank RelToken
deriving (Eq, Ord, Generic)
instance Hashable Identifier
data IdentKw
= IdentKwURI (URI Gen)
| IdentKwBlank RelToken
| IdentKwC KeywordC
| IdentKwNC KeywordNC
deriving (Eq, Ord, Generic)
instance Hashable IdentKw
data TaggedKey
= TaggedKeyKeyword ObjectKey KeywordNC
| TaggedKeyOther Ref Identifier
-- Now, here's the new structure we'll be producing:
data TaggedItem
= TaggedItemScalar TScalar
| TaggedItemNull
| TaggedItemObject TaggedObject (HashMap Text (Multi TaggedItem))
data TaggedObject = TaggedObject
{ taggedContext :: ActiveContext
, taggedKwMap :: HashMap KeywordNC (ObjectKey, Multi TaggedItem)
, taggedIdMap :: HashMap Identifier (NonEmpty (Ref, Multi TaggedItem))
}
-- Now, the functions for ref expansion (what the spec calls IRI expansion,
-- except we don't handle null or keyword inputs, which are simply expanded
-- into themselves). This includes expansion functions we'll use in later
-- steps.
uriConcat'' :: URI Gen -> RelNoAuth -> Either BereniceError (URI Gen)
uriConcat'' base label =
case uriConcat base label of
Nothing -> Left $ InvalidConcatenatedURI base label
Just uri -> Right uri
uriConcat' :: URI Gen -> RelNoAuth -> Either BereniceError Identifier
uriConcat' base label = IdentURI <$> uriConcat'' base label
blankConcat :: RelToken -> RelNoAuth -> Either BereniceError RelToken
blankConcat base label =
case relTokenFromRel label of
Nothing -> Left $ InvalidConcatenatedBlank base label
Just rt -> Right $ base <> rt
blankConcat' :: RelToken -> RelNoAuth -> Either BereniceError Identifier
blankConcat' base label = IdentBlank <$> blankConcat base label
resolveRel
:: Bool
-> ActiveContext
-> RelNoAuth
-> Either BereniceError Identifier
resolveRel docRel active rel =
case (docRel, activeBaseURI active) of
(True, Just b) ->
Right $ IdentURI $
resolveRelativeReference b $ relativeFromNoAuth rel
(_, _) ->
Left $ TermOrRelativeNotExpanded rel
expandTermOrRelative
:: Bool
-> Bool
-> ActiveContext
-> RelNoAuth
-> Either BereniceError Identifier
expandTermOrRelative docRel vocab active rel =
case (vocab, activeVocabulary active) of
(True, Just v) -> case v of
IdentURI b -> uriConcat' b rel
IdentBlank b -> blankConcat' b rel
_ -> resolveRel docRel active rel
expandCompact
:: Maybe (URI Gen)
-> CompactURI
-> ActiveContext
-> Either BereniceError Identifier
expandCompact muri c active =
case M.lookup (RefTerm $ compactURIPrefix c) $ activeTerms active of
Just td ->
case termTarget td of
IdentKwURI u ->
case expandCompactURI u c of
Nothing -> Left $ InvalidExpandedURI u c
Just v -> Right $ IdentURI v
IdentKwC _ -> Left $ KeywordAliasUsedAsPrefix c
Nothing ->
case muri of
Just uri -> Right $ IdentURI uri
Nothing -> Left $ NoTermDefForCompactRefPrefix c
expandRefNonVocab'
:: Bool -> Bool -> ActiveContext -> Ref -> Either BereniceError Identifier
expandRefNonVocab' docRel vocab active ref =
case ref of
RefURI ac -> case ac of
JustAbsolute u -> Right $ IdentURI u
JustCompact c -> expandCompact Nothing c active
AbsoluteOrCompact u c -> expandCompact (Just u) c active
RefBlank l -> Right $ IdentBlank l
RefTerm u -> expandTermOrRelative docRel vocab active u
-- | IRI expansion, when done not during context processing. The 'vocab'
-- parameter is assumed to be false, which means no keyword aliases.
--
-- * The null case isn't handled here (if null, just return null)
-- * The keyword case isn't handled here (if keyword, just return as is)
-- * 'vocab' is assumed to be false, so the term can't end up expanding to a
-- keyword
expandRefNonVocab :: Bool -> ActiveContext -> Ref -> Either BereniceError Identifier
expandRefNonVocab docRel = expandRefNonVocab' docRel False
-- | IRI expansion, when done not during context processing. The 'vocab'
-- parameter is assumed to be true, which means keyword aliases may be found
-- and expanded into keywords.
--
-- * The null case isn't handled here (if null, just return null)
-- * The keyword case isn't handled here (if keyword, just return as is)
-- * 'vocab' is assumed to be true, so the term may expand into a keyword
expandRefVocab :: Bool -> ActiveContext -> Ref -> Either BereniceError IdentKw
expandRefVocab docRel active ref =
case M.lookup ref $ activeTerms active of
Just td -> Right $ termTarget td
Nothing -> do
k <- expandRefNonVocab' docRel True active ref
return $ case k of
IdentURI u -> IdentKwURI u
IdentBlank b -> IdentKwBlank b
-- | IRI expansion, when done not during context processing.
--
-- * The null case isn't handled here (if null, just return null)
-- * The keyword case isn't handled here (if keyword, just return as is)
expandRef
:: Bool -> Bool -> ActiveContext -> Ref -> Either BereniceError IdentKw
expandRef docRel vocab active ref =
if vocab
then expandRefVocab docRel active ref
else do
k <- expandRefNonVocab docRel active ref
return $ case k of
IdentURI u -> IdentKwURI u
IdentBlank b -> IdentKwBlank b
-- | IRI expansion, when done during context processing.
--
-- * The null case isn't handled here (if null, just return null)
-- * The keyword case isn't handled here (if keyword, just return as is)
expandRef'
:: Context
-> HashMap Ref Bool
-> Bool
-> Bool
-> ActiveContext
-> Ref
-> Either BereniceError (IdentKw, ActiveContext, HashMap Ref Bool)
expandRef' local defined docRel vocab active ref = do
let updated v = createTermDefinition local ref v defined active
(a, d) <- case (M.lookup ref $ contextValues local, M.lookup ref defined) of
(Just v, Nothing) -> updated v
(Just v, Just False) -> updated v
_ -> pure (active, defined)
case (vocab, M.lookup ref $ activeTerms a) of
(True, Just td) -> Right $ t3 a d $ termTarget td
_ -> case ref of
RefURI ac -> case ac of
JustAbsolute u -> Right (IdentKwURI u, a, d)
JustCompact c ->
useContext a d (RefTerm $ compactURIPrefix c) $
fmap k'2kx . expandCompact Nothing c
AbsoluteOrCompact u c ->
useContext a d (RefTerm $ compactURIPrefix c) $
fmap k'2kx . expandCompact (Just u) c
RefBlank l -> Right (IdentKwBlank l, a, d)
RefTerm t ->
t3 a d . k'2kx <$> expandTermOrRelative docRel vocab a t
where
t3 y z x = (x, y, z)
k'2kx (IdentURI u) = IdentKwURI u
k'2kx (IdentBlank b) = IdentKwBlank b
useContext a d k expand =
let orig = t3 a d <$> expand a
updated v = do
(a', d') <- createTermDefinition local k v d a
t3 a' d' <$> expand a'
in case (M.lookup k $ contextValues local, M.lookup k d) of
(Just v, Nothing) -> updated v
(Just v, Just False) -> updated v
_ -> orig
-- Now come the functions for context processing, which is how we produce the
-- active contexts for all the objects:
createTermDefinition
:: Context
-> Ref
-> ContextValue
-> HashMap Ref Bool
-> ActiveContext
-> Either BereniceError (ActiveContext, HashMap Ref Bool)
createTermDefinition local key value defined active =
case M.lookup key defined of
Just True -> Right (active, defined)
Just False -> Left $ SpecError (CyclicURIMapping key) ""
Nothing ->
let defined' = M.insert key False defined
active' = active { activeTerms = M.delete key $ activeTerms active }
in case value of
ContextValueId i -> case i of
-- Algo step 6 says we should set the term definition
-- to null. But in the expansion algos, it seems
-- lookups just check whether a given term has a term
-- def, so no def and null def seem to mean the same
-- thing. So, instead of allowing some unnecessary
-- annoying null value in the active context, we simply
-- delete the term from the active context, because
-- that's the same result as mapping term to null.
IdNull -> Right (active', M.insert key True defined)
_ -> ctd active' defined' ExpandedTermDefinition
{ expandedIdOrReverse =
Left ExpandedTermDefinitionId
{ expandedId = Just i
, expandedContainer = Nothing
}
, expandedType = Nothing
, expandedLanguage = Nothing
}
ContextValueExpanded etd -> case expandedIdOrReverse etd of
Left etdi -> case expandedId etdi of
Just IdNull ->
Right (active', M.insert key True defined)
_ -> ctd active' defined' etd
_ -> ctd active' defined' etd
where
ctd a d etd = do
(mtyp, a2, d2) <- case expandedType etd of
Nothing -> Right (Nothing, a, d)
Just t -> case t of
TypeURI ac -> expand $ RefURI ac
TypeTerm t -> expand $ RefTerm t
TypeId -> Right (Just TermDefinitionTypeId, a, d)
TypeVocab -> Right (Just TermDefinitionTypeVocab, a, d)
TypeNull -> Left $ SpecError InvalidTypeMapping ""
let expand2 = expandRef' local d2 False True a2
case expandedIdOrReverse etd of
Right etdr -> do
(k, a3, d3) <- expand2 $ expandedReverse etdr
tt <- case k of
IdentKwURI _ -> Right k
IdentKwBlank _ -> Right k
IdentKwC _ -> Left $ SpecError InvalidURIMapping ""
IdentKwNC _ -> Left $ SpecError InvalidURIMapping ""
let td = TermDefinition
{ termTarget = tt
, termReverse = True
, termTypeOrLang = Left <$> mtyp
, termContainer =
case expandedContainerReverse etdr of
Nothing -> Nothing
Just c -> Just $ case c of
ContainerReverseNull -> ContainerNull
ContainerReverseSet -> ContainerSet
ContainerReverseIndex -> ContainerIndex
}
Right
( a3
{ activeTerms =
M.insert key td $ activeTerms a3
}
, M.insert key True d3
)
Left etdi -> do
(tt, a3, d3) <- case (id &&& eq key) <$> expandedId etdi of
Just (i, False) -> do
(k, a3', d3') <- case i of
IdNull -> Left $ SpecError InvalidURIMapping ""
IdRef ck -> expand2 ck
IdKeyword kw -> case kw of
KeywordContext -> Left $ SpecError InvalidKeywordAlias ""
KeywordC kwc -> Right (IdentKwC kwc, a2, d2)
KeywordNC kwnc -> Right (IdentKwNC kwnc, a2, d2)
Right (k, a3', d3')
_ -> case key of
RefURI ac -> case ac of
JustAbsolute u -> Right (IdentKwURI u, a2, d2)
JustCompact c -> case M.lookup (RefTerm $ compactURIPrefix c) $ contextValues local of
Nothing -> Left $ NoTermDefForCompactRefPrefix c
Just v -> do
(a3', d3') <- createTermDefinition local (RefTerm $ compactURIPrefix c) v d2 a2
case termTarget <$> M.lookup (RefTerm $ compactURIPrefix c) (activeTerms a3') of
Just tt -> case tt of
IdentKwURI u -> t3 a3' d3' . IdentKwURI <$> expandCompactURI' u c
where
expandCompactURI' u c =
case expandCompactURI u c of
Nothing -> Left $ InvalidExpandedURI u c
Just v -> Right v
IdentKwC _ -> Left $ KeywordAliasUsedAsPrefix c
Nothing -> Left $ NoTermDefForCompactRefPrefix c
AbsoluteOrCompact u c -> case M.lookup (RefTerm $ compactURIPrefix c) $ contextValues local of
Nothing -> Right (IdentKwURI u, a2, d2)
Just v -> do
(a3', d3') <- createTermDefinition local (RefTerm $ compactURIPrefix c) v d2 a2
case termTarget <$> M.lookup (RefTerm $ compactURIPrefix c) (activeTerms a3') of
Just tt -> case tt of
IdentKwURI w -> t3 a3' d3' . IdentKwURI <$> expandCompactURI' w c
where
expandCompactURI' u c =
case expandCompactURI u c of
Nothing -> Left $ InvalidExpandedURI u c
Just v -> Right v
IdentKwC _ -> Left $ KeywordAliasUsedAsPrefix c
Nothing -> Left $ NoTermDefForCompactRefPrefix c
RefBlank b -> Right (IdentKwBlank b, a2, d2)
RefTerm t -> case activeVocabulary a2 of
Just (IdentURI u) ->
t3 a2 d2 . IdentKwURI <$> uriConcat'' u t
Just (IdentBlank b) ->
t3 a2 d2 . IdentKwBlank <$> blankConcat b t
Nothing -> Left $ SpecError InvalidURIMapping ""
container <- case expandedContainer etdi of
Nothing -> Right Nothing
Just c -> case c of
ContainerNull -> Left $ SpecError InvalidContainerMapping ""
_ -> Right $ Just c
let td = TermDefinition
{ termTarget = tt
, termReverse = False
, termTypeOrLang = case (mtyp, expandedLanguage etd) of
(Nothing, Just cl) -> Just $ Right cl
(Nothing, Nothing) -> Nothing
(Just typ, _) -> Just $ Left typ
, termContainer = container
}
Right
( a3
{ activeTerms =
M.insert key td $ activeTerms a3
}
, M.insert key True d3
)
where
t3 y z x = (x, y, z)
expand r = do
(k, a', d') <- expandRef' local d False True a r
case k of
IdentKwURI u ->
Right (Just $ TermDefinitionTypeAbsolute u, a', d')
IdentKwNC KeywordId ->
Right (Just $ TermDefinitionTypeId, a', d')
IdentKwC KeywordVocab ->
Right (Just $ TermDefinitionTypeVocab, a', d')
_ -> Left $ SpecError InvalidTypeMapping ""
eq ck (IdRef ck') = ck == ck'
eq _ _ = False
updateActiveContext
:: Monad m
=> (URI Gen -> ExceptT BereniceError m (Multi Dictionary))
-> URI Gen
-> Multi LocalContextItem
-> ActiveContext
-> ExceptT BereniceError m ActiveContext
updateActiveContext fetch base = go S.empty
where
liftE = ExceptT . pure
updateBase
:: Bool
-> Maybe ContextBase
-> ActiveContext
-> Either BereniceError ActiveContext
updateBase True (Just base) active =
case base of
ContextBaseNull -> Right active { activeBaseURI = Nothing }
ContextBaseAbsolute uri -> Right active { activeBaseURI = Just uri }
ContextBaseRelative uri ->
case activeBaseURI active of
Nothing -> Left $ SpecError InvalidBaseURI ""
Just bu ->
let uri' = resolveRelativeReference bu uri
in Right active { activeBaseURI = Just uri' }
updateBase _ _ active = Right active
updateVocab
:: Maybe ContextVocab
-> ActiveContext
-> Either BereniceError ActiveContext
updateVocab Nothing active = Right active
updateVocab (Just vocab) active =
case vocab of
ContextVocabNull -> Right active { activeVocabulary = Nothing }
ContextVocabAbsolute uri -> Right active { activeVocabulary = Just $ IdentURI uri }
ContextVocabBlank label -> Right active { activeVocabulary = Just $ IdentBlank label }
updateLanguage
:: Maybe ContextLanguage
-> ActiveContext
-> Either BereniceError ActiveContext
updateLanguage Nothing active = Right active
updateLanguage (Just lang) active =
case lang of
ContextLanguageNull -> Right active { activeLanguage = Nothing }
ContextLanguageTag tag -> Right active { activeLanguage = Just tag }
go rs ls a = foldlM (flip $ go' rs) a ls
go' remotes local active =
case local of
LocalContextNull -> pure $ active { activeBaseURI = Just base }
LocalContextString u -> do
let base' = fromMaybe base $ activeBaseURI active
uAbs = resolveURIReference base' u
b = renderURI uAbs
when (b `S.member` remotes) $
throwE $ SpecError RecursiveContextInclusion ""
ls <- do
dicts <- fetch uAbs
liftE $ case dicts of
Many _ -> Left $ SpecError InvalidRemoteContext ""
One dict -> case M.lookup "@context" dict of
Nothing -> Left $ SpecError InvalidRemoteContext ""
Just vals -> traverse parseContextItem vals
go (S.insert b remotes) ls active
LocalContextObject c -> ExceptT $ pure $
updateBase (S.null remotes) (contextBase c) active
>>= updateVocab (contextVocab c)
>>= updateLanguage (contextLanguage c)
>>= makeTerms c
makeTerms c a =
fst <$> foldrM (ctd c) (a, M.empty) (M.toList $ contextValues c)
ctd c (k, v) (a, d) = createTermDefinition c k v d a
-- Finally, the functions for the tagging step, transforming our document from
-- 'UnknownObject' to 'TaggedObject':
tagDocument
:: Monad m
=> (URI Gen -> ExceptT BereniceError m (Multi Dictionary))
-> URI Gen
-> UnknownObject
-> ExceptT BereniceError m TaggedObject
tagDocument fetch base = fmap fst . tagObject initialActiveContext
where
initialActiveContext = ActiveContext Nothing Nothing Nothing M.empty
expandRefKW _ _ _ (ObjectKeyKeyword kw) =
Right $ TaggedKeyKeyword (ObjectKeyKeyword kw) kw
expandRefKW dr v a (ObjectKeyOther ck) = do
k <- expandRef dr v a ck
case k of
IdentKwURI u -> Right $ TaggedKeyOther ck $ IdentURI u
IdentKwBlank t -> Right $ TaggedKeyOther ck $ IdentBlank t
IdentKwC kwc -> Left $ InvalidTaggedKey kwc
IdentKwNC kwnc ->
Right $ TaggedKeyKeyword (ObjectKeyOther ck) kwnc
tagItem active i = case i of
UnknownItemScalar s -> return $ TaggedItemScalar s
UnknownItemNull -> return TaggedItemNull
UnknownItemObject o -> uncurry TaggedItemObject <$> tagObject active o
tagObject initialActive (UnknownObject mlocal initialValues) = do
active <- case mlocal of
Nothing -> return initialActive
Just local -> updateActiveContext fetch base local initialActive
taggedValues <- for initialValues $ \ (k, v) -> do
tk <- ExceptT . pure $ expandRefKW False True active k
tv <- traverse (tagItem active) v
return (tk, tv)
let (kws, ids) =
partitionEithers $ map (uncurry decide) $ M.elems taggedValues
kws' <- traverse detectDup $ M.fromListWith combineKws kws
let ids' = M.fromListWith (<>) ids
limap = M.map snd taggedValues
return (TaggedObject active kws' ids', limap)
where
decide (TaggedKeyKeyword ok kw) tis = Left (kw, (False, (ok, tis)))
decide (TaggedKeyOther ref i) tis = Right (i, (ref, tis) :| [])
combineKws (_, new) (_, old) = (True, new)
detectDup (True , _) = throwE $ SpecError CollidingKeywords ""
detectDup (False, tv) = pure tv
-- The next step is specialization. For each JSON object, we detect its type:
-- Node object, or value object, or list object, and so on.
--
-- First, new types we'll need:
data ValueValue
= ValueValueScalar TScalar
| ValueValueNull
-- | A type for use in values associated with keywords, allowing to attach
-- keyword aliases to them.
--
-- It's basically like turning type @a@ into @(Maybe Ref, a)@ where the @fst@
-- is the alias. It sounds trivial, and indeed it is. I'm just using it to make
-- it clearer that the @Maybe Ref@ is an alias. Sadly, even in this
-- type-expressive Haskell library, JSON-LD is a scary mess and horror, so I
-- feel like using some helper types makes things a little bit less scary.
data AliasAnd a = AliasAnd
{ aaAlias :: Maybe Ref
, aaValue :: a
}
data ValueObject = ValueObject
{ valueValue :: AliasAnd ValueValue
, valueTypeOrLang :: Maybe (AliasAnd (Either Ref Text))
, valueIndex :: Maybe (AliasAnd Text)
, valueContext :: ActiveContext
}
data NodeReverseItem
= NodeReverseId Ref
| NodeReverseObject NodeObject
{-
-- TODO move this type elsewhere if it's used somewhere above
data RefMap a
= RefMapEmpty Ref
| RefMap (NonEmpty (Ref, a))
-}
data NodeReverse = NodeReverse
{ reverseContext :: ActiveContext
, reverseMap ::
HashMap Identifier (NonEmpty (Ref, Multi NodeReverseItem))
-- ^ Each NonEmpty has a unique Ref, and each Ref appears in one NonEmpty,
-- under a single Identifier, can't appear under different Identifiers
-- because each Ref has a single Identifier matching it via IRI expansion
}
data Item
= ItemScalar TScalar
| ItemNull
| ItemNode NodeObject
| ItemValue ValueObject
-- These are the non expanded versions
data ListObject' = ListObject'
{ listArray' :: AliasAnd (Multi Item)
, listContext' :: ActiveContext
, listIndex' :: Maybe (AliasAnd Text)
}
data SetObject' = SetObject'
{ setArray' :: AliasAnd (Multi Item)
, setContext' :: ActiveContext
, setIndex' :: Maybe (AliasAnd Text)
}
data NodeItem
= NodeItemOne Item
| NodeItemList ListObject'
| NodeItemSet SetObject'
-- ^ Same as below, using Vector for now as long as no need for lookups
data LanguageItem
= LanguageItemNull
| LanguageItemString Text
data NodeValue
= Items (Multi NodeItem)
-- ^ by default arrays are unordered in JSON-LD, but I'm using Vector to
-- avoid hashing that HashSet would need, especially since probably I won't
-- need to do lookups. And if I do, I guess I'll switch to HashSet
| LangMap (HashMap Text (Multi LanguageItem))
| IndexMap (HashMap Text (Multi NodeItem))
data NodeObject = NodeObject
{ nodeContext :: ActiveContext
, nodeId :: Maybe (AliasAnd Ref)
, nodeGraph :: Maybe (AliasAnd (Multi NodeObject))
, nodeType :: Maybe (AliasAnd (Multi Ref))
, nodeReverse :: Maybe (AliasAnd NodeReverse)
, nodeIndex :: Maybe (AliasAnd Text)
, nodeValues ::
HashMap
Identifier
(NonEmpty (Ref, (Either (Multi NodeItem) NodeValue)))
-- ^ Each NonEmpty has a unique Ref, I could use a HashMap but wondering if
-- that helps anything if I won't really be doing any lookups in it, just
-- compacting or expanding into other structures.
--
-- 'Left' means that the unexpanded key's container mapping is @list,
-- otherwise 'Right'.
--
-- TODO perhaps change that type? Either have a dedicated datatype or just
-- use 'NodeValue' with additional 'Bool' field in its 'Items' ctor to say
-- whether container is '@list' or not.
}
data DetectedObject
= DetectedValueObject ValueObject
| DetectedListObject ListObject'
| DetectedSetObject SetObject'
| DetectedNodeObject NodeObject
-- And now the functions for the specialization step:
specializeDocument :: TaggedObject -> Either BereniceError NodeObject
specializeDocument = specializeNode
where
alias (ObjectKeyKeyword _) = Nothing
alias (ObjectKeyOther ref) = Just ref
withAlias act (ok, tv) = AliasAnd (alias ok) <$> act tv
aliasAnd ok act = AliasAnd (alias ok) <$> act
specializeNode (TaggedObject active kws vals) =
NodeObject
<$> pure active
<*> (for (M.lookup KeywordId kws) $ withAlias $ \ tv -> case tv of
One (TaggedItemScalar (TScalarString _ token)) -> case token of
TWeird w -> Left $ NodeInvalidId $ Right w
TKeywordNC kw -> Left $ NodeInvalidId $ Left kw
TRef ck -> Right ck
_ -> Left $ SpecError InvalidIdValue ""
)
<*> (for (M.lookup KeywordGraph kws) $ withAlias $ \ tv ->
for tv $ \ ti -> case ti of
TaggedItemObject to _ -> do
o <- specializeObject to
case o of
DetectedNodeObject no -> Right no
_ -> Left NodeInvalidGraph
_ -> Left NodeInvalidGraph
)
<*> (for (M.lookup KeywordType kws) $ withAlias $ \ tv ->
for tv $ \ ti -> case ti of
TaggedItemScalar (TScalarString _ token) ->
case token of
TWeird w -> Left $ NodeInvalidType $ Right w
TKeywordNC kw -> Left $ NodeInvalidType $ Left kw
TRef ck -> Right ck
_ -> Left $ SpecError InvalidTypeValue ""
)
<*> (for (M.lookup KeywordReverse kws) $ withAlias $ \ tv -> case tv of
One (TaggedItemObject (TaggedObject a k m) _) ->
if M.null k
then fmap (NodeReverse a) $ for m $ \ ne -> for ne $ \ (ref, tv) -> fmap ((,) ref) $
for tv $ \ ti -> case ti of
TaggedItemScalar (TScalarString _ token) -> case token of
TWeird w -> Left $ NodeInvalidReverse $ Right w
TKeywordNC kw -> Left $ NodeInvalidReverse $ Left kw
TRef ck -> Right $ NodeReverseId ck
TaggedItemObject to _ -> do
o <- specializeObject to
case o of
DetectedNodeObject no -> Right $ NodeReverseObject no
_ -> Left $ SpecError InvalidReversePropertyValue ""
_ -> Left $ SpecError InvalidReversePropertyValue ""
else Left $ SpecError InvalidReversePropertyMap ""
_ -> Left $ SpecError InvalidReverseValue ""
)
<*> (for (M.lookup KeywordIndex kws) $ withAlias $ \ tv -> case tv of
One (TaggedItemScalar (TScalarString t _)) -> Right t
_ -> Left $ SpecError InvalidIndexValue ""
)
<*> (for vals $ \ ne -> for ne $ \ (ref, tv) -> (,) ref <$>
let ti2ni ti = case ti of
TaggedItemScalar s -> Right $ NodeItemOne $ ItemScalar s
TaggedItemNull -> Right $ NodeItemOne $ ItemNull
TaggedItemObject to _ -> do
o <- specializeObject to
Right $ case o of
DetectedValueObject vo -> NodeItemOne $ ItemValue vo
DetectedListObject lo -> NodeItemList lo
DetectedSetObject so -> NodeItemSet so
DetectedNodeObject no -> NodeItemOne $ ItemNode no
in case (M.lookup ref (activeTerms active) >>= termContainer, tv) of
(Just ContainerList, _) -> Left <$> traverse ti2ni tv
(Just ContainerLanguage, One (TaggedItemObject _ m)) -> fmap (Right . LangMap) $ for m $ \ mti -> for mti $ \ ti -> case ti of
TaggedItemScalar (TScalarString t _) -> Right $ LanguageItemString t
TaggedItemNull -> Right LanguageItemNull
_ -> Left $ SpecError InvalidLanguageMapValue ""
(Just ContainerIndex, One (TaggedItemObject _ m)) -> Right . IndexMap <$> traverse (traverse ti2ni) m
_ -> Right . Items <$> traverse ti2ni tv
)
specializeObject tobj@(TaggedObject active kws vals) =
case M.lookup KeywordValue kws of
Just (ok, tval) -> fmap DetectedValueObject $
ValueObject
<$> (aliasAnd ok $ case tval of
One ti -> case ti of
TaggedItemScalar s -> Right $ ValueValueScalar s
TaggedItemNull -> Right ValueValueNull
TaggedItemObject _ _ -> Left $ SpecError InvalidValueObjectValue ""
Many _ -> Left $ SpecError InvalidValueObjectValue ""
)
<*> case ( M.lookup KeywordType kws
, M.lookup KeywordLanguage kws) of
(Nothing, Nothing) -> Right Nothing
(Just (ok, tv), Nothing) -> fmap (Just . AliasAnd (alias ok) . Left) $ tv2vt tv
where
tv2vt (One (TaggedItemScalar (TScalarString _ token))) =
case token of
TWeird w -> Left $ ValueInvalidType $ Right w
TKeywordNC kw -> Left $ ValueInvalidType $ Left kw
TRef ck -> Right ck
tv2vt (One _) =
Left $ SpecError InvalidTypeValue ""
tv2vt (Many _) =
Left $ SpecError InvalidTypedValue ""
(Nothing, Just (ok, tv)) -> case tv of
One (TaggedItemScalar (TScalarString t _)) ->
Right $ Just $ AliasAnd (alias ok) $ Right t
_ ->
Left $ SpecError InvalidLanguageTaggedString ""
(Just _, Just _) -> Left $ SpecError InvalidValueObject ""
<*> case M.lookup KeywordIndex kws of
Nothing -> Right Nothing
Just (ok, tv) -> case tv of
One (TaggedItemScalar (TScalarString t _)) ->
Right $ Just $ AliasAnd (alias ok) t
_ ->
Left $ SpecError InvalidIndexValue ""
<*> pure active
Nothing -> case M.lookup KeywordList kws of
Just p -> DetectedListObject <$> detectListSet True ListObject' p
Nothing -> case M.lookup KeywordSet kws of
Just p -> DetectedSetObject <$> detectListSet False SetObject' p
Nothing -> DetectedNodeObject <$> specializeNode tobj
where
detectListSet list f (ok, tv) = f
<$> (aliasAnd ok $ for tv $ \ ti -> case ti of
TaggedItemScalar s -> Right $ ItemScalar s
TaggedItemNull -> Right ItemNull
TaggedItemObject to _ -> do
o <- specializeObject to
case (list, o) of
(_, DetectedValueObject vo) ->
Right $ ItemValue vo
(_, DetectedNodeObject no) ->
Right $ ItemNode no
(True, _) -> Left $ SpecError ListOfLists ""
(False, _) -> Left ListInvalidMember
)
<*> pure active
<*> (for (M.lookup KeywordIndex kws) $ withAlias $ \ tv -> case tv of
One (TaggedItemScalar (TScalarString t _)) -> Right t
_ -> Left $ SpecError InvalidIndexValue ""
)
-- The next step is expansion. In the tagging step we did context processing
-- and ref expansion on properties, and now we'll do the rest of the expansion
-- algorithm.
--
-- First, some helper types:
data ExpandedScalar
= ExpandedScalarId Identifier
| ExpandedScalarVal ExpandedScalarValue
deriving Eq
data ExpandedReverseObject = ExpandedReverseObject
{ expandedReverseReverse :: HashMap Identifier (Vector ExpandedNodeObject)
, expandedReverseValues :: HashMap Identifier (Vector ExpandedNodeObject)
}
data ActiveProperty
= ActivePropertyNull
| ActivePropertyURI AbsoluteOrCompact
| ActivePropertyBlank RelToken
| ActivePropertyTerm RelNoAuth
| ActivePropertyKeyword KeywordNC
-- Now, here are the types for the new structure we'll be producing:
data ExpandedStringValue = ExpandedStringValue
{ esText :: Text
, esToken :: Token
, esTypeOrLang :: Maybe (Either (URI Gen) Text)
}
deriving Eq
data ExpandedNBValue = ExpandedNBValue
{ enbValue :: Either Scientific Bool
, enbType :: Maybe (URI Gen)
}
deriving Eq
data ExpandedScalarValue = ExpandedScalarValue
{ scalarValue :: Either ExpandedNBValue ExpandedStringValue
, scalarIndex :: Maybe Text
}
deriving Eq
data ExpandedAtom
= ExpandedAtomScalar ExpandedScalarValue
| ExpandedAtomNode ExpandedNodeObject
deriving Eq
data ExpandedList = ExpandedList
{ elistIndex :: Maybe Text
, elistArray :: Vector ExpandedAtom
}
deriving Eq
data ExpandedItem
= ExpandedItemOne ExpandedAtom
| ExpandedItemList ExpandedList
deriving Eq
-- "This algorithm expands a JSON-LD document, such that all context
-- definitions are removed, all terms and compact IRIs are expanded to absolute
-- IRIs, blank node identifiers, or keywords and all JSON-LD values are
-- expressed in arrays in expanded form."
data ExpandedNodeObject = ExpandedNodeObject
{ enodeId :: Maybe Identifier
, enodeGraph :: Maybe (Multi ExpandedNodeObject)
-- ^ TODO do we need the distinction between one and many? If we really do,
-- replace this Either with a Multi. Otherwise, replace with a Vector.
, enodeType :: Maybe (Vector Identifier)
, enodeReverse :: Maybe (HashMap Identifier (Vector ExpandedNodeObject))
, enodeIndex :: Maybe Text
, enodeValues :: HashMap Identifier (Vector ExpandedItem)
}
deriving Eq
-- And now the functions:
ck2ap (RefURI ac) = ActivePropertyURI ac
ck2ap (RefBlank t) = ActivePropertyBlank t
ck2ap (RefTerm t) = ActivePropertyTerm t
{-
ap2ck ActivePropertyNull = Nothing
ap2ck (ActivePropertyURI ac) = Just $ RefURI ac
ap2ck (ActivePropertyBlank t) = Just $ RefBlank t
ap2ck (ActivePropertyTerm t) = Just $ RefTerm t
ap2ck (ActivePropertyKeyword _) = Nothing
-}
idOrValue
:: ActiveContext
-> Ref
-> Either Bool (Maybe (Either (URI Gen) ContextLanguage))
idOrValue active property =
let mtorl = M.lookup property (activeTerms active) >>= termTypeOrLang
in case mtorl of
Just torl -> case torl of
Left tt -> case tt of
TermDefinitionTypeId -> Left False
TermDefinitionTypeVocab -> Left True
TermDefinitionTypeAbsolute u -> Right $ Just $ Left u
Right tl -> Right $ Just $ Right tl
Nothing -> Right Nothing
expandStringToId
:: Bool -> ActiveContext -> Ref -> Either BereniceError Identifier
expandStringToId vocab active ck = do
k <- expandRef True vocab active ck
case k of
IdentKwURI u -> Right $ IdentURI u
IdentKwBlank b -> Right $ IdentBlank b
IdentKwC kwc -> Left $ ENodeInvalidId $ Right $ WKeywordC kwc
IdentKwNC kwnc -> Left $ ENodeInvalidId $ Left kwnc
k2eno k = ExpandedNodeObject (Just k) Nothing Nothing Nothing Nothing M.empty
expandReverseItem
:: ActiveContext
-> Ref
-> NodeReverseItem
-> Either BereniceError ExpandedNodeObject
expandReverseItem active property nri =
case nri of
NodeReverseId ck ->
case idOrValue active property of
Left vocab -> k2eno <$> expandStringToId vocab active ck
Right _ -> Left $ SpecError InvalidReversePropertyValue ""
NodeReverseObject n ->
-- The active property can be neither null nor a keyword, so
-- specifically not @graph. Therefore we can use
-- 'expandNodeObjectNotNull' below, knowing the check that may
-- result with null isn't supposed to happen.
expandNodeObjectNotNull (ck2ap property) n
expandReverseObject
:: NodeReverse
-> Either BereniceError ExpandedReverseObject
expandReverseObject reverse = do
-- 1-4 - It's a JSON object
-- 5 - We already did that in tagDocument
let active = reverseContext reverse
-- 6-7
initial = ExpandedReverseObject M.empty M.empty
foldlM (processProperty' active) initial (M.toList $ reverseMap reverse)
-- 8-11 - There are no keys that are keywords
-- 12 - Active property is @reverse, so it's neither null nor @graph
where
processProperty' active obj (ep, vals) =
foldlM (processProperty active ep) obj vals
processProperty active ep (ExpandedReverseObject rrs rvs) (k, v) = do
-- 7.1 - We already handled @context separately
-- 7.2 - We already did that in tagDocument
-- 7.3 - No such option, expandRef returns only keywords or things with
-- a colon
-- 7.4 - We already did this in specializeDocument
let mtd = M.lookup k $ activeTerms active
mc = mtd >>= termContainer
ev <- case (mc, isObj v) of
-- 7.5 - The value of a reverse property has to point to a
-- node, it can't be just a scalar. Instead of waiting for
-- later when the expansion of the containing node object
-- discovers this, we exclude the possibility of a language map
-- from the datatype for reverse property value, and just raise
-- the error here.
(Just ContainerLanguage, True) -> Left $ SpecError InvalidReversePropertyValue ""
-- 7.6 - Same as above
(Just ContainerIndex, True) -> Left $ SpecError InvalidReversePropertyValue ""
-- 7.7
_ -> traverse (expandReverseItem active k) v
-- 7.8 - It can't be null
case mc of
-- 7.9 - The value of a reverse property (at least unless the
-- reverse property is @reverse, the spec algo mentions this case,
-- anyway this is not the case because we already raised an error
-- on the property being a keyword) mustn't be a value object or
-- list object. If we add it here, the expansion of the node object
-- (that contains the reverse object we're expanding) will discover
-- it and given an error. Instead, we're raising the error early.
Just ContainerList -> Left $ SpecError InvalidReversePropertyValue ""
_ -> Right $ case termReverse <$> mtd of
-- 7.10
Just True ->
case ev of
One n ->
let upd Nothing = Just $ V.singleton n
upd (Just l) = Just $ V.cons n l
in ExpandedReverseObject (M.alter upd ep rrs) rvs
Many ns ->
ExpandedReverseObject
(M.insertWith (<>) ep ns rrs)
rvs
-- 7.11 - The spec is unclear, what does it mean to append?
-- Appending a scalar to an array is clear, but what if we have
-- an array to append to an array? Do we add it as a new single
-- array item, or do we (++) the two arrays? It seems the JS
-- impementation does the latter. Let's do the same.
_ -> case ev of
One n ->
let upd Nothing = Just $ V.singleton n
upd (Just v) = Just $ V.cons n v
in ExpandedReverseObject rrs $ M.alter upd ep rvs
Many ns ->
ExpandedReverseObject rrs (M.insertWith (<>) ep ns rvs)
isObj (One (NodeReverseObject _)) = True
isObj _ = False
nbOrString :: TScalar -> Either (Either Scientific Bool) (Text, Token)
nbOrString (TScalarString t token) = Right (t, token)
nbOrString (TScalarNumber s) = Left $ Left s
nbOrString (TScalarBool b) = Left $ Right b
expandValue
:: ActiveContext -> Ref -> TScalar -> Either BereniceError ExpandedScalar
expandValue active property value =
case idOrValue active property of
Left vocab -> case value of
TScalarString _ token ->
case token of
TWeird w -> Left $ ENodeInvalidId $ Right w
TKeywordNC kw -> Left $ ENodeInvalidId $ Left kw
TRef ck ->
ExpandedScalarId <$> expandStringToId vocab active ck
_ -> Left $ ExpandValueSetIdToNonString active property value
Right m -> Right $ ExpandedScalarVal ExpandedScalarValue
{ scalarValue = case nbOrString value of
Left nb -> Left $ ExpandedNBValue nb $
case m of
Just (Left u) -> Just u
_ -> Nothing
Right (t, token) -> Right $ ExpandedStringValue t token $
case m of
Just uorl -> case uorl of
Left u -> Just $ Left u
Right tl -> case tl of
ContextLanguageTag t -> Just $ Right t
ContextLanguageNull -> Nothing
Nothing -> Right <$> activeLanguage active
, scalarIndex = Nothing
}
expandValueObject
:: ValueObject
-> Either BereniceError (Maybe ExpandedScalarValue)
expandValueObject (ValueObject valueA mtorlA mindexA active) = do
let value = aaValue valueA
mtorl = aaValue <$> mtorlA
mindex = aaValue <$> mindexA
-- 1-4 - It's a JSON object
-- 5 - We already did this in tagDocument
-- 6-7 - No need to loop, just handle the few keys relevant to value object
-- 7.4 - Yes, all properties are keywords
-- 7.5-7.11 - Irrelevant, we're only handling keywords
for (v2m value) $ \ s -> do
mtorl' <- for mtorl $ \ torl -> case torl of
Left typ -> do
ik <- expandRefVocab True active typ
case ik of
IdentKwURI u -> Right $ Left u
_ -> Left $ SpecError InvalidTypedValue ""
Right lang -> Right $ Right lang
value' <- case nbOrString s of
Left nb -> Left . ExpandedNBValue nb <$>
case mtorl' of
Nothing -> Right Nothing
Just torl -> case torl of
Left u -> Right $ Just u
Right _ -> Left $ SpecError InvalidLanguageTaggedValue ""
Right (t, token) -> Right $ Right $ ExpandedStringValue t token mtorl'
Right $ ExpandedScalarValue value' mindex
where
v2m (ValueValueScalar s) = Just s
v2m ValueValueNull = Nothing
expandItem
:: ActiveContext
-> Ref
-> Item
-> Either BereniceError (Maybe ExpandedAtom)
expandItem _ _ ItemNull = Right Nothing
expandItem _ property (ItemNode n) =
fmap ExpandedAtomNode <$> expandNodeObject (ck2ap property) n
expandItem _ _ (ItemValue v) =
fmap ExpandedAtomScalar <$> expandValueObject v
expandItem active property (ItemScalar s) = do
es <- expandValue active property s
return $ Just $
case es of
ExpandedScalarId k -> ExpandedAtomNode $ k2eno k
ExpandedScalarVal v -> ExpandedAtomScalar v
toArray :: Multi a -> Vector a
toArray (One x) = V.singleton x
toArray (Many v) = v
expandListObject
:: Ref
-> ListObject'
-> Either BereniceError ExpandedList
expandListObject property (ListObject' items active mindex) =
-- 1-4 - It's a JSON object
-- 5 - We already did context processing in the tagging step
-- 6-7 - Handle the specific possible keywords
ExpandedList (aaValue <$> mindex) . V.mapMaybe id <$>
traverse (expandItem active property) (toArray $ aaValue items)
expandSetObject
:: Ref
-> SetObject'
-> Either BereniceError (Vector ExpandedAtom)
expandSetObject property (SetObject' items active _mindex) =
-- 1-4 - It's a JSON object
-- 5 - We already did context processing in the tagging step
-- 6-7 - Handle the specific possible keywords
V.mapMaybe id <$> traverse (expandItem active property) (toArray $ aaValue items)
expandLangMap
:: HashMap Text (Multi LanguageItem)
-> Either BereniceError (Vector ExpandedAtom)
expandLangMap lm =
let lv2t LanguageItemNull = Nothing
lv2t (LanguageItemString t) = Just t
lv2v (One li) = V.singleton <$> lv2t li
lv2v (Many v) = traverse lv2t v
in case traverse lv2v lm of
Nothing -> Left $ SpecError InvalidLanguageMapValue ""
Just lm' ->
let t2i tag val = ExpandedAtomScalar ExpandedScalarValue
{ scalarValue = Right ExpandedStringValue
{ esText = val
, esToken = parseToken val
, esTypeOrLang = Just $ Right tag
}
, scalarIndex = Nothing
}
p2i (tag, vals) = V.map (t2i tag) vals
in return $ V.concat $ map p2i $ M.toList lm'
expandIndexMap
:: ActiveContext
-> Ref
-> HashMap Text (Multi NodeItem)
-> Either BereniceError (Vector ExpandedItem)
expandIndexMap active property im = do
let e2v (Left ei) = V.singleton ei
e2v (Right eas) = V.map ExpandedItemOne eas
im' = M.map toArray im
expand = expandNodeItem active property
fmap V.concat $ for (M.toList im') $ \ (index, items) -> do
-- We need to expand the items array. Let's do this
-- according to step 3 of this algorith. We already
-- know that the active property we'll be using, k,
-- isn't a keyword, and that the container mapping
-- is @index, not @list, so we can skip step 3.2.2.
-- Let's do the rest of what it says though.
items' <- traverse expand items
let items'' = fold $ V.map e2v $ V.mapMaybe id items'
-- Supposed to add @index-index mapping to the
-- expanded node item!! If it doesn't already
-- contain one
addIndex ind ei = case ei of
ExpandedItemOne a -> ExpandedItemOne $ case a of
ExpandedAtomScalar s -> ExpandedAtomScalar s
{ scalarIndex =
Just $ fromMaybe ind $ scalarIndex s
}
ExpandedAtomNode n -> ExpandedAtomNode n
{ enodeIndex =
Just $ fromMaybe ind $ enodeIndex n
}
ExpandedItemList (ExpandedList mi a) ->
let mi' = Just $ fromMaybe ind mi
in ExpandedItemList $ ExpandedList mi' a
return $ V.map (addIndex index) items''
-- TODO since we have special handling for properties that are keywords,
-- doesn't that mean that ActiveProperty can never be a keyword? And if null
-- and @graph are always checked together, perhaps we can unite them? And if
-- the specific value isn't used, we can unite all Ref APs into a single
-- param-less ctor?
expandNodeItem
:: ActiveContext
-> Ref
-> NodeItem
-> Either BereniceError (Maybe (Either ExpandedItem (Vector ExpandedAtom)))
expandNodeItem active property (NodeItemOne i) =
fmap (Left . ExpandedItemOne) <$> expandItem active property i
expandNodeItem _ property (NodeItemList lo) =
Just . Left . ExpandedItemList <$> expandListObject property lo
expandNodeItem _ property (NodeItemSet so) =
Just . Right <$> expandSetObject property so
-- | Expand a value associated with a non-keyword key of a node object. This
-- version is for the case the active property's container mapping isn't @list.
expandNodeValueNonLO
:: ActiveContext
-> Ref
-> NodeValue
-> Either BereniceError (Maybe (Multi ExpandedItem))
expandNodeValueNonLO active property nv = case nv of
Items (One ni) -> do
mr <- expandNodeItem active property ni
return $ mr <&> \ r -> case r of
Left ei -> One ei
Right eas -> Many $ V.map ExpandedItemOne eas
Items (Many nis) ->
fmap (Just . Many . fold . V.mapMaybe id) $ for nis $ \ ni -> do
mr <- expandNodeItem active property ni
return $ mr <&> \ r -> case r of
Left ei -> V.singleton ei
Right eas -> V.map ExpandedItemOne eas
LangMap lm -> Just . Many . V.map ExpandedItemOne <$> expandLangMap lm
IndexMap im -> Just . Many <$> expandIndexMap active property im
-- | Expand a value associated with a non-keyword key of a node object. This
-- version is for the case the active property's container mapping is @list.
expandNodeValueForLO
:: ActiveContext
-> Ref
-> Multi NodeItem
-> Either BereniceError (Maybe (Either ExpandedItem (Vector ExpandedAtom)))
expandNodeValueForLO active property nv = case nv of
One ni -> expandNodeItem active property ni
Many nis -> do
fmap (Just . Right . V.mapMaybe id) $ for nis $ \ ni -> do
mr <- expandNodeItem active property ni
for mr $ \ r -> case r of
Left ei -> case ei of
ExpandedItemOne ea -> Right ea
ExpandedItemList _ -> Left $ SpecError ListOfLists ""
Right _eas -> Left $ SpecError ListOfLists ""
-- | Expand a node object, but skip the last step, 12, which may set the result
-- to null under certain conditions. If you use this function, you must do step
-- 12 manually, or guarantee that those conditions aren't met.
expandNodeObjectNotNull
:: ActiveProperty
-> NodeObject
-> Either BereniceError ExpandedNodeObject
expandNodeObjectNotNull property obj = do
-- 1-4 - It's a JSON object
-- 5 - We already did this in tagDocument
let active = nodeContext obj
-- 6-7 - The spec says to go over the keys in lexicographical order.
-- Technically we can do that if we take care to notice and maintain the
-- textual form of all the keys, including ones we represent with albegraic
-- datatypes. But since the order shouldn't matter, I'm wondering why the
-- spec requires a specific order. To avoid unnecessary ugliness, let's
-- just go over them in a convenient order: Handle each 'NodeObject' field
-- that corresponds to a keyword, and then traverse the hashmap containing
-- the rest of the keys.
{-
let emptyObj = ExpandedNodeObject
{ enodeId = Nothing
, enodeGraph = Nothing
, enodeType = Nothing
, enodeReverse = Nothing
, enodeIndex = Nothing
, enodeValues = M.empty
}
-}
-- For each key that is a keyword, other than @context that we already
-- handled:
-- 7.1 - Okay, it's not @context, so keep going
-- 7.2 - No need, keywords are expanded to themselves
-- 7.3 - Expanded property is a keyword, so keep going
-- 7.4 - Indeed, we're handling the keywords
-- 7.4.1 - That's what 'checkReverse' does below for each keyword
-- 7.4.2 - No chance for collision right now
-- 7.4.3 - 7.4.11 - Handle the keywords below
-- 7.4.12 - Be careful below, notice when the value ends up being null!
-- 7.4.13 - Indeed, do them one by one
mnid <- for (aaValue <$> nodeId obj) $ checkReverse $ expandRefNonVocab True active
mgraph <- case aaValue <$> nodeGraph obj of
Nothing -> return Nothing
Just g -> checkReverse' $
let p = ActivePropertyKeyword KeywordGraph
expand = expandNodeObject p
in case g of
One n -> do
mn <- expand n
return $ One <$> mn
Many v ->
Just . Many . V.mapMaybe id <$> traverse expand v
mtype <- for (aaValue <$> nodeType obj) $ checkReverse $ \ nt ->
let expand ck = do
k <- expandRefVocab True active ck
case k of
IdentKwURI u -> Right $ IdentURI u
IdentKwBlank b -> Right $ IdentBlank b
IdentKwC kw -> Left $ ENodeInvalidType $ Left kw
IdentKwNC kw -> Left $ ENodeInvalidType $ Right kw
in case nt of
One ti -> V.singleton <$> expand ti
Many tis -> traverse expand tis
mrv <- for (aaValue <$> nodeReverse obj) $ checkReverse $ \ r -> do
-- 7.4.11 - We already parsed the value, we know it's an object
-- 7.4.11.1
ExpandedReverseObject rrs rvs <- expandReverseObject r
-- 7.4.11.2 - In the double reverse map, which we got from the
-- expansion, all the values are already arrays. The spec is
-- unclear here: Given the existing array in the expanded node
-- object we're building, and the array from the double reverse
-- map, do we add the latter as an array item of the former, or do
-- we (++) the latter to the former? The JS implementation and the
-- Racket implementation (by cwebber) seem to do (++), and it makes
-- more sense to me, so let's do it here too.
-- Since we're just making initial values for our expanded node
-- object, there's nothing to do: The initial map is going to be
-- simply 'rrs' itself.
-- 7.4.11.3 - We're making the initial value and we already handled
-- value and list objects early, so we can just use the value map
-- as is.
let rvs' = if M.null rvs then Nothing else Just rvs
rrs' = M.map (V.map $ ExpandedItemOne . ExpandedAtomNode) rrs
return (rvs', rrs')
let (mreverse, values) = fromMaybe (Nothing, M.empty) mrv
mindex <- for (aaValue <$> nodeIndex obj) $ checkReverse pure
-- Keywords @value, @language, @list and @set aren't allowed in a node
-- object, so we're done with keywords.
let initialEnode = ExpandedNodeObject
{ enodeId = mnid
, enodeGraph = mgraph
, enodeType = mtype
, enodeReverse = mreverse
, enodeIndex = mindex
, enodeValues = values
}
-- Now do 7 on the rest of the keys in the node object
enode <-
foldlM
(processProperty' active)
initialEnode
(M.toList $ nodeValues obj)
-- 8 - This is a node object, it doesn't have @value
-- 9 - We already made sure earlier that @type maps to an array
-- 10 - This is a node object, it doesn't have @list or @set
-- 11 - This is a node object, it doesn't have @language
-- 12 - Skipping, we do this step in 'expandNodeObject'
return enode
where
processProperty' active n (ep, vals) =
foldlM (processProperty active ep) n vals
processProperty active ep n (k, v) = do
-- 7.1 - Nope, it's not @context
-- 7.2 - We already did this in tagDocument
-- 7.3 - ep can't be null (not sure if because the spec makes sure
-- it never happens, or because my code dropped the nulls
-- somewhere), and it has to be either a keyword or stuff that has
-- a colon, because its type guarantees that. So, nothing to check
-- here.
let setKW = setKeyword n
nonKW = processPropertyNonKW active k v n
case ep of
-- 7.4
-- 7.4.1
{-
KeyKeyword kw -> checkReverseM' $ case kw of
-- 7.4.2
KeywordId -> setKW enodeId setEnodeId $ case v of
-- 7.4.3
-- 7.4.12 - can't be null, or we raised error on nulls?
-- 7.4.13
Items (One (NodeItemOne (ItemScalar (ScalarString _ ck)))) -> Just <$> expandRef True False active ck
_ -> Left InvalidIdValue
-- 7.4.2
KeywordType -> setKW enodeType setEnodeType $ case v of
-- 7.4.4
-- 7.4.12 - can't be null, or we raised error on nulls?
-- 7.4.13
Items (One (NodeItemOne (ItemScalar (ScalarString _ ck)))) -> Just . V.singleton <$> expandRef True True active ck
Items (Many nis) -> fmap Just $ for nis $ \ ni -> case ni of
One (ItemScalar (ScalarString _ ck)) -> expandRef True True active ck
_ -> Left InvalidTypeValue
_ -> Left InvalidTypeValue
-- 7.4.2
KeywordGraph -> setKWM enodeGraph setEnodeGraph $ do
-- 7.4.5
-- 7.4.12 - Return Nothing on null so that setKWM drops it
-- 7.4.13
let p = ActivePropertyKeyword KeywordGraph
expand = expandNodeObject fetch base active p
err = GraphIsntNodeObject obj k v
case v of
Items (One (NodeItemOne (ItemNode o))) -> fmap Left <$> expand o
Items (Many nis) -> do
let decide (One (ItemNode o)) = return o
decide _ = throwE err
catMaybesV = V.mapMaybe id
Just . Right . catMaybesV <$>
traverse (decide >=> expand) nis
_ -> throwE err
-- The algo just says to handle it, but node objects aren't
-- supposed to have a @value and the spec says to ignore it
-- when processing. At least for now, just to avoid missing
-- bugs, let's make this cause an error.
KeywordValue -> throwE $ NodeObjectIgnoredField obj kw v
-- Same as above
KeywordLanguage -> throwE $ NodeObjectIgnoredField obj kw v
-- 7.4.2
KeywordIndex -> setKW enodeIndex setEnodeIndex $ case v of
-- 7.4.8
-- 7.4.12 - can't be null, has to be a string
-- 7.4.13
Items (One (NodeItemOne (ItemScalar (ScalarString t _)))) -> Right $ Just t
_ -> Left InvalidIndexValue
-- Same as above
KeywordList -> throwE $ NodeObjectIgnoredField obj kw v
-- Same as above
KeywordSet -> throwE $ NodeObjectIgnoredField obj kw v
-- 7.4.2
KeywordReverse -> case enodeReverse n of
Just _ -> throwE $ SpecError CollidingKeywords ""
Nothing -> case v of
-- 7.4.11
Items (One (NodeItemOne (ItemNode o))) -> do
r <- liftE $ node2reverse o
-- 7.4.11.1
ExpandedReverseObject rrs rvs <-
expandReverseObject fetch base active r
-- 7.4.11.2
let rrs' = M.map (V.map $ ExpandedItemOne . ExpandedAtomNode) rrs
vs = M.unionWith (V.++) (enodeValues n) rrs'
-- 7.4.11.3
rs = if M.null rvs
then Nothing
-- result cannot already have a
-- @reverse member, the spec algo
-- already checked for colliding
-- keywords, and now unnecessarily
-- wants us to check again. No need for
-- that.
else Just rvs
return n { enodeValues = vs, enodeReverse = rs }
_ -> throwE $ SpecError InvalidReverseValue ""
-- 7.4.12
-- Spec implies we skip, haven't checked playground, let's
-- just raise an error here
KeywordContainer -> throwE $ NodeObjectIgnoredField obj kw v
KeywordBase -> throwE $ NodeObjectIgnoredField obj kw v
KeywordVocab -> throwE $ NodeObjectIgnoredField obj kw v
-}
IdentURI u -> nonKW $ IdentURI u
IdentBlank t -> nonKW $ IdentBlank t
processPropertyNonKW active k v n erp =
-- 7.5-7.11
case v of
Left nis -> do
mev <- expandNodeValueForLO active k nis
return $ case mev of
Nothing -> n
Just ev ->
let lo = case ev of
Left ei -> case ei of
ExpandedItemOne ea ->
ExpandedList Nothing $ V.singleton ea
ExpandedItemList lo' -> lo'
Right eas -> ExpandedList Nothing eas
in n { enodeValues =
M.insertWith
(flip (V.++))
erp
(V.singleton $ ExpandedItemList lo) $
enodeValues n
}
Right nv -> case M.lookup k $ activeTerms active of
Just TermDefinition { termReverse = True } -> do
mev <- expandNodeValueNonLO active k nv
case mev of
Nothing -> return n
Just ev -> do
ev' <- case ev of
One i -> return $ V.singleton i
Many v -> return v
eris <- for ev' $ \ i -> case i of
ExpandedItemOne a -> case a of
ExpandedAtomScalar _ -> Left $ SpecError InvalidReversePropertyValue ""
ExpandedAtomNode m -> return m
ExpandedItemList _ -> Left $ SpecError InvalidReversePropertyValue ""
return n
{ enodeReverse = Just $ case enodeReverse n of
Nothing -> M.singleton erp eris
Just m -> M.insertWith (flip (V.++)) erp eris m
}
_ -> do
mev <- expandNodeValueNonLO active k nv
case mev of
Nothing -> return n
Just ev -> do
eis <- case ev of
One i -> return $ V.singleton i
Many v -> return v
return n
{ enodeValues =
M.insertWith (flip (V.++)) erp eis $ enodeValues n
}
isReverse (ActivePropertyKeyword KeywordReverse) = True
isReverse _ = False
checkReverse' a =
if isReverse property
then Left $ SpecError InvalidReversePropertyMap ""
else a
checkReverse f = checkReverse' . f
setKeyword n g s a =
case g n of
Just _ -> Left $ SpecError CollidingKeywords ""
Nothing -> maybe n (s n . Just) <$> a
checkCollision n f a =
case f n of
Just _ -> Left $ SpecError CollidingKeywords ""
Nothing -> a
setEnodeId n v = n { enodeId = v }
setEnodeType n v = n { enodeType = v }
setEnodeGraph n v = n { enodeGraph = v }
setEnodeIndex n v = n { enodeIndex = v }
setEnodeReverse n v = n { enodeReverse = v }
joinReverseValues = Many . V.concatMap toVector
where
toVector (One i) = V.singleton i
toVector (Many v) = v
expandNodeObject
:: ActiveProperty
-> NodeObject
-> Either BereniceError (Maybe ExpandedNodeObject)
expandNodeObject property obj = do
enode <- expandNodeObjectNotNull property obj
-- 12 - Can't contain @value or @list, but check for the other conditions
return $
if nullOrGraph property && emptyOrOnlyId enode
then Nothing
else Just enode
where
nullOrGraph ActivePropertyNull = True
nullOrGraph (ActivePropertyKeyword KeywordGraph) = True
nullOrGraph _ = False
emptyOrOnlyId (ExpandedNodeObject _id Nothing Nothing Nothing Nothing m) =
M.null m
emptyOrOnlyId _ = False
-- | Expand a JSON-LD document.
expandDocument
:: Multi NodeObject -> Either BereniceError (Vector ExpandedNodeObject)
expandDocument (One no) = do
meno <- expandNodeObject ActivePropertyNull no
Right $ case meno of
Nothing -> V.empty
Just eno -> case eno of
ExpandedNodeObject Nothing (Just g) Nothing Nothing Nothing m ->
if M.null m
then toArray g
else V.singleton eno
_ -> V.singleton eno
expandDocument (Many nos) =
V.mapMaybe id <$> traverse (expandNodeObject ActivePropertyNull) nos
-- We're done with the expansion step. The next step in our process of parsing
-- JSON-LD into RDF is flattening. In this step we don't have to do the actual
-- flattening algorithm described in the JSON-LD API spec though. We just do
-- the part where we produce a node map.
--
-- The types for the structure we'll be producing:
data MappedAtom
= MappedAtomScalar ExpandedScalarValue
| MappedAtomNode Identifier
deriving Eq
data MappedNodeObject = MappedNodeObject
{ mnodeId :: Identifier
, mnodeType :: Maybe (Vector Identifier)
, mnodeIndex :: Maybe Text
, mnodeValues :: HashMap Identifier (Vector (Multi MappedAtom))
-- ^ In the @Multi MappedAtom@, a @Many@ means a list object, and a @One@
-- means single plain atom
}
data NodeMap = NodeMap
{ nodeMapDefault :: Maybe (HashMap Identifier MappedNodeObject)
, nodeMapNamed :: HashMap Identifier (HashMap Identifier MappedNodeObject)
}
-- And now the functions:
int2blank :: Int -> RelToken
int2blank counter =
case parseRelToken $ BC.pack label of
Nothing ->
error $
"berenice: int2blank: parseRelToken failed to parse: " ++ label
Just rt -> rt
where
label = 'b' : show counter
newBlank :: Monad m => StateT (HashMap RelToken RelToken, Int) m RelToken
newBlank = do
counter <- gets snd
modify $ second (+ 1)
return $ int2blank counter
replaceBlank
:: Monad m
=> RelToken
-> StateT (HashMap RelToken RelToken, Int) m RelToken
replaceBlank label = do
(idmap, counter) <- get
case M.lookup label idmap of
Just entry -> return entry
Nothing -> do
let new = int2blank counter
put (M.insert label new idmap, counter + 1)
return new
generateNodeMap
:: Monad m
=> Vector ExpandedNodeObject
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) NodeMap
generateNodeMap =
foldlM (flip $ gnmNode activeGraphOpsDef Nothing) $ NodeMap Nothing M.empty
where
--newBlankK' :: Monad m => StateT (HashMap RelToken RelToken, Int) m Key'
newBlankK' = IdentBlank <$> newBlank
replaceBlankK' (IdentBlank b) = IdentBlank <$> replaceBlank b
replaceBlankK' k = return k
idObject i = MappedNodeObject i Nothing Nothing M.empty
{-
initNode
:: Monad m
=> ( NodeMap -> HashMap Key' MappedNodeObject
, HashMap Key' MappedNodeObject -> NodeMap -> NodeMap
)
-> NodeMap
-> ExpandedNodeObject
-> StateT
(HashMap Text Text, Int)
m
( HashMap Key' MappedNodeObject
, Key'
, MappedNodeObject
)
-}
initNode ops nodeMap object = do
let (getGraph, _) = ops
graph = getGraph nodeMap
id_ <- maybe newBlankK' replaceBlankK' $ enodeId object
let idObj = idObject id_
node = M.lookupDefault idObj id_ graph
return (graph, id_, node)
insertSubject ap node s =
let ins p v n = n { mnodeValues = M.insert p v $ mnodeValues n }
in case M.lookup ap $ mnodeValues node of
Nothing -> ins ap (V.singleton s) node
Just eis -> if s `elem` eis
then node
else ins ap (eis `V.snoc` s) node
insertSubjectValue ap node =
insertSubject ap node . One . MappedAtomScalar
insertSubjectNode ap node =
insertSubject ap node . One . MappedAtomNode
insertSubjectList ap node =
insertSubject ap node . Many
activeGraphOpsDef =
( fromMaybe M.empty . nodeMapDefault
, \ g nm -> nm { nodeMapDefault = Just g }
)
activeGraphOpsNamed gid =
( M.lookupDefault M.empty gid . nodeMapNamed
, \ g nm -> nm { nodeMapNamed = M.insert gid g $ nodeMapNamed nm }
)
{-
processNode
:: Monad m
=> NodeMap
-> ( NodeMap -> HashMap Key' MappedNodeObject
, HashMap Key' MappedNodeObject -> NodeMap -> NodeMap
)
-> HashMap Key' MappedNodeObject
-> MappedNodeObject
-> ExpandedNodeObject
-> Key'
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) NodeMap
-}
processNode nodeMap ops graph node object id_ = do
mtypes <- lift $ traverse (traverse replaceBlankK') $ enodeType object
let (_getGraph, setGraph) = ops
node' = case mtypes of
Nothing -> node
Just ks -> case mnodeType node of
Nothing -> node { mnodeType = Just ks }
Just ks' -> node { mnodeType = Just $ ks' V.++ (V.filter (`notElem` ks') ks) }
node'' <- case enodeIndex object of
Nothing -> return node'
Just i -> case mnodeIndex node' of
Nothing -> return node' { mnodeIndex = Just i }
Just i' -> if i == i'
then return node'
else throwE $ SpecError ConflictingIndexes ""
let nodeMap' = setGraph (M.insert id_ node'' graph) nodeMap
nodeMap'' <- case enodeReverse object of
Nothing -> return nodeMap'
Just r ->
let f p nm eno = gnmNode ops (Just (p, Just id_)) eno nm
g nm (p, enos) = foldlM (f p) nm enos
in foldlM g nodeMap' $ M.toList r
nodeMap''' <- case enodeGraph object of
Nothing -> return nodeMap''
Just gs ->
let f nm g = gnmNode (activeGraphOpsNamed id_) Nothing g nm
in foldlM f nodeMap'' gs
let f p nm v = case v of
ExpandedItemOne a -> case a of
ExpandedAtomScalar vo -> gnmValue ops id_ p vo nm
ExpandedAtomNode no -> gnmNode ops (Just (p, Nothing)) no nm
ExpandedItemList l -> gnmList ops id_ p (elistArray l) nm
g nm (p, vs) = do
p' <- lift $ replaceBlankK' p
foldlM (f p') nm vs
foldlM g nodeMap''' $ M.toList $ enodeValues object
gnmValue ops as ap value nodeMap = do
let (getGraph, setGraph) = ops
graph = getGraph nodeMap
node <- case M.lookup as graph of
Just n -> return n
-- TODO here and in gnmList below, if it's proven that the key does
-- exist, maybe we can somehow change the code such that we never
-- have to deal with the Nothing? If changing the code as is
-- doesn't work, try the justified-containers package, see if it
-- can be used here
Nothing -> throwE $ error "Oops, active subject not found!!!"
let node' = insertSubjectValue ap node value
return $ setGraph (M.insert as node' graph) nodeMap
gnmValueWithList list value nodeMap =
return (nodeMap, list `V.snoc` MappedAtomScalar value)
{-
gnmList
:: Monad m
=> ( NodeMap -> HashMap Key' MappedNodeObject
, HashMap Key' MappedNodeObject -> NodeMap -> NodeMap
)
-> Key'
-> Key'
-> Vector ExpandedAtom
-> NodeMap
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) NodeMap
-}
gnmList ops as ap arr nodeMap = do
-- TODO 5.2 says to pass the active subject, is there a reason? Where
-- does it get used? for value objects it isn't used, and for node
-- objects only the object form is used (for @reverse), while here it
-- seems rquired the active subject is a string. Below I'm simply not
-- passing it for this reason; is it possible I missed something? Or
-- the algo is being silly? Because unless I missed something, the algo
-- could have said to pass 'null' for active subject, and behavior
-- would be the same, beside the seemingly unnecessary lookup in step 2
let f (nm, l) (ExpandedAtomScalar vo) = gnmValueWithList l vo nm
f (nm, l) (ExpandedAtomNode no) = gnmNodeWithList ops ap l no nm
(nodeMap', list) <- foldlM f (nodeMap, V.empty) arr
let (getGraph, setGraph) = ops
graph = getGraph nodeMap'
node <- case M.lookup as graph of
Just n -> return n
Nothing -> throwE $ error "Oops, active subject not found!!!"
let node' = insertSubjectList ap node list
return $ setGraph (M.insert as node' graph) nodeMap'
{-
gnmNode
:: Monad m
=> ( NodeMap -> HashMap Key' MappedNodeObject
, HashMap Key' MappedNodeObject -> NodeMap -> NodeMap
)
-> Maybe (Key', Maybe Key')
-> ExpandedNodeObject
-> NodeMap
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) NodeMap
-}
gnmNode ops activeSubjectAndProperty object nodeMap = do
(graph, id_, node) <- lift $ initNode ops nodeMap object
let node' = case activeSubjectAndProperty of
Nothing -> node
Just (ap, mas) -> insertSubjectNode ap node $ case mas of
Just as -> as
Nothing -> id_
processNode nodeMap ops graph node' object id_
{-
gnmNodeWithList
:: Monad m
=> ( NodeMap -> HashMap Key' MappedNodeObject
, HashMap Key' MappedNodeObject -> NodeMap -> NodeMap
)
-> Key'
-> Vector MappedAtom
-> ExpandedNodeObject
-> NodeMap
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) (NodeMap, Vector MappedAtom)
-}
gnmNodeWithList ops ap l object nodeMap = do
(graph, id_, node) <- lift $ initNode ops nodeMap object
let list = l `V.snoc` MappedAtomNode id_
flip (,) list <$> processNode nodeMap ops graph node object id_
-- Finally, now we do the deserialization step (which could enjoy a better
-- name), converting the node map into RDF triples.
--
-- Some helper types:
data TypedLiteralString = TypedLiteralString
{ lsText :: Text
, lsLang :: Maybe Text
}
data TypedLiteralNB = TypedLiteralNumber Scientific | TypedLiteralBool Bool
data TypedLiteral = TypedLiteral
{ tlitValue :: Either TypedLiteralNB TypedLiteralString
, tlitType :: URI Gen
}
-- Now the types for the structure we'll produce:
data PlainLiteral = PlainLiteral
{ literalLexical :: Text
, literalTypeLang :: Either (URI Gen) Text
}
data Node
= NodeIdent Identifier
| NodeLiteral PlainLiteral
data Triple = Triple
{ tripleSubject :: Identifier
, tripleProperty :: URI Gen
, tripleObject :: Node
}
data RdfDataset = RdfDataset
{ rdfDefault :: [Triple]
, rdfNamed :: HashMap Identifier [Triple]
}
-- And the functions:
u_langString, u_string, u_boolean, u_integer, u_double, u_first, u_rest, u_nil,
u_type, u_List :: URI Gen
u_langString = $(rdf "langString")
u_string = $(xsd "string")
u_boolean = $(xsd "boolean")
u_integer = $(xsd "integer")
u_double = $(xsd "double")
u_first = $(rdf "first")
u_rest = $(rdf "rest")
u_nil = $(rdf "nil")
u_type = $(rdf "type")
u_List = $(rdf "List")
-- | Serialize a Haskell typed literal into a plain textual one. The textual
-- literal can then be used when serializing RDF data for storage, or for
-- sending over the network, or for rendering a query, and so on. The plain
-- literal is the canonical representation, while the typed one is for
-- convenience of manipulating and using data in Haskell.
typed2plain :: TypedLiteral -> Either BereniceError PlainLiteral
typed2plain (TypedLiteral vl typ) =
case (typ == u_langString, mlang) of
(True, Nothing) -> Left $ LangStringNoLanguage vl
(False, Just lang) -> Left $ NonLangStringHasLanguage vl lang
_ -> case vl of
Left nb -> flip PlainLiteral (Left typ) <$>
case nb of
TypedLiteralBool b -> Right $ if b then "true" else "false"
TypedLiteralNumber s ->
if isInteger s && typ /= u_double
then case toInt64 s of
Nothing -> Left $ SuspiciouslyBigInteger s
Just i -> Right $ T.pack $ show i
else case toDouble s of
Nothing -> Left $ OutOfRangeOfDouble s
Just d ->
Right $ T.pack $ showEFloat Nothing d ""
Right (TypedLiteralString t ml) ->
Right $ PlainLiteral t $
case mlang of
Nothing -> Left typ
Just l -> Right l
where
mlang = case vl of
Left _ -> Nothing
Right s -> lsLang s
toInt64 :: Scientific -> Maybe Int64
toInt64 = toBoundedInteger
toDouble :: Scientific -> Maybe Double
toDouble = either (const Nothing) Just . toBoundedRealFloat
value2resource :: ExpandedScalarValue -> TypedLiteral
value2resource (ExpandedScalarValue value _index) =
case value of
Left (ExpandedNBValue v mt) ->
case v of
Left s -> TypedLiteral
{ tlitValue = Left $ TypedLiteralNumber s
, tlitType =
case mt of
Just u -> u
Nothing ->
if isInteger s
then u_integer
else u_double
}
Right b -> TypedLiteral
{ tlitValue = Left $ TypedLiteralBool b
, tlitType = fromMaybe u_boolean mt
}
Right (ExpandedStringValue text _token mtorl) -> TypedLiteral
{ tlitValue = Right TypedLiteralString
{ lsText = text
, lsLang =
case mtorl of
Just (Right l) -> Just l
_ -> Nothing
}
, tlitType =
case mtorl of
Just (Left u) -> u
Just (Right _) -> u_langString
Nothing -> u_string
}
object2rdf :: MappedAtom -> Either BereniceError Node
object2rdf (MappedAtomScalar s) =
fmap NodeLiteral $ typed2plain $ value2resource s
object2rdf (MappedAtomNode i) = Right $ NodeIdent i
list2rdf
:: Monad m
=> Vector MappedAtom
-> ExceptT BereniceError (StateT (HashMap RelToken RelToken, Int) m) [Triple]
list2rdf =
((foldr mkTriples [] . foldr mkRest []) <$>) .
traverse (\ a -> (,) <$> lift newBlank <*> liftE (object2rdf a))
where
liftE = ExceptT . pure
rest [] = IdentURI u_nil
rest ((s, _, _):_) = IdentBlank s
mkRest (s, o) l = (s, o, NodeIdent $ rest l) : l
mkTriples (s, o, r) l =
Triple (IdentBlank s) u_first o : Triple (IdentBlank s) u_rest r : l
deserializeToRDF
:: Vector ExpandedNodeObject -> Either BereniceError RdfDataset
deserializeToRDF = flip evalState (M.empty, 0) . runExceptT . doc2triples
where
liftE :: Monad m => Either a b -> ExceptT a m b
liftE = ExceptT . pure
eno2triples subject node = do
nonkw <- fmap concat $ for (M.toList $ mnodeValues node) $
\ (p, mis) -> case p of
IdentBlank b -> throwE $ BlankProperty subject b
IdentURI u -> fmap fold $ for mis $ \ mi -> case mi of
One ma -> do
object <- liftE $ object2rdf ma
return [Triple subject u object]
Many ml -> do
ts <- list2rdf ml
return $ case ts of
[] ->
[Triple
subject
u $
NodeIdent $ IdentURI u_nil
]
(t:_) ->
let triple =
Triple
subject
u
(NodeIdent $ tripleSubject t)
in triple : ts
let typ2triple = Triple subject u_type . NodeIdent
typ = maybe [] (V.toList . V.map typ2triple) $ mnodeType node
return $ typ ++ nonkw
graph2triples = fmap concat . traverse (uncurry eno2triples) . M.toList
doc2triples enos = do
NodeMap mdef named <- generateNodeMap enos
RdfDataset
<$> maybe (pure []) graph2triples mdef
<*> traverse graph2triples named
-- Finally, we define a function that does the whole process, from plain JSON
-- all the way to RDF triples.
jsonld2rdf
:: Monad m
=> (URI Gen -> ExceptT BereniceError m (Multi Dictionary))
-> URI Gen
-> Multi Dictionary
-> ExceptT BereniceError m RdfDataset
jsonld2rdf fetch base
= ExceptT . pure . traverse parseDocument
>=> traverse
(tagDocument fetch base >=> ExceptT . pure . specializeDocument)
>=> ExceptT . pure . (expandDocument >=> deserializeToRDF)
-- TODO after expansion, drop the esToken field? Do we ever need it? I mean, if
-- expansion produces ExpandedStringValue etc. perhaps remove the esToken field
-- because we use it *during* expansion, and after that iirc it's never really
-- used?
-- We're done with the code for converting from JSON to RDF. Below comes the
-- other direction, serializing an RDF dataset into a JSON-LD document.
--
-- The first step in our process of serializing RDF to JSON-LD is to build a
-- node map. In other words, we apply the RDF to JSON-LD serialization
-- algorithm, but for each graph separately, not combining them yet into a
-- single JSON-LD document.
--
-- We're converting between existing types, 'RdfDataset' and 'NodeMap', but we
-- still define some helper types:
-- | A type for tracking references to a node. Given a node with identifier
-- /i/, are there triples in which /i/ is used as the object? For each such
-- occurence we keep a 'Usage' value. It specifies the node object that makes
-- use of /i/, and in which property. In other words, it means that node object
-- /n/ has a property /p/ and the value associated with it is /i/.
data Usage = Usage
{ usageNode :: Identifier
, usageProperty :: URI Gen
-- , usageValue :: Identifier
}
data Single a = None | Single a | More deriving Functor
data UseNativeTypes = UseNativeTypes | DontUseNativeTypes
data UseRdfType = UseRdfType | UseTypeKeyword
-- And now the functions for building the node map:
-- | Convert the object of an RDF triple into an object suitable for use as a
-- value in a node object's key-value map.
--
-- The conversion into native types is stricter than required by the spec:
--
-- * For xsd:boolean, the spec converts @true@ and @false@ into boolean
-- literals and anything else remains as a string literal. But in this
-- implementation, @true@, @false@, 0 and 1 are converted into boolean
-- literals (because that's the standard lexical space for xsd:boolean) and
-- anything else raises an error.
-- * For xsd:double and xsd:integer, the spec converts a valid lexical form
-- into a numeric literal, and otherwise keeps it as a string literal. But in
-- this implementation, an invalid lexical form results with an error.
rdf2object :: UseNativeTypes -> Node -> Either BereniceError MappedAtom
rdf2object _ (NodeIdent i) = Right $ MappedAtomNode i
rdf2object native (NodeLiteral (PlainLiteral lexical torl)) =
MappedAtomScalar . flip ExpandedScalarValue Nothing <$>
case native of
UseNativeTypes ->
case torl of
Left typ
| typ == u_boolean ->
case lexical of
"true" -> Right $ Left $ boolean True
"false" -> Right $ Left $ boolean False
"1" -> Right $ Left $ boolean True
"0" -> Right $ Left $ boolean False
_ -> Left $ InvalidXsdBoolean lexical
| typ == u_integer ->
case readMaybe $ T.unpack lexical of
Nothing -> Left $ InvalidXsdInteger lexical
Just i ->
Right $ Left $ number $ scientific i 0
| typ == u_double ->
case readMaybe $ T.unpack lexical of
Nothing -> Left $ InvalidXsdDouble lexical
Just s -> Right $ Left $ number s
| otherwise -> Right $ Right string
Right _ -> Right $ Right string
DontUseNativeTypes -> Right $ Right string
where
string = ExpandedStringValue
{ esText = lexical
, esToken = parseToken lexical
, esTypeOrLang =
case torl of
Left typ ->
if typ == u_string
then Nothing
else Just $ Left typ
Right lang -> Just $ Right lang
}
boolean b = ExpandedNBValue
{ enbValue = Right b
, enbType = Nothing
}
number s = ExpandedNBValue
{ enbValue = Left s
, enbType = Nothing
}
-- TODO in the property-objects map below we use Vector but we really mean
-- NonEmpty i.e. assume it can't be an empty Vector, no need to use V.null
-- because we assume it's guaranteed to be False
-- | Build a node object from the RDF triples that describe it.
--
-- This function takes RDF triples as input, all of which share the same
-- subject. The triples are passed in the form of 3 parameters:
--
-- * The 'Identifier' is the shared subject of the triples.
-- * The 'Vector' is a list of types the subject is a member of. For each type
-- identifier /i/, the triple *s - rdf:type - i* is represented, where *s* is
-- the shared subject. Note that the 'Vector' must be non-empty! If there are
-- no triples with the rdf:type property, pass 'Nothing'.
-- * The 'HashMap' represents the rest of the triples, i.e. all the ones whose
-- property isn't rdf:type. For each key *k* in the map, for each node *n* in
-- the 'Vector' associated with it, the triple @s - k - n@ is represented.
-- Note that all these 'Vector' must be non-empty! If some property isn't
-- used, simply don't insert it into the 'HashMap' at all.
mappedNodeObject
:: UseNativeTypes
-- ^ Whether to try converting literals to native JSON literals, or
-- serialize all literals as JSON strings
-> UseRdfType
-- ^ Whether to serialize the rdf:type property as rdf:type or as @type in
-- the produced JSON-LD document
-> Identifier
-- ^ The shared subject of all the triples
-> Maybe (Vector Identifier)
-- ^ A list of nodes for rdf:type, i.e. each node /i/ represents the triple
-- @s - rdf:type - i@ where /s/ is the shared subject
-> HashMap (URI Gen) (Vector Node)
-- ^ The rest of the triples, i.e. everything except for rdf:type. The map
-- keys are properties and each property maps to a 'Vector' of triple
-- objects. For each property /p/ and a node /n/ from the vector it maps
-- to, the triple @s - p - n@ is expressed, where /s/ is the shared
-- subject.
-> Either BereniceError MappedNodeObject
mappedNodeObject native rdftype subject mtypes properties =
let mno vs = MappedNodeObject
{ mnodeId = subject
, mnodeType =
case UseRdfType of
UseRdfType -> Nothing
UseTypeKeyword -> mtypes
, mnodeIndex = Nothing
, mnodeValues = vs
}
prop2value property nodes =
(,) (IdentURI property) . V.map One <$>
traverse (rdf2object native) nodes
insertTypes =
case (UseRdfType, mtypes) of
(UseRdfType, Just types) ->
M.insert
(IdentURI u_type)
(V.map (One . MappedAtomNode) types)
_ -> id
in mno . insertTypes . M.fromList <$>
traverse (uncurry prop2value) (M.toList properties)
nodeObjectUsageOf
:: Identifier
-> HashMap (URI Gen) (Vector Node)
-> [URI Gen]
nodeObjectUsageOf ident =
M.keys . M.filter (V.elem ident . V.mapMaybe node2ident)
where
node2ident (NodeIdent i) = Just i
node2ident (NodeLiteral _) = Nothing
-- | This function takes RDF triples as input, all of which share the same
-- subject. That shared subject isn't passed because it isn't used here. And
-- the 'HashMap' maps from properties to objects. So each 'Node' in the vector
-- represents a separate triple.
--
-- The output is a mapping containing all the references it makes to other node
-- objects (or to itself), i.e. triples in which the object is an IRI or a
-- blank node. The mapping is organized by referred object, i.e. for each
-- referred object there's a non-empty list of the properties through which it
-- is referred to.
--
-- NOTE: Since for each referred object we'll only need at most a single usage
-- (because when there's more than one, we don't build a list object), we're
-- only keeping one usage per object, via the 'Single' type. To be more
-- precise, we need to generate all the usages that target rdf:nil, and just
-- one usage targetting each blank node identifier, and for URIs which aren't
-- rdf:nil we don't need to generate usages at all.
--
-- So, here we just generate a usage for each blank node identifier. We'll
-- separately generate the usages for rdf:nil.
nodeObjectBlankUsage
:: HashMap (URI Gen) (Vector Node)
-> HashMap RelToken (Single (URI Gen))
nodeObjectBlankUsage =
M.fromList .
-- ^ Turn into a 'HashMap', the identifiers are unique so we don't lose any
-- data here
-- Result: HashMap Text (Single (URI Gen))
map (second $ fmap snd) .
-- -- ^ Since each 'NonEmpty' shares the identifier, extract it from there to
-- be attached just once, to a 'NonEmpty' of properties
-- Result: [(Text, Single (URI Gen))]
groupAllWith fst .
-- ^ Sort and group by identifier i.e. the referred node
-- Result: [(Text, Single (Text, URI Gen))]
concat .
-- ^ Each list is for one property, concatenate them all together now
-- Result: [(Text, URI Gen)]
map (NE.toList . (\ (prop, idents) -> flip (,) prop <$> idents)) .
-- ^ Attach the property to all identifiers
-- Result: [[(Text, URI Gen)]]
M.toList .
-- ^ Result: [URI Gen, NonEmpty Text]
M.mapMaybe (nonEmpty . V.toList . V.mapMaybe node2blank)
-- ^ Filter out nodes that are literals or URIs, keep only blanks
-- Result: HashMap (URI Gen) (NonEmpty Text)
where
node2blank (NodeIdent (IdentBlank b)) = Just b
node2blank _ = Nothing
groupWith :: Eq b => (a -> b) -> [a] -> [(b, Single a)]
groupWith f = go . map (\ x -> (x, f x))
where
go [] = []
go ((x, y) : ps) = go' x y ps
go' x y ps =
case ps of
[] -> [(y, Single x)]
((u, v) : qs) ->
if y == v
then (y, More) : go (dropWhile ((== y) . snd) qs)
else (y, Single x) : go' u v qs
groupAllWith :: Ord b => (a -> b) -> [a] -> [(b, Single a)]
groupAllWith f = groupWith f . sortOn f
-- | This function combines 'mappedNodeObject', 'nodeObjectBlank Usage' and
-- 'nodeObjectUsageOf' for rdf:nil. It extracts the rdf:type triples, applies
-- those 3 functions and returns a tuple containing their results.
--
-- NOTE 1: In the specification's algorithm, if use-rdf-type is set to false,
-- we place the type(s) under a @type property only if their values are URIs or
-- blank nodes. If one of the types is found to be a literal, we keep it
-- silently under rdf:type. The implementation here, however, *always* checks
-- the types, regardless of the use-rdf-type option, and if any of them are
-- found to be literals, it raises an error instead of quietly accepting them.
--
-- NOTE 2: In the specification's algorithm, if use-rdf-type is set to false,
-- i.e. we use @type, it's excluded from the list of usages. In other words,
-- the URIs associated with rdf:type are assumed not to be RDF lists and aren't
-- checked for that. The implementation here follows that, but it excludes
-- rdf:type even in the case use-rdf-type is true, i.e. regardless of whether
-- we encode it as @type or as rdf:type, we assume here that the type of
-- something can't be a list (a resource can be of more than one type, but, we
-- assume none of those types can be lists).
--
-- By the way, the RDF schema spec says the range of the rdf:type property is
-- rdfs:Class.
nodeAndUsage
:: UseNativeTypes
-> UseRdfType
-> Identifier
-> HashMap (URI Gen) (Vector Node)
-> Either
BereniceError
(MappedNodeObject, ([URI Gen], HashMap RelToken (Single (URI Gen))))
nodeAndUsage native rdftype subject properties = do
let (mtypes, properties') = lookupAndDelete u_type properties
mtypes' <- traverse (traverse $ node2type subject) mtypes
let result mno =
( mno
, ( nodeObjectUsageOf (IdentURI u_nil) properties'
, nodeObjectBlankUsage properties'
)
)
result <$> mappedNodeObject native rdftype subject mtypes' properties'
where
node2type _ (NodeIdent id_) = Right id_
node2type i (NodeLiteral pl) = Left $ InvalidRdfType i pl
-- | For each node object /n/ in the graph, attach its usages, i.e. node
-- objects that have a property associated with the @id of /n/. The input is a
-- graph in which each node object specifies its usage of other node objects:
-- It's a map from such an object's identifier to the list of properties by
-- which it's referred to.
attachUsages
:: HashMap
Identifier
(MappedNodeObject, ([URI Gen], HashMap RelToken (Single (URI Gen))))
-> (HashMap Identifier (MappedNodeObject, Single Usage), [Usage])
attachUsages g =
let initial = M.map (second $ const None) g
mkUsages (mno, um) = M.map (fmap $ Usage $ mnodeId mno) $ snd um
glist = M.toList g
blankUsages = map (mkUsages . snd) glist
nilUsages =
concatMap (\ (i, (_, (props, _))) -> map (Usage i) props) glist
midObj i = MappedNodeObject i Nothing Nothing M.empty
combine (_, new) (mno, old) = (mno, old `add` new)
f = M.foldlWithKey' $ \ m b su ->
let i = IdentBlank b
in M.insertWith combine i (midObj i, su) m
in (foldl' f initial blankUsages, nilUsages)
where
add None s = s
add (Single a) None = Single a
add _ _ = More
gatherListObjectItems
:: HashMap Identifier (MappedNodeObject, Single Usage)
-> Usage
-> [(Identifier, MappedAtom)]
-> (Usage, Identifier, [(Identifier, MappedAtom)])
gatherListObjectItems graph (Usage i p) = go (graph M.! i) p $ IdentURI u_nil
where
oneMaybe v =
if V.length v == 1
then Nothing
else Just $ V.head v
optionalMaybe v =
case V.length v of
0 -> Just Nothing
1 -> Just $ Just $ V.head v
_ -> Nothing
rdfList Nothing = False
rdfList (Just mi) =
case mi of
Nothing -> True
Just i -> i /= IdentURI u_List
listNode (MappedNodeObject _ mtypes mi vals) su p =
let (firsts, vals1) = lookupAndDelete (IdentURI u_first) vals
(rests, vals2) = lookupAndDelete (IdentURI u_rest) vals1
in case ( p == u_rest
, su
, oneMaybe =<< firsts
, oneMaybe =<< rests
, M.null vals2
, mi
, rdfList $ optionalMaybe =<< mtypes
) of
(True, Single u, Just (One first), Just _, True, Nothing, True)
-> Just (u, first)
_ -> Nothing
go (mno, su) property value items =
let ident = mnodeId mno
in case listNode mno su property of
Nothing -> (Usage ident property, value, items)
Just (u'@(Usage ident' property'), first) ->
let items' = (ident, first) : items
in case ident' of
IdentURI _ -> (u', ident, items')
IdentBlank _ ->
go (graph M.! ident') property' ident items'
buildListObjects
:: HashMap Identifier (MappedNodeObject, Single Usage)
-> Usage
-> HashMap Identifier (MappedNodeObject, Single Usage)
buildListObjects graph usage =
let (Usage nodeIdent property, headId, items) =
gatherListObjectItems graph usage []
canConvert =
if property == u_first
then case items of
[] -> Nothing
((s, _) : is) ->
Just $ case is of
[] -> (headId, u_rest, s, is)
((h, _) : _) -> (s , u_rest, h, is)
else Just (nodeIdent, property, headId, items)
in case canConvert of
Nothing -> graph
Just (subj, prop, hid, is) ->
let (subjects, array) = unzip is
adjustProp vals =
case V.elemIndex (One $ MappedAtomNode hid) vals of
Nothing -> error "Impossible!"
Just i -> vals V.// [(i, Many $ V.fromList array)]
adjustMNO (mno, us) =
( mno
{ mnodeValues =
M.adjust adjustProp (IdentURI prop) $
mnodeValues mno
}
, us
)
in foldl'
(flip M.delete)
(M.adjust adjustMNO subj graph)
subjects
serializeGraph
:: UseNativeTypes
-> UseRdfType
-> HashMap Identifier (HashMap (URI Gen) (Vector Node))
-> Either BereniceError (HashMap Identifier MappedNodeObject)
serializeGraph native rdftype graph = do
nodesUsages <- M.traverseWithKey (nodeAndUsage native rdftype) graph
let (graphWithUsage, nilUsages) = attachUsages nodesUsages
return $ M.map fst $ foldl' buildListObjects graphWithUsage nilUsages
-- TODO somewhere above there's a usage of groupAllWith iirc? Maybe I can
-- replace it with the groupAllExtract defined below?
triples2map :: [Triple] -> HashMap Identifier (HashMap (URI Gen) (Vector Node))
triples2map = recordMap3 tripleSubject tripleProperty tripleObject
rdf2nm
:: UseNativeTypes
-> UseRdfType
-> RdfDataset
-> Either BereniceError NodeMap
rdf2nm native rdftype (RdfDataset def named) =
NodeMap
<$> (Just <$> serialize def)
<*> traverse serialize named
where
serialize = serializeGraph native rdftype . triples2map
-- Once we've built the node map, the next step is to produce a JSON-LD
-- document from it. Let's call it render, we render the node map. It's the
-- opposite of generating a node map.
--
-- Building the node map and then rendering it, these two steps together,
-- correspond to the RDF-to-JSONLD serialization algorithm in the
-- specification.
renderNodeMap :: NodeMap -> Vector ExpandedNodeObject
renderNodeMap (NodeMap mdef named) =
V.fromList $
map (uncurry named2eno) (M.toList named) ++ maybe [] graph2enos mdef
where
onlyId (MappedNodeObject _ Nothing Nothing m) = M.null m
onlyId _ = False
ma2ea (MappedAtomScalar esv) = ExpandedAtomScalar esv
ma2ea (MappedAtomNode i) = ExpandedAtomNode ExpandedNodeObject
{ enodeId = Just i
, enodeGraph = Nothing
, enodeType = Nothing
, enodeReverse = Nothing
, enodeIndex = Nothing
, enodeValues = M.empty
}
mma2ei (One a) = ExpandedItemOne $ ma2ea a
mma2ei (Many as) = ExpandedItemList $ ExpandedList Nothing $ V.map ma2ea as
mno2eno (MappedNodeObject id_ mtypes mindex vals) = ExpandedNodeObject
{ enodeId = Just id_
, enodeGraph = Nothing
, enodeType = mtypes
, enodeReverse = Nothing
, enodeIndex = mindex
, enodeValues = M.map (V.map mma2ei) vals
}
graph2enos = map mno2eno . filter (not . onlyId) . M.elems
named2eno i g = ExpandedNodeObject
{ enodeId = Just i
, enodeGraph = Just $ Many $ V.fromList $ graph2enos g
, enodeType = Nothing
, enodeReverse = Nothing
, enodeIndex = Nothing
, enodeValues = M.empty
}
-- After rendering the node map, we've obtained a valid JSON-LD document.
-- However, we may wish to compact it, transform it into compacted form.
-- JSON-LD compaction is done here in three steps. Much like the expansion
-- algorithm occurs in three steps here, tagging and specialization and
-- expansion, the compaction algorithm occurs in three steps as well:
-- Compaction, generalization and untagging.
--
-- Let's start with the compaction step. In this step we tag properties with
-- their compacted versions (determined by the IRI compaction algorithm) and do
-- some transformations on the values associated by them. The rest of the
-- process, the actual reorganizing of the objects to be mapped by compacted
-- properties, will happen during generalization and untagging.
--
-- Here are some helper types for the compaction step:
data InverseTerm = InverseTerm
{ itTerm :: Ref
, itReverse :: Bool
, itTypeOrLang :: Maybe (Either TermDefinitionType ContextLanguage)
}
data ByContainer i = ByContainer
{ bcNone :: i
, bcNull :: i
, bcSet :: i
, bcList :: i
, bcLang :: i
, bcIndex :: i
}
deriving Functor
data TypeMap = TypeMap
{ tmNone :: Maybe Ref
, tmId :: Maybe Ref
, tmVocab :: Maybe Ref
, tmReverse :: Maybe Ref
, tmValues :: HashMap (URI Gen) Ref
}
data LanguageMap = LanguageMap
{ lmNone :: Maybe Ref
, lmNull :: Maybe Ref
, lmValues :: HashMap Text Ref
}
type ContainerMap = ByContainer (LanguageMap, TypeMap)
type InverseContext = HashMap IdentKw ContainerMap
-- And here are the functions for compacting an expanded document:
renderAC :: AbsoluteOrCompact -> ByteString
renderAC (JustAbsolute u) = renderURI u
renderAC (JustCompact c) = renderCompactURI c
renderAC (AbsoluteOrCompact _ c) = renderCompactURI c
showRef :: Ref -> Text
showRef = decodeUtf8 . showRef'
where
showRef' (RefURI ac) = renderAC ac
showRef' (RefBlank rt) = "_:" <> renderRelToken rt
showRef' (RefTerm rel) = renderRelNoAuth rel
inverseContext :: ActiveContext -> InverseContext
inverseContext (ActiveContext _mbase _mvocab mlang terms) =
M.fromList $
-- ^ Result: HashMap IdentKw (ByContainer (LanguageMap, TypeMap))
map (second $
fmap (foldl' update emptyLT . sortBy (order `on` itTerm)) .
-- ^ Result: ByContainer (LanguageMap, TypeMap)
foldr collect emptyITM
-- ^ Result: ByContainer [InverseTerm]
-- Input: NonEmpty (Ref, TermDefinition)
) $
-- ^ Result: [(IdentKw, ByContainer (LanguageMap, TypeMap))]
groupAllExtract (termTarget . snd) id $
-- ^ Result: [(IdentKw, NonEmpty (Ref, TermDefinition))]
M.toList terms
-- ^ Result: [(Ref, TermDefinition)]
where
emptyITM :: ByContainer [InverseTerm]
emptyITM = ByContainer [] [] [] [] [] []
collect :: (Ref, TermDefinition) -> ByContainer [InverseTerm] -> ByContainer [InverseTerm]
collect (term, TermDefinition _ reverse mtorl mcontainer) itm =
let it = InverseTerm term reverse mtorl
in case mcontainer of
Nothing -> itm { bcNone = it : bcNone itm }
Just container -> case container of
ContainerNull -> itm { bcNull = it : bcNull itm }
ContainerSet -> itm { bcSet = it : bcSet itm }
ContainerList -> itm { bcList = it : bcList itm }
ContainerLanguage -> itm { bcLang = it : bcLang itm }
ContainerIndex -> itm { bcIndex = it : bcIndex itm }
order :: Ref -> Ref -> Ordering
order s t =
case s' `T.compareLength` T.length t' of
LT -> LT
GT -> GT
EQ -> compare s' t'
where
s' = showRef s
t' = showRef t
emptyLT =
( LanguageMap Nothing Nothing M.empty
, TypeMap Nothing Nothing Nothing Nothing M.empty
)
update :: (LanguageMap, TypeMap) -> InverseTerm -> (LanguageMap, TypeMap)
update lt@(lm, tm) (InverseTerm term reverse mtorl) =
if reverse
then case tmReverse tm of
Just _ -> lt
Nothing -> (lm, tm { tmReverse = Just term })
else case mtorl of
Just torl -> case torl of
Left typ -> case typ of
TermDefinitionTypeId -> case tmId tm of
Just _ -> lt
Nothing -> (lm, tm { tmId = Just term })
TermDefinitionTypeVocab -> case tmVocab tm of
Just _ -> lt
Nothing -> (lm, tm { tmVocab = Just term })
TermDefinitionTypeAbsolute u -> case M.lookup u $ tmValues tm of
Just _ -> lt
Nothing -> (lm, tm { tmValues = M.insert u term $ tmValues tm })
Right lang -> case lang of
ContextLanguageNull -> case lmNull lm of
Just _ -> lt
Nothing -> (lm { lmNull = Just term }, tm)
ContextLanguageTag t -> case M.lookup t $ lmValues lm of
Just _ -> lt
Nothing -> (lm { lmValues = M.insert t term $ lmValues lm }, tm)
Nothing -> case (lmNone lm, maybe (lmNone lm) (flip M.lookup $ lmValues lm) mlang, tmNone tm) of
(Just _, Just _, Just _) -> lt
_ -> ( lm { lmNone = lmNone lm <|> Just term
, lmValues = case mlang of
Nothing -> lmValues lm
Just lang -> M.insertWith (const id) lang term $ lmValues lm
}
, tm { tmNone = tmNone tm <|> Just term }
)
selectTerm
:: ContainerMap
-> [ContainerMap -> (LanguageMap, TypeMap)]
-> Either [TypeMap -> Maybe Ref] [LanguageMap -> Maybe Ref]
-> Maybe Ref
selectTerm cm containers preferred =
let select (lm, tm) = case preferred of
Left types -> asum $ map ($ tm) types
Right langs -> asum $ map ($ lm) langs
in asum $ map (select . ($ cm)) containers
data LanguageMapField
= LMNone
| LMNull
| LMValue Text
deriving Eq
data CommonLanguage = CommonLanguage
{ clGet :: LanguageMap -> Maybe Ref
, clTag :: LanguageMapField
}
data TypeMapField
= TMNone
| TMId
| TMVocab
| TMReverse
| TMValue (URI Gen)
deriving Eq
data CommonType = CommonType
{ ctGet :: TypeMap -> Maybe Ref
, ctTag :: TypeMapField
}
updateCommon
:: (Maybe CommonType, Maybe CommonLanguage)
-> ExpandedAtom
-> (Maybe CommonType, Maybe CommonLanguage)
updateCommon (mct, mcl) item =
let il = clNone
it = ctNone
(il', it') = case item of
ExpandedAtomScalar esv -> case scalarValue esv of
Left enb -> case enbType enb of
Just u -> (il, ctValue u)
Nothing -> (clNull, it)
Right es -> case esTypeOrLang es of
Just torl -> case torl of
Left u -> (il, ctValue u)
Right l -> (clValue l, it)
Nothing -> (clNull, it)
ExpandedAtomNode _ -> (il, ctId)
cl' = case mcl of
Nothing -> Just il
Just cl -> Just $
case (clTag cl /= clTag il, item) of
(True, ExpandedAtomScalar _) -> clNone
_ -> cl
ct' = case mct of
Nothing -> Just it
Just ct -> Just $
if ctTag ct /= ctTag it
then ctNone
else ct
in (ct', cl')
where
clNone = CommonLanguage lmNone LMNone
clNull = CommonLanguage lmNull LMNull
clValue l = CommonLanguage (M.lookup l . lmValues) $ LMValue l
ctNone = CommonType tmNone TMNone
ctId = CommonType tmId TMId
ctValue u = CommonType (M.lookup u . tmValues) $ TMValue u
-- TODO since we're building containers array by consing, remember to reverse
-- it before passing it to 'selectTerm'
handleList
:: CommonLanguage
-> [ContainerMap -> (LanguageMap, TypeMap)]
-> ExpandedList
-> ( Either CommonType CommonLanguage
, [ContainerMap -> (LanguageMap, TypeMap)]
)
handleList defaultLanguage containers (ExpandedList mindex items) =
let containers' = case mindex of
Nothing -> bcList : containers
Just _ -> containers
-- list = items
commonType = Nothing
commonLanguage =
if V.null items
then Just defaultLanguage
else Nothing
bothNone (ct, cl) = case (ctTag <$> ct, clTag <$> cl) of
(Just TMNone, Just LMNone) -> True
_ -> False
(commonType', commonLanguage') = foldlUntil' bothNone updateCommon (commonType, commonLanguage) items
cl = fromMaybe clNone commonLanguage'
ct = fromMaybe ctNone commonType'
in ( if ctTag ct /= TMNone
then Left ct
else Right cl
, containers'
)
where
clNone = CommonLanguage lmNone LMNone
ctNone = CommonType tmNone TMNone
data AssociatedValue
= AssocItem ExpandedItem
| AssocArray
compactToTerm
:: Maybe AssociatedValue
-> Bool
-> ActiveContext
-> InverseContext
-> ContainerMap
-> Maybe Ref
compactToTerm mvalue reverze active inverse cm =
let defLang = case activeLanguage active of
Nothing -> clNone
Just lang -> clValue lang
containers =
case mvalue of
Just (AssocItem (ExpandedItemOne (ExpandedAtomScalar (ExpandedScalarValue _ (Just _))))) -> [bcIndex]
Just (AssocItem (ExpandedItemOne (ExpandedAtomNode (ExpandedNodeObject { enodeIndex = Just _ })))) -> [bcIndex]
Just (AssocItem (ExpandedItemList (ExpandedList (Just _) _))) -> [bcIndex]
_ -> []
-- Initialize type/language to @language
-- and type/language value to @null
-- In other words, our default is Right clNull
(pref, containers') =
if reverze
then (Left ctReverse, bcSet : containers)
else case mvalue of
Just (AssocItem (ExpandedItemList el)) ->
handleList defLang containers el
Just (AssocItem (ExpandedItemOne (ExpandedAtomScalar esv))) ->
second (bcSet :) $
case scalarValue esv of
Left enb -> case enbType enb of
Just u -> (Left $ ctValue u, containers)
Nothing -> (Right clNull, containers)
Right es -> case esTypeOrLang es of
Nothing -> (Right clNull, containers)
Just torl -> case torl of
Left u -> (Left $ ctValue u, containers)
Right l -> case scalarIndex esv of
Nothing -> (Right $ clValue l, bcLang : containers)
Just _ -> (Right clNull, containers)
_ -> (Left ctId, bcSet : containers)
containers'' = bcNone : containers'
preferred :: Either [TypeMap -> Maybe Ref] [LanguageMap -> Maybe Ref]
preferred = case pref of
Left ct ->
let addReverse = case ctTag ct of
TMId -> Just id
TMReverse -> Just (tmReverse :)
_ -> Nothing
in case addReverse of
Just rev -> Left $ rev $ case mvalue of
Just (AssocItem (ExpandedItemOne (ExpandedAtomNode (ExpandedNodeObject { enodeId = Just i })))) ->
let i' = case i of
IdentURI u -> IdentKwURI u
IdentBlank b -> IdentKwBlank b
--t2r (TRef r) = Just r
--t2r _ = Nothing
mref = compactIdentKw Nothing True False active inverse i'
in case mref >>= flip M.lookup (activeTerms active) of
Just tdef | termTarget tdef == i' ->
[tmVocab, tmId, tmNone]
Nothing ->
[tmId, tmVocab, tmNone]
_ -> [ctGet ct, tmNone]
Nothing -> Left [ctGet ct, tmNone]
Right cl -> Right [clGet cl, lmNone]
in selectTerm cm containers'' preferred
where
clNone = CommonLanguage lmNone LMNone
clNull = CommonLanguage lmNull LMNull
clValue l = CommonLanguage (M.lookup l . lmValues) $ LMValue l
ctNone = CommonType tmNone TMNone
ctId = CommonType tmId TMId
ctValue u = CommonType (M.lookup u . tmValues) $ TMValue u
ctReverse = CommonType tmReverse TMReverse
compactIdentKwAsTerm
:: Maybe AssociatedValue
-> Bool
-> Bool
-> ActiveContext
-> InverseContext
-> IdentKw
-> Maybe Ref
compactIdentKwAsTerm mvalue vocab reverze active inverse target = do
guard vocab
cm <- M.lookup target inverse
compactToTerm mvalue reverze active inverse cm
compactIdentKwNotTerm
:: Bool
-> Bool
-> ActiveContext
-> IdentKw
-> Maybe Ref
compactIdentKwNotTerm valueNull vocab active target =
asVocab <|> asCompact <|> asRelative
where
asVocab = do
guard vocab
ident <- activeVocabulary active
case (target, ident) of
(IdentKwURI u, IdentURI v) -> asSuffixU u v
(IdentKwBlank b, IdentBlank v) -> asSuffixB b v
_ -> Nothing
where
asSuffixU u v = do
ref <- RefTerm <$> splitURI v u
guard $ isNothing $ M.lookup ref $ activeTerms active
Just ref
asSuffixB b v = do
ref <- RefTerm . relFromRelToken <$> splitRelToken v b
guard $ isNothing $ M.lookup ref $ activeTerms active
Just ref
asCompact = do
candidate <-
case target of
IdentKwURI u -> Just $ candidateU u
IdentKwBlank b -> Just $ candidateB b
-- TODO
-- There are weird cases in JSON-LD in which a compact URI can get
-- expanded into a keyword. These weird cases aren't supported in
-- this library. So if we have a keyword here, we shouldn't compact
-- it into a compact URI. One thing we could do is, check if it
-- works, and if yes, raise an error because we don't want to
-- produce that sort of compact URI. For now, we're just assuming
-- here that compaction into compact URI didn't succeed. Perhaps
-- later if test results suggest, update the code here to detect
-- such weird cases and raise an error.
_ -> Nothing
let candidates = M.mapMaybeWithKey candidate $ activeTerms active
if M.null candidates
then Nothing
else Just $ snd $ minimumBy (order `on` fst) $ M.elems candidates
where
candidateCheck compact =
let ref = RefURI $ compact2ac compact
in case M.lookup ref $ activeTerms active of
Just tdef
| termTarget tdef /= target || not valueNull -> Nothing
_ -> Just (compact, ref)
candidateU u (RefTerm t) tdef = do
p <- case termTarget tdef of
IdentKwURI q -> Just q
_ -> Nothing
c <- compactURI (renderRelNoAuth t) p u
guard $ p /= u
candidateCheck c
candidateU _ _ _ = Nothing
candidateB b (RefTerm t) tdef = do
p <- case termTarget tdef of
IdentKwBlank q -> Just q
_ -> Nothing
rt <- splitRelToken p b
c <- parseCompactFromRel (renderRelNoAuth t) (relFromRelToken rt)
candidateCheck c
candidateB _ _ _ = Nothing
order s t =
case B.length s' `compare` B.length t' of
LT -> LT
GT -> GT
EQ -> compare s' t'
where
s' = renderCompactURI s
t' = renderCompactURI t
asRelative = do
guard $ not vocab
case (target, activeBaseURI active) of
(IdentKwURI u, Just b) ->
RefTerm <$> (relNoAuthFromRelative =<< makeRelative b u)
_ -> Nothing
ident2ikw :: Identifier -> IdentKw
ident2ikw (IdentURI u) = IdentKwURI u
ident2ikw (IdentBlank b) = IdentKwBlank b
ident2ref :: Identifier -> Ref
ident2ref (IdentBlank b) = RefBlank b
ident2ref (IdentURI u) = RefURI $ uri2ac u
-- | This function tries to compact a URI, a blank node identifier or a keyword
-- into a 'Ref'. It implements the IRI compaction algorithm.
compactIdentKw
:: Maybe AssociatedValue
-> Bool
-> Bool
-> ActiveContext
-> InverseContext
-> IdentKw
-> Maybe Ref
compactIdentKw mvalue vocab reverze active inverse target =
compactIdentKwAsTerm mvalue vocab reverze active inverse target <|>
compactIdentKwNotTerm (isNothing mvalue) vocab active target
compactLoneId :: Bool -> ActiveContext -> InverseContext -> Identifier -> Ref
compactLoneId vocab active inverse ident =
case compactIdentKw Nothing vocab False active inverse $ ident2ikw ident of
Just ref -> ref
Nothing -> ident2ref ident
-- Compact a value given it's a node object (and not a value object)
compactValueNode :: ActiveContext -> InverseContext -> Ref -> ExpandedNodeObject -> Maybe Ref
compactValueNode active inverse property eno = do
ident <- onlyId eno
torl <- termTypeOrLang =<< mtd
case torl of
Left TermDefinitionTypeId -> Just $ loneId False ident
Left TermDefinitionTypeVocab -> Just $ loneId True ident
_ -> Nothing
where
mtd = M.lookup property $ activeTerms active
ignoreIndex Nothing = True
ignoreIndex (Just i) =
case termContainer =<< mtd of
Just ContainerIndex -> True
_ -> False
onlyId (ExpandedNodeObject (Just id_) Nothing Nothing Nothing mi vals) =
if M.null vals && ignoreIndex mi
then Just id_
else Nothing
onlyId _ = Nothing
loneId vocab = compactLoneId vocab active inverse
-- Nothing means we couldn't compact, just use the value as-is
compactValue :: ActiveContext -> InverseContext -> Ref -> ExpandedAtom -> Maybe TScalar
compactValue active inverse property ea =
case ea of
ExpandedAtomNode eno ->
ref2scalar <$> compactValueNode active inverse property eno
ExpandedAtomScalar (ExpandedScalarValue v mi) -> do
guard $ ignoreIndex mi
case v of
Left (ExpandedNBValue enb mt) ->
case (mt, termTypeOrLang =<< mtd) of
(Just u, Just (Left (TermDefinitionTypeAbsolute v)))
| u == v -> Just scalar
(Nothing, _) -> Just scalar
_ -> Nothing
where
scalar = enb2scalar enb
Right (ExpandedStringValue t tok mtorl) ->
case (mtorl, termTypeOrLang =<< mtd) of
(Just (Left u), Just (Left (TermDefinitionTypeAbsolute v)))
| u == v -> Just scalar
(Just (Right l), Just (Right (ContextLanguageTag l')))
| l == l' -> Just scalar
(Nothing, Just (Right ContextLanguageNull)) ->
Just scalar
(Nothing, _)
| isNothing $ activeLanguage active -> Just scalar
_ -> Nothing
where
scalar = TScalarString t tok
where
mtd = M.lookup property $ activeTerms active
ignoreIndex Nothing = True
ignoreIndex (Just i) =
case termContainer =<< mtd of
Just ContainerIndex -> True
_ -> False
ref2scalar ref = TScalarString (showRef ref) (TRef ref)
enb2scalar (Left s) = TScalarNumber s
enb2scalar (Right b) = TScalarBool b
compactDocument
:: Bool
-> ActiveContext
-> Vector ExpandedNodeObject
-> Either BereniceError NodeObject
compactDocument compactArrays active enos = do
nos <- traverse (compactNodeObject $ Left False) enos
let mno =
if compactArrays && V.length nos == 1
then Just $ V.head nos
else Nothing
mref = compactIdentKw' Nothing False False $ IdentKwNC KeywordGraph
Right $ case mno of
Just no -> no
Nothing -> NodeObject
{ nodeContext = active
, nodeId = Nothing
, nodeGraph = Just $ AliasAnd mref $ Many nos
, nodeType = Nothing
, nodeReverse = Nothing
, nodeIndex = Nothing
, nodeValues = M.empty
}
where
inverse = inverseContext active
compactIdentKw' mv v r = compactIdentKw mv v r active inverse
ne2multi iap (x :| []) =
if compactArrays && not (containerSetOrList iap)
then One x
else Many $ V.singleton x
where
containerSetOrList iap =
case termContainer =<< M.lookup iap (activeTerms active) of
Just ContainerSet -> True
Just ContainerList -> True
_ -> False
ne2multi _ (x :| xs) = Many $ V.fromList $ x : xs
compactReverseObject
:: HashMap Identifier (Vector ExpandedNodeObject)
-> Either
BereniceError
(HashMap Identifier (NonEmpty (Ref, (Bool, Multi NodeReverseItem))))
compactReverseObject = M.traverseWithKey compact
where
compactIdentKwAsTerm' mv v r = compactIdentKwAsTerm mv v r active inverse
compactIdentKwNotTerm' vn v = compactIdentKwNotTerm vn v active
n2ai = AssocItem . ExpandedItemOne . ExpandedAtomNode
compactAsTerm value =
compactIdentKwAsTerm' (Just $ n2ai value) True True . ident2ikw
compactNotTerm =
compactIdentKwNotTerm' False True . ident2ikw
iapNotTerm ep =
case compactNotTerm ep of
Just ref -> ref
Nothing -> ident2ref ep
compactENO iap eno =
case compactValueNode active inverse iap eno of
Nothing ->
NodeReverseObject <$> compactNodeObject (Right iap) eno
Just ref ->
Right $ NodeReverseId ref
isReverse iap =
case termReverse <$> M.lookup iap (activeTerms active) of
Just True -> True
_ -> False
ne2container iap ne =
case termContainer =<< M.lookup iap (activeTerms active) of
-- TODO we need to support language and index maps! When
-- container is @language or @index. Similarly, support them
-- when parsing and expanding a document. The idea is that when
-- we have a NonEmpty NRI, instead of making a Multi out of it,
-- we make an index map. Language maps don't make sense here;
-- the playground can tell me which error to give in that case.
Just ContainerLanguage -> error "TODO"
Just ContainerIndex -> error "TODO"
_ -> ne2multi iap ne
compactE ep =
let iap =
case compactIdentKw' (Just AssocArray) True True $
ident2ikw ep of
Just ref -> ref
Nothing -> ident2ref ep
in (iap, (isReverse iap, Many V.empty)) :| []
compactNE ep =
traverse
(\ (iap, enos) ->
(,) iap . (,) (isReverse iap) . ne2container iap <$>
traverse (compactENO iap) enos
) .
groupAllExtractDefault1 (iapNotTerm ep) (flip compactAsTerm ep) id
compact ep = maybe (pure $ compactE ep) (compactNE ep) . nonEmpty . V.toList
-- property: Either Bool Ref; Bool says whether it's @reverse
compactNodeObject property eno = do
let id_ =
enodeId eno <&> \ i -> (findAlias KeywordId, typeOrId False i)
typ =
enodeType eno <&> \ is ->
( findAlias KeywordType
, vec2multi $ V.map (typeOrId True) is
)
reverze' <- for (enodeReverse eno) $ \ m -> do
revmap <- compactReverseObject m
Right
( ( findAlias KeywordReverse
, filterReverse False revmap
)
, M.map (NE.map $ second $ Left . fmap nri2ni) $ filterReverse True revmap
)
let reverze = do
p <- fst <$> reverze'
if M.null $ snd p
then Nothing
else Just $ second (NodeReverse active) p
rrs = maybe M.empty snd reverze'
index =
if apContainerIndex property
then Nothing
else enodeIndex eno <&> \ t -> (findAlias KeywordIndex, t)
graph <- for (enodeGraph eno) $ \ gs -> do
gs' <- traverse (compactNodeObject $ Left False) gs
Right (findAlias KeywordGraph, gs')
values <- M.traverseWithKey compact $ enodeValues eno
Right NodeObject
{ nodeContext = active
, nodeId = uncurry AliasAnd <$> id_
, nodeGraph = uncurry AliasAnd <$> graph
, nodeType = uncurry AliasAnd <$> typ
, nodeReverse = uncurry AliasAnd <$> reverze
, nodeIndex = uncurry AliasAnd <$> index
, nodeValues = M.unionWith (error "TODO") values rrs
}
where
vec2multi v =
if V.length v == 1
then One $ V.unsafeHead v
else Many v
typeOrId typ i =
case compactIdentKw' Nothing typ False $ ident2ikw i of
Just ref -> ref
Nothing -> ident2ref i
findAlias = compactIdentKw' Nothing True False . IdentKwNC
filterReverse rev = M.mapMaybe $ \ ne ->
let filt (ref, (r, nris)) =
if r == rev
then Just (ref, nris)
else Nothing
in nonEmpty $ mapMaybe filt $ NE.toList ne
nri2ni (NodeReverseId ref) = NodeItemOne $ ItemScalar $ TScalarString (showRef ref) (TRef ref)
nri2ni (NodeReverseObject n) = NodeItemOne $ ItemNode n
apContainerIndex (Left _) = False
apContainerIndex (Right ap) =
case termContainer =<< M.lookup ap (activeTerms active) of
Just ContainerIndex -> True
_ -> False
compactIdentKwAsTerm' mv v r = compactIdentKwAsTerm mv v r active inverse
compactIdentKwNotTerm' vn v = compactIdentKwNotTerm vn v active
compactAsTerm value =
compactIdentKwAsTerm' (Just $ AssocItem value) True True . ident2ikw
compactNotTerm =
compactIdentKwNotTerm' False True . ident2ikw
iapNotTerm ep =
case compactNotTerm ep of
Just ref -> ref
Nothing -> ident2ref ep
compactEA iap ea =
case compactValue active inverse iap ea of
Just scalar -> Right $ ItemScalar scalar
Nothing -> case ea of
ExpandedAtomScalar esv ->
Right $ ItemValue ValueObject
{ valueValue =
AliasAnd (findAlias KeywordValue) $
ValueValueScalar $
case scalarValue esv of
Left enb ->
case enbValue enb of
Left s -> TScalarNumber s
Right b -> TScalarBool b
Right es ->
TScalarString (esText es) (esToken es)
, valueTypeOrLang =
case scalarValue esv of
Left enb -> AliasAnd (findAlias KeywordType) . Left . compactType <$> enbType enb
Right es -> uncurry AliasAnd . first findAlias . either ((,) KeywordType . Left . compactType) ((,) KeywordLanguage . Right) <$> esTypeOrLang es
, valueIndex =
if apContainerIndex $ Right iap
then Nothing
else AliasAnd (findAlias KeywordIndex) <$> scalarIndex esv
, valueContext = active
}
ExpandedAtomNode eno ->
ItemNode <$> compactNodeObject (Right iap) eno
where
compactType u =
case compactIdentKw' Nothing True False $ IdentKwURI u of
Just ref -> ref
Nothing -> ident2ref $ IdentURI u
ne2container iap nis =
case termContainer =<< M.lookup iap (activeTerms active) of
Just ContainerList ->
case mapMaybe loArr $ NE.toList nis of
[] -> Right $ Right $ Items $ Many $ ne2v nis
[items] -> Right $ Left $ NodeItemOne <$> items
_ -> Left $ SpecError CompactionToListOfLists ""
where
loArr (NodeItemList lo) = Just $ aaValue $ listArray' lo
loArr _ = Nothing
-- TODO it seems that the IRI compaction algo is supposed to
-- guarantee that all the 'NodeItem' in nis are valid for
-- language map / index map. But here this is not expressed in
-- the type. So for now, I'm using 'error' for these cases. If
-- they arise at run time, it probably means I have a bug. But
-- ideally, express those cases through the type. Perhaps
-- instead of a general (Ref, value) type, have a sum type with
-- ctors based on the container mapping of the Ref.
Just ContainerLanguage -> Right $ Right $ LangMap $ M.fromList $ map (second $ vec2multi . ne2v) $ groupAllExtract fst snd $ map langVal $ NE.toList nis
where
langVal (NodeItemOne (ItemValue (ValueObject (AliasAnd _ (ValueValueScalar (TScalarString t _))) (Just (AliasAnd _ (Right l))) _ _))) = (l, LanguageItemString t)
langVal _ = error "Hmmm should be impossible, do I have a bug?"
-- TODO we're supposed to extract the niIndex from the expanded
-- item, not from the compacted one, because the latter already
-- had it removed. Instead, something we can do is to keep the
-- indexes when compacting the node items, and *now* when we
-- build this index map, remove index values.
--
-- In the current situations, there are going to be errors
-- because the compacted items already have index removed.
Just ContainerIndex -> Right $ Right $ IndexMap $ M.fromList $ map (second $ vec2multi . ne2v) $ groupAllExtract fst snd $ map (niIndex &&& id) $ NE.toList nis
where
niIndex (NodeItemOne (ItemNode (NodeObject { nodeIndex = Just (AliasAnd _ i) }))) = i
niIndex (NodeItemList (ListObject' { listIndex' = Just (AliasAnd _ i) })) = i
niIndex (NodeItemSet (SetObject' { setIndex' = Just (AliasAnd _ i) })) = i
niIndex _ = error "Hmmm should be impossible, do I have a bug?"
_ -> Right $ Right $ Items $ ne2multi iap nis
where
ne2v = V.fromList . NE.toList
compactEI iap (ExpandedItemOne ea) = NodeItemOne <$> compactEA iap ea
compactEI iap (ExpandedItemList (ExpandedList mi v)) =
nilo <$> traverse (compactEA iap) v
where
nilo items = NodeItemList ListObject'
{ listArray' = AliasAnd (findAlias KeywordList) $ vec2multi items
, listContext' = active
, listIndex' = AliasAnd (findAlias KeywordIndex) <$> mi
}
compactE ep =
let iap =
case compactIdentKw' (Just AssocArray) True True $
ident2ikw ep of
Just ref -> ref
Nothing -> ident2ref ep
in (iap, Right $ Items $ Many V.empty) :| []
compactNE ep =
traverse
(\ (iap, eis) ->
(,) iap <$> (ne2container iap =<< (traverse (compactEI iap) eis))
) .
groupAllExtractDefault1 (iapNotTerm ep) (flip compactAsTerm ep) id
compact ep = maybe (pure $ compactE ep) (compactNE ep) . nonEmpty . V.toList
-- The code for the expansion algorithm used to look scary to me for most of
-- the period of time I was writing it. Eventually I started feeling
-- comfortable with it. Now, the compaction code seems scary to me. I wonder if
-- this feeling will pass too.
--
-- We're done compacting! The next step is generalization. We convert the
-- various object types (node object, value object, etc.) info a uniform
-- representation. We're getting a step closer to something that looks like
-- plain JSON.
-- | Take a representation that is aware of specific object types (such as node
-- object, value object, etc.) and produce a general object representation.
--
-- NOTE: The 'TaggedItemObject' constructor of the 'TaggedItem' has 2
-- parameters. When expanding, one is used for language maps and index maps,
-- and the other for everything else. Here, when compacting. we produce only
-- one of these two values, and leave the other empty: When compacting a
-- language map or index map, we produce the same map used for expanding them.
-- Otherwise, we produce the other value.
--
-- When untagging the produced 'TaggedObject', simply check which one of the
-- two values in an empty map, and use the other for the actual untagging.
generalizeDocument :: NodeObject -> TaggedObject
generalizeDocument = generalizeNode
where
generalizeNode (NodeObject active mid mg mt mr mi vs) = TaggedObject
{ taggedContext = active
, taggedKwMap = M.fromList $ catMaybes
[ kw mid KeywordId $ One . ref2ti
, kw mg KeywordGraph $ fmap node2ti
, kw mt KeywordType $ fmap ref2ti
, kw mr KeywordReverse $ One . flip TaggedItemObject M.empty . reverse2ti
, kw mi KeywordIndex $ One . text2ti
]
, taggedIdMap = M.map (NE.map $ second $ nv2tv active) vs
}
where
okey kw Nothing = ObjectKeyKeyword kw
okey _ (Just ref) = ObjectKeyOther ref
kw' (AliasAnd ma v) kwnc f = (kwnc, (okey kwnc ma, f v))
kw Nothing _ _ = Nothing
kw (Just aa) kwnc f = Just $ kw' aa kwnc f
ref2ti ref = TaggedItemScalar $ TScalarString (showRef ref) (TRef ref)
text2ti t = TaggedItemScalar $ TScalarString t $ parseToken t
node2ti n = TaggedItemObject (generalizeNode n) M.empty
reverse2ti (NodeReverse c m) = TaggedObject
{ taggedContext = c
, taggedKwMap = M.empty
, taggedIdMap = M.map (NE.map $ second $ fmap nri2ti) m
}
where
nri2ti (NodeReverseId ref) = ref2ti ref
nri2ti (NodeReverseObject n) = node2ti n
nv2tv a v =
case v of
Left nis -> ni2ti <$> nis
Right nv ->
case nv of
Items nis -> ni2ti <$> nis
LangMap lm -> One $ TaggedItemObject (emptyTO a) $ M.map (fmap li2ti) lm
IndexMap im -> One $ TaggedItemObject (emptyTO a) $ M.map (fmap ni2ti) im
where
emptyTO a = TaggedObject a M.empty M.empty
loso2to ais c mi = TaggedObject
{ taggedContext = c
, taggedKwMap = M.fromList $ catMaybes
[ kw (Just ais) KeywordList $ fmap i2ti
, kw mi KeywordIndex $ One . text2ti
]
, taggedIdMap = M.empty
}
lo2to (ListObject' ais c mi) = loso2to ais c mi
so2to (SetObject' ais c mi) = loso2to ais c mi
vo2to (ValueObject valueA mtorlA mindexA active) = TaggedObject
{ taggedContext = active
, taggedKwMap = M.fromList $ catMaybes
[ kw (Just valueA) KeywordValue $ One . vv2ti
, kw mindexA KeywordIndex $ One . text2ti
, mtorlA <&> \ (AliasAnd aa torl) ->
case torl of
Left typ -> kw' (AliasAnd aa typ) KeywordType $ One . ref2ti
Right lang -> kw' (AliasAnd aa lang) KeywordLanguage $ One . text2ti
]
, taggedIdMap = M.empty
}
where
vv2ti (ValueValueScalar s) = TaggedItemScalar s
vv2ti ValueValueNull = TaggedItemNull
i2ti (ItemScalar s) = TaggedItemScalar s
i2ti ItemNull = TaggedItemNull
i2ti (ItemNode n) = node2ti n
i2ti (ItemValue vo) = TaggedItemObject (vo2to vo) M.empty
ni2ti (NodeItemOne i) = i2ti i
ni2ti (NodeItemList lo) = TaggedItemObject (lo2to lo) M.empty
ni2ti (NodeItemSet so) = TaggedItemObject (so2to so) M.empty
li2ti (LanguageItemNull) = TaggedItemNull
li2ti (LanguageItemString t) = text2ti t
-- Now comes the untagging step, in which we drop the expanded properties and
-- organize values under their compacted properties.
renderKwNC :: KeywordNC -> Text
renderKwNC KeywordId = "@id"
renderKwNC KeywordValue = "@value"
renderKwNC KeywordLanguage = "@language"
renderKwNC KeywordType = "@type"
renderKwNC KeywordList = "@list"
renderKwNC KeywordSet = "@set"
renderKwNC KeywordReverse = "@reverse"
renderKwNC KeywordIndex = "@index"
renderKwNC KeywordGraph = "@graph"
untagDocument :: Multi LocalContextItem -> TaggedObject -> UnknownObject
untagDocument c o = UnknownObject (Just c) (untagObject o)
where
untagObject (TaggedObject _ kws vals) =
let kws' = M.elems kws
vals' = map (first ObjectKeyOther) $ concatMap NE.toList $ M.elems vals
in M.fromList $ map ((renderObjectKey . fst) &&& second (fmap untagItem)) $ kws' ++ vals'
where
renderObjectKey (ObjectKeyKeyword kw) = renderKwNC kw
renderObjectKey (ObjectKeyOther ref) = showRef ref
untagItem (TaggedItemScalar s) = UnknownItemScalar s
untagItem TaggedItemNull = UnknownItemNull
untagItem (TaggedItemObject to m) =
UnknownItemObject $ UnknownObject Nothing $
if nullTO to
then M.map ((,) okey . fmap untagItem) m
else untagObject to
where
nullTO (TaggedObject _ kws vals) = M.null kws && M.null vals
okey = error "BUG: ObjectKey in compacted unknownMap used"
-- Finally comes the rendering step, which produces a JSON-like structure you
-- can serialize with a JSON library such as @aeson@.
renderDocument :: UnknownObject -> Dictionary
renderDocument = renderObject
where
uri2scalar = ScalarString . decodeUtf8 . renderURI
renderContextItem LocalContextNull = ValueNull
renderContextItem (LocalContextString u) = ValueScalar $ uri2scalar u
renderContextItem (LocalContextObject (Context ml mb mv vs)) =
ValueDictionary $ M.fromList $
consField "@language" ml renderLang $
consField "@base" mb renderBase $
consField "@vocab" mv renderVocab $
map (bimap showRef $ One . renderContextValue) $ M.toList vs
where
consField _ Nothing _ = id
consField t (Just v) f = (:) (t, One $ f v)
renderLang (ContextLanguageTag t) = ValueScalar $ ScalarString t
renderLang ContextLanguageNull = ValueNull
renderBase (ContextBaseAbsolute u) = ValueScalar $ uri2scalar u
renderBase (ContextBaseRelative u) = ValueScalar $ uri2scalar u
renderBase ContextBaseNull = ValueNull
renderVocab (ContextVocabAbsolute u) = ValueScalar $ uri2scalar u
renderVocab ContextVocabNull = ValueNull
renderVocab (ContextVocabBlank rt) =
ValueScalar $ ScalarString $ decodeUtf8 $ "_:" <> renderRelToken rt
renderId IdNull = ValueNull
renderId (IdRef ref) = ValueScalar $ ScalarString $ showRef ref
renderId (IdKeyword kw) = ValueScalar $ ScalarString $ renderKw kw
where
renderKw KeywordContext = "@context"
renderKw (KeywordNC kw) = renderKwNC kw
renderKw (KeywordC kw) = renderKwC kw
where
renderKwC KeywordContainer = "@container"
renderKwC KeywordBase = "@base"
renderKwC KeywordVocab = "@vocab"
renderETD (ExpandedTermDefinition idr mt ml) =
ValueDictionary $ M.fromList $ consIdReverse idr $ catMaybes
[ field "@type" mt renderType
, field "@language" ml renderLang
]
where
field _ Nothing _ = Nothing
field t (Just v) f = Just (t, One $ f v)
bs2value = ValueScalar . ScalarString . decodeUtf8
renderType (TypeURI ac) = bs2value $ renderAC ac
renderType (TypeTerm rna) = bs2value $ renderRelNoAuth rna
renderType TypeId = ValueScalar $ ScalarString "@id"
renderType TypeVocab = ValueScalar $ ScalarString "@vocab"
renderType TypeNull = ValueNull
t2v = ValueScalar . ScalarString
consIdReverse (Left (ExpandedTermDefinitionId mid mc)) =
consField "@id" mid renderId .
consField "@container" mc renderContainer
where
renderContainer ContainerNull = ValueNull
renderContainer ContainerSet = t2v "@set"
renderContainer ContainerList = t2v "@list"
renderContainer ContainerLanguage = t2v "@language"
renderContainer ContainerIndex = t2v "@index"
consIdReverse (Right (ExpandedTermDefinitionReverse r mc)) =
consField "@reverse" (Just r) (t2v . showRef) .
consField "@container" mc renderContainerReverse
where
renderContainerReverse ContainerReverseNull = ValueNull
renderContainerReverse ContainerReverseSet = t2v "@set"
renderContainerReverse ContainerReverseIndex = t2v "@index"
renderContextValue (ContextValueId id_) = renderId id_
renderContextValue (ContextValueExpanded etd) = renderETD etd
renderObject (UnknownObject mc m) =
maybe id (M.insert "@context" . fmap renderContextItem) mc $
M.map (fmap renderItem . snd) m
renderItem UnknownItemNull = ValueNull
renderItem (UnknownItemObject uo) = ValueDictionary $ renderObject uo
renderItem (UnknownItemScalar s) =
ValueScalar $
case s of
TScalarString t _ -> ScalarString t
TScalarNumber n -> ScalarNumber n
TScalarBool b -> ScalarBool b
-- Now we can define a function that does the whole process, from RDF to JSON:
rdf2jsonld
:: UseNativeTypes
-> UseRdfType
-> Bool
-> Multi LocalContextItem
-> ActiveContext
-> RdfDataset
-> Either BereniceError Dictionary
rdf2jsonld native rdftype compactArrays context active rdf = do
nm <- rdf2nm native rdftype rdf
let enos = renderNodeMap nm
no <- compactDocument compactArrays active enos
let to = generalizeDocument no
uo = untagDocument context to
Right $ renderDocument uo
|