zenparse.js
Also see Raw version, Minified version.
1 /* First, the JSON parser */
2
3 /*
4 Copyright (c) 2005 JSON.org
5 Permission is hereby granted, free of charge, to any person obtaining a copy
6 of this software and associated documentation files (the "Software"), to deal
7 in the Software without restriction, including without limitation the rights
8 to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9 copies of the Software, and to permit persons to whom the Software is
10 furnished to do so, subject to the following conditions:
11 The Software shall be used for Good, not Evil.
12 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
13 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
14 FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
15 AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
16 LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
17 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
18 SOFTWARE.
19 */
20 Array.prototype.______array = '______array';
21 var JSON = {
22 org: 'http://www.JSON.org',
23 copyright: '(c)2005 JSON.org',
24 license: 'http://www.crockford.com/JSON/license.html',
25 stringify: function (arg) {
26 var c, i, l, s = '', v;
27 switch (typeof arg) {
28 case 'object':
29 if (arg) {
30 if (arg.______array == '______array') {
31 for (i = 0; i < arg.length; ++i) {
32 v = this.stringify(arg[i]);
33 if (s) {
34 s += ',';
35 }
36 s += v;
37 }
38 return '[' + s + ']';
39 } else if (typeof arg.toString != 'undefined') {
40 for (i in arg) {
41 v = arg[i];
42 if (typeof v != 'undefined' && typeof v != 'function') {
43 v = this.stringify(v);
44 if (s) {
45 s += ',';
46 }
47 s += this.stringify(i) + ':' + v;
48 }
49 }
50 return '{' + s + '}';
51 }
52 }
53 return 'null';
54 case 'number':
55 return isFinite(arg) ? String(arg) : 'null';
56 case 'string':
57 l = arg.length;
58 s = '"';
59 for (i = 0; i < l; i += 1) {
60 c = arg.charAt(i);
61 if (c >= ' ') {
62 if (c == '\\' || c == '"') {
63 s += '\\';
64 }
65 s += c;
66 } else {
67 switch (c) {
68 case '\b':
69 s += '\\b';
70 break;
71 case '\f':
72 s += '\\f';
73 break;
74 case '\n':
75 s += '\\n';
76 break;
77 case '\r':
78 s += '\\r';
79 break;
80 case '\t':
81 s += '\\t';
82 break;
83 default:
84 c = c.charCodeAt();
85 s += '\\u00' + Math.floor(c / 16).toString(16) +
86 (c % 16).toString(16);
87 }
88 }
89 }
90 return s + '"';
91 case 'boolean':
92 return String(arg);
93 default:
94 return 'null';
95 }
96 },
97 parse: function (text) {
98 var at = 0;
99 var ch = ' ';
100 function error(m) {
101 throw {
102 name: 'JSONError',
103 message: m,
104 at: at - 1,
105 text: text
106 };
107 }
108 function next() {
109 ch = text.charAt(at);
110 at += 1;
111 return ch;
112 }
113 function white() {
114 while (ch) {
115 if (ch <= ' ') {
116 next();
117 } else if (ch == '/') {
118 switch (next()) {
119 case '/':
120 while (next() && ch != '\n' && ch != '\r') {}
121 break;
122 case '*':
123 next();
124 for (;;) {
125 if (ch) {
126 if (ch == '*') {
127 if (next() == '/') {
128 next();
129 break;
130 }
131 } else {
132 next();
133 }
134 } else {
135 error("Unterminated comment");
136 }
137 }
138 break;
139 default:
140 error("Syntax error");
141 }
142 } else {
143 break;
144 }
145 }
146 }
147 function string() {
148 var i, s = '', t, u;
149 if (ch == '"') {
150 outer: while (next()) {
151 if (ch == '"') {
152 next();
153 return s;
154 } else if (ch == '\\') {
155 switch (next()) {
156 case 'b':
157 s += '\b';
158 break;
159 case 'f':
160 s += '\f';
161 break;
162 case 'n':
163 s += '\n';
164 break;
165 case 'r':
166 s += '\r';
167 break;
168 case 't':
169 s += '\t';
170 break;
171 case 'u':
172 u = 0;
173 for (i = 0; i < 4; i += 1) {
174 t = parseInt(next(), 16);
175 if (!isFinite(t)) {
176 break outer;
177 }
178 u = u * 16 + t;
179 }
180 s += String.fromCharCode(u);
181 break;
182 default:
183 s += ch;
184 }
185 } else {
186 s += ch;
187 }
188 }
189 }
190 error("Bad string");
191 }
192 function array() {
193 var a = [];
194 if (ch == '[') {
195 next();
196 white();
197 if (ch == ']') {
198 next();
199 return a;
200 }
201 while (ch) {
202 a.push(value());
203 white();
204 if (ch == ']') {
205 next();
206 return a;
207 } else if (ch != ',') {
208 break;
209 }
210 next();
211 white();
212 }
213 }
214 error("Bad array");
215 }
216 function object() {
217 var k, o = {};
218 if (ch == '{') {
219 next();
220 white();
221 if (ch == '}') {
222 next();
223 return o;
224 }
225 while (ch) {
226 k = string();
227 white();
228 if (ch != ':') {
229 break;
230 }
231 next();
232 o[k] = value();
233 white();
234 if (ch == '}') {
235 next();
236 return o;
237 } else if (ch != ',') {
238 break;
239 }
240 next();
241 white();
242 }
243 }
244 error("Bad object");
245 }
246 function number() {
247 var n = '', v;
248 if (ch == '-') {
249 n = '-';
250 next();
251 }
252 while (ch >= '0' && ch <= '9') {
253 n += ch;
254 next();
255 }
256 if (ch == '.') {
257 n += '.';
258 while (next() && ch >= '0' && ch <= '9') {
259 n += ch;
260 }
261 }
262 if (ch == 'e' || ch == 'E') {
263 n += 'e';
264 next();
265 if (ch == '-' || ch == '+') {
266 n += ch;
267 next();
268 }
269 while (ch >= '0' && ch <= '9') {
270 n += ch;
271 next();
272 }
273 }
274 v = +n;
275 if (!isFinite(v)) {
276 ////error("Bad number");
277 } else {
278 return v;
279 }
280 }
281 function word() {
282 switch (ch) {
283 case 't':
284 if (next() == 'r' && next() == 'u' && next() == 'e') {
285 next();
286 return true;
287 }
288 break;
289 case 'f':
290 if (next() == 'a' && next() == 'l' && next() == 's' &&
291 next() == 'e') {
292 next();
293 return false;
294 }
295 break;
296 case 'n':
297 if (next() == 'u' && next() == 'l' && next() == 'l') {
298 next();
299 return null;
300 }
301 break;
302 }
303 error("Syntax error");
304 }
305 function value() {
306 white();
307 switch (ch) {
308 case '{':
309 return object();
310 case '[':
311 return array();
312 case '"':
313 return string();
314 case '-':
315 return number();
316 default:
317 return ch >= '0' && ch <= '9' ? number() : word();
318 }
319 }
320 return value();
321 }
322 };
323
324
325
326
327 /*** end JSON ***/
328
329
330
331
332
333
334
335 /* Copyright 2005 Tom W.M. */
336
337 if (!document.body) { document.body = document.getElementsByTagName("body")[0]; }
338
339 /** queue data **/
340 var queue = [
341 // student pages are handled specially
342 new QueueItem("student","http://www.mezzoblue.com/zengarden/alldesigns/student/"),
343 new QueueItem("student","http://www.mezzoblue.com/zengarden/alldesigns/student/002/")
344 ]; // array of pages to visit
345 var completed = []; // queue entries are added to this array after they are successfully (no exceptions thrown) completed
346 var failed = []; // and vice-versa
347
348
349 var output = []; // array of ZenEntry objects
350
351 var categories = {
352 "conceptual" : 12,
353 "fun" : null,
354 "minimal" : 4,
355 "official" : 9,
356 "specialeffects" : null,
357 // "student" : 2, //student designs require special processing
358 "themes" : 2,
359 "zengarden" : 2,
360 "allothers" : 3,
361 };
362
363 function QueueItem(cat,url) {
364 var self = this;
365 this.toString = function() {
366 return self.url;
367 };
368 this.url = url;
369 this.category = cat;
370 };
371
372 function createQueue() { //add some pages to the queue:
373 var baseString = "http://www.mezzoblue.com/zengarden/alldesigns/";
374 function lpad(num) {
375 num = num.toString();
376 return (num.length == 2) ? "0" + num : (num.length == 1) ? "00" + num : num;
377 };
378 for (cat in categories) {
379 queue[queue.length] = new QueueItem(cat,baseString + cat + "/");
380 if (categories[cat] != null) {
381 for (var i = 1; i < (categories[cat]+1); i++) {
382 queue[queue.length] = new QueueItem(cat,baseString + cat + "/" + lpad(i) + "/");
383 }
384 }
385 }
386 };
387
388 createQueue(); //go ahead
389
390 /*
391 var nations = { // this isn't complete--or possibly even necessary?
392 "United States" : "American",
393 "Canada" : "Canadian",
394 "Latvia" : "Latvian",
395 "Argentina" : "Argentinan",
396 "Italy" : "Italian",
397 "Poland" : "Polish",
398 "Japan" : "Japanese",
399 "France" : "French",
400 "United Kingdom" : "British",
401 "Malaysia" : "Malaysian",
402 "Australia" : "Australian",
403 "Germany" : "",
404 "Israel" : "",
405 "Croatia" : "",
406 "Singapore" : "",
407 "Belgium" : "",
408 "Czech Republic" : "",
409 "" : "",
410 "" : "",
411 "" : "",
412 }; */
413
414 function ZenEntry(category,number,name,url,nation,cssfile,title) {
415 this.number = number;
416 this.name = name;
417 this.url = url;
418 /*this.nationality = nations[nation] || null;
419 if (this.nationality == null) {
420 throw "Nationality of " + nation + " is not defined.";
421 }*/
422 this.nation = nation;
423 this.cssfile = cssfile;
424 this.title = title;
425 };
426
427 function locationContains(doc,needle) {
428 return (doc.location.href.toString().indexOf(needle) > -1);
429 };
430
431 /**** PROCESS THE PAGE ****/
432
433 function processPage(queueItem,doc) {
434 var content = doc.getElementById("mainContent");
435 if (!content) throw "#mainContent not found";
436 for (var i = 0; i<content.childNodes.length; i++) {
437 if (content.childNodes[i].className && content.childNodes[i].className.match(/\s*contentarea\s*/i)) {
438 content = content.childNodes[i];
439 break;
440 }
441 }
442 if (content == document.getElementById("mainContent")) throw "contentarea div not found";
443 lis = content.getElementsByTagName("ul")[1].getElementsByTagName("li");
444 for (var i = 0; i < lis.length; i++) {
445 lis[i].style.border = "1px dotted purple";
446 processListItem(doc,queueItem,lis[i]);
447 }
448
449 };
450
451
452 function processListItem(doc,queueItem,li) {
453 var nation = (li.getElementsByTagName("img")[0]) ? li.getElementsByTagName("img")[0].getAttribute("alt") : undefined;
454 as = li.getElementsByTagName("a");
455
456 var range = doc.createRange();
457 range.selectNodeContents(li);
458 var name = range.toString().replace(/^.*by |\s$/,"");
459
460 var cssfile = as[0].getAttribute("href").replace(/^\s*http:\/\/(www\.)?csszengarden\.com\/\?cssfile=/i,"");
461 if (cssfile.match(/^\//)) { cssfile = "http://www.csszengarden.com" + cssfile; }
462
463 var number = (li.innerHTML.replace(/^\s*<img[^\>]*>\s*/i,"").match(/[0-9]{0,2}[1-9]\s/g)) ? parseInt(li.innerHTML.replace(/^\s*<img[^\>]*>\s*([0-9]{0,2}[1-9]).*/i,"$1"),10) : undefined;
464
465 var category = queueItem.category;
466
467 var range = doc.createRange();
468 range.selectNodeContents(as[0]);
469 var title = range.toString().replace(/^\s|\s$/g,"");
470
471 var url = (as[1]) ? as[1].getAttribute("href") : undefined;
472
473 output.push(new ZenEntry(category,number,name,url,nation,cssfile,title));
474 sendInfo(JSON.stringify(output[output.length-1]));
475 };
476
477 /**** PROCESS THE QUEUE ****/
478
479 var step = 0;
480
481 function doStage() {
482 // check for a page load
483 if (!iframe.contentDocument ||
484 !iframe.contentDocument.body ||
485 !iframe.contentDocument.body.innerHTML ||
486 !iframe.contentDocument.body.innerHTML.indexOf("© Copyright 2001 – 2005, Dave Shea, all rights reserved.") < 0
487 ) {
488 setTimeout(doStage,200);
489 return;
490 } else if (iframe.contentDocument.body.innerHTML.indexOf("error 404 - file not found") > -1) {
491 alert("404 error accessing\n" + queue[0].toString() + "\n\nMoving on to next item.");
492 failed.push(queue.shift());
493 iframe.setLocation(queue[0].toString());
494 setTimeout(doStage,500);
495 }{
496 var success = true;
497 try {
498 processPage(queue[0],iframe.contentDocument);
499 } catch (e) {
500 processException(e);
501 success = false;
502 }
503 // advance the queue
504 var item = queue.shift();
505 if (success) {
506 completed.push(item);
507 } else {
508 failed.push(item);
509 }
510 if (queue[0]) {
511 iframe.contentDocument.body.innerHTML = "";
512 step++;
513 //if (step > 6) return; // limits to a certain number of pages
514 iframe.setLocation(queue[0].toString());
515 setTimeout(doStage,500);
516 }
517 }
518 };
519
520
521 /**** HANDLE ERRORS ****/
522
523 function processException(e) {
524 sendInfo(e);
525 };
526
527
528 function sendInfo(info) {
529 data.value += "\n\n" + info.toString();
530 data.selectionEnd = data.length-1;
531 };
532
533
534 function setInfo(string) {
535 data.value = string;
536 };
537
538 /**** SET IT ALL UP ****/
539
540 document.body.innerHTML = "";
541
542 var iframe = document.createElement("iframe");
543 iframe.setAttribute("style","position:absolute;right: 10px;height:300px;left:10px;top:10px;");
544 document.body.appendChild(iframe);
545 iframe.setLocation = function(newLocation) {
546 this.setAttribute("src",newLocation);
547 };
548
549 var data = document.createElement("textarea");
550 //data.setAttribute("type","text");
551 data.setAttribute("name","data");
552 data.style.position = "absolute";
553 data.style.left = "10px";
554 data.style.right = "10px";
555 data.style.top = "320px";
556 data.style.height = "200px";
557 document.body.appendChild(data);
558
559 /**** BEGIN ****/
560 iframe.setLocation(queue[0]);
561 setInfo(JSON.stringify(queue).replace(/,\{"url"/g,',\n{"url"'));
562 setTimeout(function() { setInfo(""); doStage(); },3000);
563
564 /* bookmarklet:
565
566 javascript:var script = document.createElement("script"); script.setAttribute("type","text/javascript"); script.setAttribute("src","http://freecog.net/zenparse.js"); document.getElementsByTagName("head")[0].appendChild(script); void(null);
567
568 */
© 2004–2010 Tom W. Most