icuSources/i18n/collationbuilder.h

   1 // © 2016 and later: Unicode, Inc. and others.
   2 // License & terms of use: http://www.unicode.org/copyright.html
   3 /*
   4 *******************************************************************************
   5 * Copyright (C) 2013-2014, International Business Machines
   6 * Corporation and others.  All Rights Reserved.
   7 *******************************************************************************
   8 * collationbuilder.h
   9 *
  10 * created on: 2013may06
  11 * created by: Markus W. Scherer
  12 */
  13
  14 #ifndef __COLLATIONBUILDER_H__
  15 #define __COLLATIONBUILDER_H__
  16
  17 #include "unicode/utypes.h"
  18
  19 #if !UCONFIG_NO_COLLATION
  20
  21 #include "unicode/uniset.h"
  22 #include "unicode/unistr.h"
  23 #include "collationrootelements.h"
  24 #include "collationruleparser.h"
  25 #include "uvectr32.h"
  26 #include "uvectr64.h"
  27
  28 struct UParseError;
  29
  30 U_NAMESPACE_BEGIN
  31
  32 struct CollationData;
  33 struct CollationTailoring;
  34
  35 class CEFinalizer;
  36 class CollationDataBuilder;
  37 class Normalizer2;
  38 class Normalizer2Impl;
  39
  40 class U_I18N_API CollationBuilder : public CollationRuleParser::Sink {
  41 public:
  42     CollationBuilder(const CollationTailoring *base, UErrorCode &errorCode);
  43     virtual ~CollationBuilder();
  44
  45     void disableFastLatin() { fastLatinEnabled = FALSE; }
  46
  47     CollationTailoring *parseAndBuild(const UnicodeString &ruleString,
  48                                       const UVersionInfo rulesVersion,
  49                                       CollationRuleParser::Importer *importer,
  50                                       UParseError *outParseError,
  51                                       UErrorCode &errorCode);
  52
  53     const char *getErrorReason() const { return errorReason; }
  54
  55 private:
  56     friend class CEFinalizer;
  57
  58     /** Implements CollationRuleParser::Sink. */
  59     virtual void addReset(int32_t strength, const UnicodeString &str,
  60                           const char *&errorReason, UErrorCode &errorCode);
  61     /**
  62      * Returns the secondary or tertiary weight preceding the current node's weight.
  63      * node=nodes[index].
  64      */
  65     uint32_t getWeight16Before(int32_t index, int64_t node, int32_t level);
  66
  67     int64_t getSpecialResetPosition(const UnicodeString &str,
  68                                     const char *&parserErrorReason, UErrorCode &errorCode);
  69
  70     /** Implements CollationRuleParser::Sink. */
  71     virtual void addRelation(int32_t strength, const UnicodeString &prefix,
  72                              const UnicodeString &str, const UnicodeString &extension,
  73                              const char *&errorReason, UErrorCode &errorCode);
  74
  75     /**
  76      * Picks one of the current CEs and finds or inserts a node in the graph
  77      * for the CE + strength.
  78      */
  79     int32_t findOrInsertNodeForCEs(int32_t strength, const char *&parserErrorReason,
  80                                    UErrorCode &errorCode);
  81     int32_t findOrInsertNodeForRootCE(int64_t ce, int32_t strength, UErrorCode &errorCode);
  82     /** Finds or inserts the node for a root CE's primary weight. */
  83     int32_t findOrInsertNodeForPrimary(uint32_t p, UErrorCode &errorCode);
  84     /** Finds or inserts the node for a secondary or tertiary weight. */
  85     int32_t findOrInsertWeakNode(int32_t index, uint32_t weight16, int32_t level,
  86                                  UErrorCode &errorCode);
  87
  88     /**
  89      * Makes and inserts a new tailored node into the list, after the one at index.
  90      * Skips over nodes of weaker strength to maintain collation order
  91      * ("postpone insertion").
  92      * @return the new node's index
  93      */
  94     int32_t insertTailoredNodeAfter(int32_t index, int32_t strength, UErrorCode &errorCode);
  95
  96     /**
  97      * Inserts a new node into the list, between list-adjacent items.
  98      * The node's previous and next indexes must not be set yet.
  99      * @return the new node's index
 100      */
 101     int32_t insertNodeBetween(int32_t index, int32_t nextIndex, int64_t node,
 102                               UErrorCode &errorCode);
 103
 104     /**
 105      * Finds the node which implies or contains a common=05 weight of the given strength
 106      * (secondary or tertiary), if the current node is stronger.
 107      * Skips weaker nodes and tailored nodes if the current node is stronger
 108      * and is followed by an explicit-common-weight node.
 109      * Always returns the input index if that node is no stronger than the given strength.
 110      */
 111     int32_t findCommonNode(int32_t index, int32_t strength) const;
 112
 113     void setCaseBits(const UnicodeString &nfdString,
 114                      const char *&parserErrorReason, UErrorCode &errorCode);
 115
 116     /** Implements CollationRuleParser::Sink. */
 117     virtual void suppressContractions(const UnicodeSet &set, const char *&parserErrorReason,
 118                                       UErrorCode &errorCode);
 119
 120     /** Implements CollationRuleParser::Sink. */
 121     virtual void optimize(const UnicodeSet &set, const char *&parserErrorReason,
 122                           UErrorCode &errorCode);
 123
 124     /**
 125      * Adds the mapping and its canonical closure.
 126      * Takes ce32=dataBuilder->encodeCEs(...) so that the data builder
 127      * need not re-encode the CEs multiple times.
 128      */
 129     uint32_t addWithClosure(const UnicodeString &nfdPrefix, const UnicodeString &nfdString,
 130                             const int64_t newCEs[], int32_t newCEsLength, uint32_t ce32,
 131                             UErrorCode &errorCode);
 132     uint32_t addOnlyClosure(const UnicodeString &nfdPrefix, const UnicodeString &nfdString,
 133                             const int64_t newCEs[], int32_t newCEsLength, uint32_t ce32,
 134                             UErrorCode &errorCode);
 135     void addTailComposites(const UnicodeString &nfdPrefix, const UnicodeString &nfdString,
 136                            UErrorCode &errorCode);
 137     UBool mergeCompositeIntoString(const UnicodeString &nfdString, int32_t indexAfterLastStarter,
 138                                    UChar32 composite, const UnicodeString &decomp,
 139                                    UnicodeString &newNFDString, UnicodeString &newString,
 140                                    UErrorCode &errorCode) const;
 141
 142     UBool ignorePrefix(const UnicodeString &s, UErrorCode &errorCode) const;
 143     UBool ignoreString(const UnicodeString &s, UErrorCode &errorCode) const;
 144     UBool isFCD(const UnicodeString &s, UErrorCode &errorCode) const;
 145
 146     void closeOverComposites(UErrorCode &errorCode);
 147
 148     uint32_t addIfDifferent(const UnicodeString &prefix, const UnicodeString &str,
 149                             const int64_t newCEs[], int32_t newCEsLength, uint32_t ce32,
 150                             UErrorCode &errorCode);
 151     static UBool sameCEs(const int64_t ces1[], int32_t ces1Length,
 152                          const int64_t ces2[], int32_t ces2Length);
 153
 154     /**
 155      * Walks the tailoring graph and overwrites tailored nodes with new CEs.
 156      * After this, the graph is destroyed.
 157      * The nodes array can then be used only as a source of tailored CEs.
 158      */
 159     void makeTailoredCEs(UErrorCode &errorCode);
 160     /**
 161      * Counts the tailored nodes of the given strength up to the next node
 162      * which is either stronger or has an explicit weight of this strength.
 163      */
 164     static int32_t countTailoredNodes(const int64_t *nodesArray, int32_t i, int32_t strength);
 165
 166     /** Replaces temporary CEs with the final CEs they point to. */
 167     void finalizeCEs(UErrorCode &errorCode);
 168
 169     /**
 170      * Encodes "temporary CE" data into a CE that fits into the CE32 data structure,
 171      * with 2-byte primary, 1-byte secondary and 6-bit tertiary,
 172      * with valid CE byte values.
 173      *
 174      * The index must not exceed 20 bits (0xfffff).
 175      * The strength must fit into 2 bits (UCOL_PRIMARY..UCOL_QUATERNARY).
 176      *
 177      * Temporary CEs are distinguished from real CEs by their use of
 178      * secondary weights 06..45 which are otherwise reserved for compressed sort keys.
 179      *
 180      * The case bits are unused and available.
 181      */
 182     static inline int64_t tempCEFromIndexAndStrength(int32_t index, int32_t strength) {
 183         return
 184             // CE byte offsets, to ensure valid CE bytes, and case bits 11
 185             INT64_C(0x4040000006002000) +
 186             // index bits 19..13 -> primary byte 1 = CE bits 63..56 (byte values 40..BF)
 187             ((int64_t)(index & 0xfe000) << 43) +
 188             // index bits 12..6 -> primary byte 2 = CE bits 55..48 (byte values 40..BF)
 189             ((int64_t)(index & 0x1fc0) << 42) +
 190             // index bits 5..0 -> secondary byte 1 = CE bits 31..24 (byte values 06..45)
 191             ((index & 0x3f) << 24) +
 192             // strength bits 1..0 -> tertiary byte 1 = CE bits 13..8 (byte values 20..23)
 193             (strength << 8);
 194     }
 195     static inline int32_t indexFromTempCE(int64_t tempCE) {
 196         tempCE -= INT64_C(0x4040000006002000);
 197         return
 198             ((int32_t)(tempCE >> 43) & 0xfe000) |
 199             ((int32_t)(tempCE >> 42) & 0x1fc0) |
 200             ((int32_t)(tempCE >> 24) & 0x3f);
 201     }
 202     static inline int32_t strengthFromTempCE(int64_t tempCE) {
 203         return ((int32_t)tempCE >> 8) & 3;
 204     }
 205     static inline UBool isTempCE(int64_t ce) {
 206         uint32_t sec = (uint32_t)ce >> 24;
 207         return 6 <= sec && sec <= 0x45;
 208     }
 209
 210     static inline int32_t indexFromTempCE32(uint32_t tempCE32) {
 211         tempCE32 -= 0x40400620;
 212         return
 213             ((int32_t)(tempCE32 >> 11) & 0xfe000) |
 214             ((int32_t)(tempCE32 >> 10) & 0x1fc0) |
 215             ((int32_t)(tempCE32 >> 8) & 0x3f);
 216     }
 217     static inline UBool isTempCE32(uint32_t ce32) {
 218         return
 219             (ce32 & 0xff) >= 2 &&  // not a long-primary/long-secondary CE32
 220             6 <= ((ce32 >> 8) & 0xff) && ((ce32 >> 8) & 0xff) <= 0x45;
 221     }
 222
 223     static int32_t ceStrength(int64_t ce);
 224
 225     /** At most 1M nodes, limited by the 20 bits in node bit fields. */
 226     static const int32_t MAX_INDEX = 0xfffff;
 227     /**
 228      * Node bit 6 is set on a primary node if there are nodes
 229      * with secondary values below the common secondary weight (05).
 230      */
 231     static const int32_t HAS_BEFORE2 = 0x40;
 232     /**
 233      * Node bit 5 is set on a primary or secondary node if there are nodes
 234      * with tertiary values below the common tertiary weight (05).
 235      */
 236     static const int32_t HAS_BEFORE3 = 0x20;
 237     /**
 238      * Node bit 3 distinguishes a tailored node, which has no weight value,
 239      * from a node with an explicit (root or default) weight.
 240      */
 241     static const int32_t IS_TAILORED = 8;
 242
 243     static inline int64_t nodeFromWeight32(uint32_t weight32) {
 244         return (int64_t)weight32 << 32;
 245     }
 246     static inline int64_t nodeFromWeight16(uint32_t weight16) {
 247         return (int64_t)weight16 << 48;
 248     }
 249     static inline int64_t nodeFromPreviousIndex(int32_t previous) {
 250         return (int64_t)previous << 28;
 251     }
 252     static inline int64_t nodeFromNextIndex(int32_t next) {
 253         return next << 8;
 254     }
 255     static inline int64_t nodeFromStrength(int32_t strength) {
 256         return strength;
 257     }
 258
 259     static inline uint32_t weight32FromNode(int64_t node) {
 260         return (uint32_t)(node >> 32);
 261     }
 262     static inline uint32_t weight16FromNode(int64_t node) {
 263         return (uint32_t)(node >> 48) & 0xffff;
 264     }
 265     static inline int32_t previousIndexFromNode(int64_t node) {
 266         return (int32_t)(node >> 28) & MAX_INDEX;
 267     }
 268     static inline int32_t nextIndexFromNode(int64_t node) {
 269         return ((int32_t)node >> 8) & MAX_INDEX;
 270     }
 271     static inline int32_t strengthFromNode(int64_t node) {
 272         return (int32_t)node & 3;
 273     }
 274
 275     static inline UBool nodeHasBefore2(int64_t node) {
 276         return (node & HAS_BEFORE2) != 0;
 277     }
 278     static inline UBool nodeHasBefore3(int64_t node) {
 279         return (node & HAS_BEFORE3) != 0;
 280     }
 281     static inline UBool nodeHasAnyBefore(int64_t node) {
 282         return (node & (HAS_BEFORE2 | HAS_BEFORE3)) != 0;
 283     }
 284     static inline UBool isTailoredNode(int64_t node) {
 285         return (node & IS_TAILORED) != 0;
 286     }
 287
 288     static inline int64_t changeNodePreviousIndex(int64_t node, int32_t previous) {
 289         return (node & INT64_C(0xffff00000fffffff)) | nodeFromPreviousIndex(previous);
 290     }
 291     static inline int64_t changeNodeNextIndex(int64_t node, int32_t next) {
 292         return (node & INT64_C(0xfffffffff00000ff)) | nodeFromNextIndex(next);
 293     }
 294
 295     const Normalizer2 &nfd, &fcd;
 296     const Normalizer2Impl &nfcImpl;
 297
 298     const CollationTailoring *base;
 299     const CollationData *baseData;
 300     const CollationRootElements rootElements;
 301     uint32_t variableTop;
 302
 303     CollationDataBuilder *dataBuilder;
 304     UBool fastLatinEnabled;
 305     UnicodeSet optimizeSet;
 306     const char *errorReason;
 307
 308     int64_t ces[Collation::MAX_EXPANSION_LENGTH];
 309     int32_t cesLength;
 310
 311     /**
 312      * Indexes of nodes with root primary weights, sorted by primary.
 313      * Compact form of a TreeMap from root primary to node index.
 314      *
 315      * This is a performance optimization for finding reset positions.
 316      * Without this, we would have to search through the entire nodes list.
 317      * It also allows storing root primary weights in list head nodes,
 318      * without previous index, leaving room in root primary nodes for 32-bit primary weights.
 319      */
 320     UVector32 rootPrimaryIndexes;
 321     /**
 322      * Data structure for assigning tailored weights and CEs.
 323      * Doubly-linked lists of nodes in mostly collation order.
 324      * Each list starts with a root primary node and ends with a nextIndex of 0.
 325      *
 326      * When there are any nodes in the list, then there is always a root primary node at index 0.
 327      * This allows some code not to have to check explicitly for nextIndex==0.
 328      *
 329      * Root primary nodes have 32-bit weights but do not have previous indexes.
 330      * All other nodes have at most 16-bit weights and do have previous indexes.
 331      *
 332      * Nodes with explicit weights store root collator weights,
 333      * or default weak weights (e.g., secondary 05) for stronger nodes.
 334      * "Tailored" nodes, with the IS_TAILORED bit set,
 335      * do not store explicit weights but rather
 336      * create a difference of a certain strength from the preceding node.
 337      *
 338      * A root node is followed by either
 339      * - a root/default node of the same strength, or
 340      * - a root/default node of the next-weaker strength, or
 341      * - a tailored node of the same strength.
 342      *
 343      * A node of a given strength normally implies "common" weights on weaker levels.
 344      *
 345      * A node with HAS_BEFORE2 must be immediately followed by
 346      * a secondary node with an explicit below-common weight, then a secondary tailored node,
 347      * and later an explicit common-secondary node.
 348      * The below-common weight can be a root weight,
 349      * or it can be BEFORE_WEIGHT16 for tailoring before an implied common weight
 350      * or before the lowest root weight.
 351      * (&[before 2] resets to an explicit secondary node so that
 352      * the following addRelation(secondary) tailors right after that.
 353      * If we did not have this node and instead were to reset on the primary node,
 354      * then addRelation(secondary) would skip forward to the the COMMON_WEIGHT16 node.)
 355      *
 356      * If the flag is not set, then there are no explicit secondary nodes
 357      * with the common or lower weights.
 358      *
 359      * Same for HAS_BEFORE3 for tertiary nodes and weights.
 360      * A node must not have both flags set.
 361      *
 362      * Tailored CEs are initially represented in a CollationDataBuilder as temporary CEs
 363      * which point to stable indexes in this list,
 364      * and temporary CEs stored in a CollationDataBuilder only point to tailored nodes.
 365      *
 366      * A temporary CE in the ces[] array may point to a non-tailored reset-before-position node,
 367      * until the next relation is added.
 368      *
 369      * At the end, the tailored weights are allocated as necessary,
 370      * then the tailored nodes are replaced with final CEs,
 371      * and the CollationData is rewritten by replacing temporary CEs with final ones.
 372      *
 373      * We cannot simply insert new nodes in the middle of the array
 374      * because that would invalidate the indexes stored in existing temporary CEs.
 375      * We need to use a linked graph with stable indexes to existing nodes.
 376      * A doubly-linked list seems easiest to maintain.
 377      *
 378      * Each node is stored as an int64_t, with its fields stored as bit fields.
 379      *
 380      * Root primary node:
 381      * - primary weight: 32 bits 63..32
 382      * - reserved/unused/zero: 4 bits 31..28
 383      *
 384      * Weaker root nodes & tailored nodes:
 385      * - a weight: 16 bits 63..48
 386      *   + a root or default weight for a non-tailored node
 387      *   + unused/zero for a tailored node
 388      * - index to the previous node: 20 bits 47..28
 389      *
 390      * All types of nodes:
 391      * - index to the next node: 20 bits 27..8
 392      *   + nextIndex=0 in last node per root-primary list
 393      * - reserved/unused/zero bits: bits 7, 4, 2
 394      * - HAS_BEFORE2: bit 6
 395      * - HAS_BEFORE3: bit 5
 396      * - IS_TAILORED: bit 3
 397      * - the difference strength (primary/secondary/tertiary/quaternary): 2 bits 1..0
 398      *
 399      * We could allocate structs with pointers, but we would have to store them
 400      * in a pointer list so that they can be indexed from temporary CEs,
 401      * and they would require more memory allocations.
 402      */
 403     UVector64 nodes;
 404 };
 405
 406 U_NAMESPACE_END
 407
 408 #endif  // !UCONFIG_NO_COLLATION
 409 #endif  // __COLLATIONBUILDER_H__