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 | 1× 1× 1× 180× 180× 180× 145× 145× 145× 431× 180× 180× 180× 180× 283× 283× 283× 257× 283× 283× 283× 143× 5× 1× 143× 1× 1× 1× 143× 143× 144× 144× 144× 283× 143× 147× 143× 143× 144× 144× 144× 143× 143× 1× 1× 1× 180× 180× 174× 1× 1× 1× 1× 1× 180× 180× 180× 143× 143× 3× 140× 140× 140× 28× 112× 97× 15× 1× 140× 140× 140× 140× 140× 180× 180× 143× 180× 180× 180× 180× 180× 180× 180× 180× 283× 158× 16× 16× 142× 142× 125× 1× 1× 124× 124× 124× 283× 283× 283× 180× 1× 22× 3× 3× 3× 2× 2× 2× 2× 1× 20× 3× 2× 1× 1× 1× 1× | import { forEach, isString } from '../util' const ENTER = 1 const EXIT = -1 const ANCHOR = -2 // Fragmenter // -------- // // An algorithm that is used to fragment overlapping structure elements // following a priority rule set. // E.g., we use this for creating DOM elements for annotations. The annotations // can partially be overlapping. However this is not allowed in general for DOM elements // or other hierarchical structures. // // Example: For the annotation use case consider a 'comment' spanning partially // over an 'emphasis' annotation. // 'The <comment>quick brown <bold>fox</comment> jumps over</bold> the lazy dog.' // We want to be able to create a valid XML structure: // 'The <comment>quick brown <bold>fox</bold></comment><bold> jumps over</bold> the lazy dog.' // // For that one would choose // // { // 'comment': 0, // 'bold': 1 // } // // as priority levels. // In case of structural violations as in the example, elements with a higher level // would be fragmented and those with lower levels would be preserved as one piece. // // TODO: If a violation for nodes of the same level occurs an Error should be thrown. // Currently, in such cases the first element that is opened earlier is preserved. class Fragmenter { constructor(options) { Object.assign(this, options) } start(rootContext, text, annotations) { Iif (!isString(text)) { throw new Error("Illegal argument: 'text' must be a String, but was " + text) } this._start(rootContext, text, annotations) } onText(context, text, entry) { // eslint-disable-line } // should return the created user context onEnter(entry, parentContext) { // eslint-disable-line return null } onExit(entry, context, parentContext) { // eslint-disable-line } _enter(entry, parentContext) { entry.counter++ return this.onEnter(entry, parentContext) } _exit(entry, context, parentContext) { this.onExit(entry, context, parentContext) } _createText(context, text, entry) { this.onText(context, text, entry) } _start(rootContext, text, annotations) { var entries = _extractEntries.call(this, annotations) var stack = [{context: rootContext, entry: null}] var pos = 0 for (var i = 0; i < entries.length; i++) { var entry = entries[i] var textFragment = text.substring(pos, entry.pos) if (textFragment) { // add the last text to the current element this._createText(stack[stack.length-1].context, textFragment, entry) } pos = entry.pos var stackLevel, idx, _entry if (entry.mode === ENTER || entry.mode === ANCHOR) { // find the correct position and insert an entry for (stackLevel = 1; stackLevel < stack.length; stackLevel++) { if (entry.level < stack[stackLevel].entry.level) { break } } // create elements which are open, and are now stacked ontop of the // entered entry for (idx = stack.length-1; idx >= stackLevel; idx--) { _entry = stack[idx].entry // compute number of characters since last 'enter' _entry.length = pos - _entry.pos this._exit(_entry, stack[idx].context, stack[idx-1].context) } stack.splice(stackLevel, 0, {entry: entry}) // create new elements for all lower entries for (idx = stackLevel; idx < stack.length; idx++) { _entry = stack[idx].entry // bump 'enter' pos _entry.pos = pos stack[idx].context = this._enter(_entry, stack[idx-1].context) } } if (entry.mode === EXIT || entry.mode === ANCHOR) { // find the according entry and remove it from the stack for (stackLevel = 1; stackLevel < stack.length; stackLevel++) { if (stack[stackLevel].entry.node === entry.node) { break } } for (idx = stack.length-1; idx >= stackLevel; idx--) { _entry = stack[idx].entry // compute number of characters since last 'enter' _entry.length = pos - _entry.pos this._exit(_entry, stack[idx].context, stack[idx-1].context) } stack.splice(stackLevel, 1) // create new elements for all lower entries for (idx = stackLevel; idx < stack.length; idx++) { _entry = stack[idx].entry // bump 'enter' pos _entry.pos = pos stack[idx].context = this._enter(_entry, stack[idx-1].context) } } } // Finally append a trailing text node var trailingText = text.substring(pos) if (trailingText) { this._createText(rootContext, trailingText) } } } Fragmenter.SHOULD_NOT_SPLIT = 0 Fragmenter.NORMAL = 10 Fragmenter.ANY = 100 Fragmenter.ALWAYS_ON_TOP = Number.MAX_VALUE // This is a sweep algorithm wich uses a set of ENTER/EXIT entries // to manage a stack of active elements. // Whenever a new element is entered it will be appended to its parent element. // The stack is ordered by the annotation types. // // Examples: // // - simple case: // // [top] -> ENTER(idea1) -> [top, idea1] // // Creates a new 'idea' element and appends it to 'top' // // - stacked ENTER: // // [top, idea1] -> ENTER(bold1) -> [top, idea1, bold1] // // Creates a new 'bold' element and appends it to 'idea1' // // - simple EXIT: // // [top, idea1] -> EXIT(idea1) -> [top] // // Removes 'idea1' from stack. // // - reordering ENTER: // // [top, bold1] -> ENTER(idea1) -> [top, idea1, bold1] // // Inserts 'idea1' at 2nd position, creates a new 'bold1', and appends itself to 'top' // // - reordering EXIT // // [top, idea1, bold1] -> EXIT(idea1)) -> [top, bold1] // // Removes 'idea1' from stack and creates a new 'bold1' // function _extractEntries(annotations) { var openers = [] var closers = [] forEach(annotations, function(a) { var isAnchor = (a.isAnchor ? a.isAnchor() : false) // special treatment for zero-width annos such as ContainerAnnotation.Anchors if (isAnchor) { openers.push({ mode: ANCHOR, pos: a.offset, id: a.id, level: Fragmenter.ALWAYS_ON_TOP, type: 'anchor', node: a, counter: -1, length: 0 }) } else { // TODO better naming, `Node.level` does not say enough // Better would be `Node.fragmentation = Fragmenter.SHOULD_NOT_SPLIT;` // meaning, that the fragmenter should try to render the fragment as one single // element, and not splitting it up on different stack levels. // E.g. When bold an link are overlapping // the fragmenter should not split the link element such as A<b>B<a>CD</a></b><a>EF</a>GH // but should instead A<b>B</b><a><b>CD</b><a>EF</a>GH // use a weak default level when not given var l = Fragmenter.NORMAL var isInline = (a.isInline ? a.isInline() : false) if (isInline) { l = Number.MAX_VALUE } else if (a.constructor.hasOwnProperty('fragmentation')) { l = a.constructor.fragmentation } else if (a.hasOwnProperty('fragmentationHint')) { l = a.fragmentationHint } var startOffset = Math.min(a.start.offset, a.end.offset) var endOffset = Math.max(a.start.offset, a.end.offset) var opener = { pos: startOffset, mode: ENTER, level: l, id: a.id, type: a.type, node: a, length: 0, counter: -1, } openers.push(opener) closers.push({ pos: endOffset, mode: EXIT, level: l, id: a.id, type: a.type, node: a, opener: opener }) } }) // sort the openers openers.sort(_compareOpeners) // store indexes for openers for (var i = openers.length - 1; i >= 0; i--) { openers[i].idx = i } closers.sort(_compareClosers) // merge openers and closers, sorted by pos var entries = new Array(openers.length+closers.length) var idx = 0 var idx1 = 0 var idx2 = 0 var opener = openers[idx1] var closer = closers[idx2] while(opener || closer) { if (opener && closer) { // close before open if (closer.pos <= opener.pos && closer.opener !== opener) { entries[idx] = closer idx2++ } else { entries[idx] = opener idx1++ } } else if (opener) { entries[idx] = opener idx1++ } else Eif (closer) { entries[idx] = closer idx2++ } opener = openers[idx1] closer = closers[idx2] idx++ } return entries } function _compareOpeners(a, b) { if (a.pos < b.pos) return -1 Iif (a.pos > b.pos) return 1 Iif (a.mode < b.mode) return -1 if (a.mode > b.mode) return 1 Eif (a.mode === b.mode) { Iif (a.level < b.level) return -1 Iif (a.level > b.level) return 1 } return 0 } // sort in inverse order of openers function _compareClosers(a, b) { if (a.pos < b.pos) return -1 if (a.pos > b.pos) return 1 // this makes closer be sorted in inverse order of openers // to reduce stack sice // HACK: a bit trial error. When we have to collapsed annotations // at the same position then we want the closers in the same order // as the openers. if (a.pos === a.opener.pos && b.pos === b.opener.pos) { Eif (a.opener.idx < b.opener.idx) { return -1 } else { return 1 } } Iif (a.opener.idx > b.opener.idx) return -1 Eif (a.opener.idx < b.opener.idx) return 1 return 0 } export default Fragmenter |