Guido van Rossum | a219efa | 1997-11-18 15:09:54 +0000 | [diff] [blame] | 1 | # A parser for XML, using the derived class as static DTD. |
Guido van Rossum | 5d68e8e | 1997-11-18 15:27:20 +0000 | [diff] [blame] | 2 | # Author: Sjoerd Mullender. |
Guido van Rossum | a219efa | 1997-11-18 15:09:54 +0000 | [diff] [blame] | 3 | |
| 4 | import re |
| 5 | import string |
| 6 | |
| 7 | |
| 8 | # Regular expressions used for parsing |
| 9 | |
| 10 | _S = '[ \t\r\n]+' |
| 11 | _opS = '[ \t\r\n]*' |
| 12 | _Name = '[a-zA-Z_:][-a-zA-Z0-9._:]*' |
| 13 | interesting = re.compile('[&<]') |
| 14 | incomplete = re.compile('&(' + _Name + '|#[0-9]*|#x[0-9a-fA-F]*)?|' |
| 15 | '<([a-zA-Z_:][^<>]*|' |
| 16 | '/([a-zA-Z_:][^<>]*)?|' |
| 17 | '![^<>]*|' |
| 18 | '\?[^<>]*)?') |
| 19 | |
| 20 | ref = re.compile('&(' + _Name + '|#[0-9]+|#x[0-9a-fA-F]+);?') |
| 21 | entityref = re.compile('&(?P<name>' + _Name + ')[^-a-zA-Z0-9._:]') |
| 22 | charref = re.compile('&#(?P<char>[0-9]+[^0-9]|x[0-9a-fA-F]+[^0-9a-fA-F])') |
| 23 | space = re.compile(_S) |
| 24 | newline = re.compile('\n') |
| 25 | |
| 26 | starttagopen = re.compile('<' + _Name) |
| 27 | endtagopen = re.compile('</') |
| 28 | starttagend = re.compile(_opS + '(?P<slash>/?)>') |
| 29 | endbracket = re.compile('>') |
| 30 | tagfind = re.compile(_Name) |
| 31 | cdataopen = re.compile('<!\[CDATA\[') |
| 32 | cdataclose = re.compile('\]\]>') |
| 33 | special = re.compile('<!(?P<special>[^<>]*)>') |
| 34 | procopen = re.compile('<\?(?P<proc>' + _Name + ')' + _S) |
| 35 | procclose = re.compile('\?>') |
| 36 | commentopen = re.compile('<!--') |
| 37 | commentclose = re.compile('-->') |
| 38 | doubledash = re.compile('--') |
| 39 | attrfind = re.compile( |
| 40 | _S + '(?P<name>' + _Name + ')' |
| 41 | '(' + _opS + '=' + _opS + |
| 42 | '(?P<value>\'[^\']*\'|"[^"]*"|[-a-zA-Z0-9.:+*%?!()_#=~]+))') |
| 43 | |
| 44 | |
| 45 | # XML parser base class -- find tags and call handler functions. |
| 46 | # Usage: p = XMLParser(); p.feed(data); ...; p.close(). |
Guido van Rossum | 5d68e8e | 1997-11-18 15:27:20 +0000 | [diff] [blame] | 47 | # The dtd is defined by deriving a class which defines methods with |
| 48 | # special names to handle tags: start_foo and end_foo to handle <foo> |
| 49 | # and </foo>, respectively. The data between tags is passed to the |
| 50 | # parser by calling self.handle_data() with some data as argument (the |
| 51 | # data may be split up in arbutrary chunks). Entity references are |
| 52 | # passed by calling self.handle_entityref() with the entity reference |
| 53 | # as argument. |
Guido van Rossum | a219efa | 1997-11-18 15:09:54 +0000 | [diff] [blame] | 54 | |
| 55 | class XMLParser: |
| 56 | |
| 57 | # Interface -- initialize and reset this instance |
| 58 | def __init__(self, verbose=0): |
| 59 | self.verbose = verbose |
| 60 | self.reset() |
| 61 | |
| 62 | # Interface -- reset this instance. Loses all unprocessed data |
| 63 | def reset(self): |
| 64 | self.rawdata = '' |
| 65 | self.stack = [] |
| 66 | self.lasttag = '???' |
| 67 | self.nomoretags = 0 |
| 68 | self.literal = 0 |
| 69 | self.lineno = 1 |
| 70 | |
| 71 | # For derived classes only -- enter literal mode (CDATA) till EOF |
| 72 | def setnomoretags(self): |
| 73 | self.nomoretags = self.literal = 1 |
| 74 | |
| 75 | # For derived classes only -- enter literal mode (CDATA) |
| 76 | def setliteral(self, *args): |
| 77 | self.literal = 1 |
| 78 | |
| 79 | # Interface -- feed some data to the parser. Call this as |
| 80 | # often as you want, with as little or as much text as you |
| 81 | # want (may include '\n'). (This just saves the text, all the |
| 82 | # processing is done by goahead().) |
| 83 | def feed(self, data): |
| 84 | self.rawdata = self.rawdata + data |
| 85 | self.goahead(0) |
| 86 | |
| 87 | # Interface -- handle the remaining data |
| 88 | def close(self): |
| 89 | self.goahead(1) |
| 90 | |
| 91 | # Interface -- translate references |
| 92 | def translate_references(self, data): |
| 93 | newdata = [] |
| 94 | i = 0 |
| 95 | while 1: |
| 96 | res = ref.search(data, i) |
| 97 | if res is None: |
| 98 | newdata.append(data[i:]) |
| 99 | return string.join(newdata, '') |
| 100 | if data[res.end(0) - 1] != ';': |
| 101 | self.syntax_error(self.lineno, |
| 102 | '; missing after entity/char reference') |
| 103 | newdata.append(data[i:res.start(0)]) |
| 104 | str = res.group(1) |
| 105 | if str[0] == '#': |
| 106 | if str[1] == 'x': |
| 107 | newdata.append(chr(string.atoi(str[2:], 16))) |
| 108 | else: |
| 109 | newdata.append(chr(string.atoi(str[1:]))) |
| 110 | else: |
| 111 | try: |
| 112 | newdata.append(self.entitydefs[str]) |
| 113 | except KeyError: |
| 114 | # can't do it, so keep the entity ref in |
| 115 | newdata.append('&' + str + ';') |
| 116 | i = res.end(0) |
| 117 | |
| 118 | # Internal -- handle data as far as reasonable. May leave state |
| 119 | # and data to be processed by a subsequent call. If 'end' is |
| 120 | # true, force handling all data as if followed by EOF marker. |
| 121 | def goahead(self, end): |
| 122 | rawdata = self.rawdata |
| 123 | i = 0 |
| 124 | n = len(rawdata) |
| 125 | while i < n: |
| 126 | if self.nomoretags: |
| 127 | data = rawdata[i:n] |
| 128 | self.handle_data(data) |
| 129 | self.lineno = self.lineno + string.count(data, '\n') |
| 130 | i = n |
| 131 | break |
| 132 | res = interesting.search(rawdata, i) |
| 133 | if res: |
| 134 | j = res.start(0) |
| 135 | else: |
| 136 | j = n |
| 137 | if i < j: |
| 138 | data = rawdata[i:j] |
| 139 | self.handle_data(data) |
| 140 | self.lineno = self.lineno + string.count(data, '\n') |
| 141 | i = j |
| 142 | if i == n: break |
| 143 | if rawdata[i] == '<': |
| 144 | if starttagopen.match(rawdata, i): |
| 145 | if self.literal: |
| 146 | data = rawdata[i] |
| 147 | self.handle_data(data) |
| 148 | self.lineno = self.lineno + string.count(data, '\n') |
| 149 | i = i+1 |
| 150 | continue |
| 151 | k = self.parse_starttag(i) |
| 152 | if k < 0: break |
| 153 | self.lineno = self.lineno + string.count(rawdata[i:k], '\n') |
| 154 | i = k |
| 155 | continue |
| 156 | if endtagopen.match(rawdata, i): |
| 157 | k = self.parse_endtag(i) |
| 158 | if k < 0: break |
| 159 | self.lineno = self.lineno + string.count(rawdata[i:k], '\n') |
| 160 | i = k |
| 161 | self.literal = 0 |
| 162 | continue |
| 163 | if commentopen.match(rawdata, i): |
| 164 | if self.literal: |
| 165 | data = rawdata[i] |
| 166 | self.handle_data(data) |
| 167 | self.lineno = self.lineno + string.count(data, '\n') |
| 168 | i = i+1 |
| 169 | continue |
| 170 | k = self.parse_comment(i) |
| 171 | if k < 0: break |
| 172 | self.lineno = self.lineno + string.count(rawdata[i:k], '\n') |
| 173 | i = k |
| 174 | continue |
| 175 | if cdataopen.match(rawdata, i): |
| 176 | k = self.parse_cdata(i) |
| 177 | if k < 0: break |
| 178 | self.lineno = self.lineno + string.count(rawdata[i:i], '\n') |
| 179 | i = k |
| 180 | continue |
| 181 | res = procopen.match(rawdata, i) |
| 182 | if res: |
| 183 | k = self.parse_proc(i, res) |
| 184 | if k < 0: break |
| 185 | self.lineno = self.lineno + string.count(rawdata[i:k], '\n') |
| 186 | i = k |
| 187 | continue |
| 188 | res = special.match(rawdata, i) |
| 189 | if res: |
| 190 | if self.literal: |
| 191 | data = rawdata[i] |
| 192 | self.handle_data(data) |
| 193 | self.lineno = self.lineno + string.count(data, '\n') |
| 194 | i = i+1 |
| 195 | continue |
| 196 | self.handle_special(res.group('special')) |
| 197 | self.lineno = self.lineno + string.count(res.group(0), '\n') |
| 198 | i = res.end(0) |
| 199 | continue |
| 200 | elif rawdata[i] == '&': |
| 201 | res = charref.match(rawdata, i) |
| 202 | if res is not None: |
| 203 | i = res.end(0) |
| 204 | if rawdata[i-1] != ';': |
| 205 | self.syntax_error(self.lineno, '; missing in charref') |
| 206 | i = i-1 |
| 207 | self.handle_charref(res.group('char')[:-1]) |
| 208 | self.lineno = self.lineno + string.count(res.group(0), '\n') |
| 209 | continue |
| 210 | res = entityref.match(rawdata, i) |
| 211 | if res is not None: |
| 212 | i = res.end(0) |
| 213 | if rawdata[i-1] != ';': |
| 214 | self.syntax_error(self.lineno, '; missing in entityref') |
| 215 | i = i-1 |
| 216 | self.handle_entityref(res.group('name')) |
| 217 | self.lineno = self.lineno + string.count(res.group(0), '\n') |
| 218 | continue |
| 219 | else: |
| 220 | raise RuntimeError, 'neither < nor & ??' |
| 221 | # We get here only if incomplete matches but |
| 222 | # nothing else |
| 223 | res = incomplete.match(rawdata, i) |
| 224 | if not res: |
| 225 | data = rawdata[i] |
| 226 | self.handle_data(data) |
| 227 | self.lineno = self.lineno + string.count(data, '\n') |
| 228 | i = i+1 |
| 229 | continue |
| 230 | j = res.end(0) |
| 231 | if j == n: |
| 232 | break # Really incomplete |
| 233 | self.syntax_error(self.lineno, 'bogus < or &') |
| 234 | data = res.group(0) |
| 235 | self.handle_data(data) |
| 236 | self.lineno = self.lineno + string.count(data, '\n') |
| 237 | i = j |
| 238 | # end while |
| 239 | if end and i < n: |
| 240 | data = rawdata[i:n] |
| 241 | self.handle_data(data) |
| 242 | self.lineno = self.lineno + string.count(data, '\n') |
| 243 | i = n |
| 244 | self.rawdata = rawdata[i:] |
| 245 | # XXX if end: check for empty stack |
| 246 | |
| 247 | # Internal -- parse comment, return length or -1 if not terminated |
| 248 | def parse_comment(self, i): |
| 249 | rawdata = self.rawdata |
| 250 | if rawdata[i:i+4] <> '<!--': |
| 251 | raise RuntimeError, 'unexpected call to handle_comment' |
| 252 | res = commentclose.search(rawdata, i+4) |
| 253 | if not res: |
| 254 | return -1 |
| 255 | # doubledash search will succeed because it's a subset of commentclose |
| 256 | if doubledash.search(rawdata, i+4).start(0) < res.start(0): |
| 257 | self.syntax_error(self.lineno, "`--' inside comment") |
| 258 | self.handle_comment(rawdata[i+4: res.start(0)]) |
| 259 | return res.end(0) |
| 260 | |
| 261 | # Internal -- handle CDATA tag, return lenth or -1 if not terminated |
| 262 | def parse_cdata(self, i): |
| 263 | rawdata = self.rawdata |
| 264 | if rawdata[i:i+9] <> '<![CDATA[': |
| 265 | raise RuntimeError, 'unexpected call to handle_cdata' |
| 266 | res = cdataclose.search(rawdata, i+9) |
| 267 | if not res: |
| 268 | return -1 |
| 269 | self.handle_cdata(rawdata[i+9:res.start(0)]) |
| 270 | return res.end(0) |
| 271 | |
| 272 | def parse_proc(self, i, res): |
| 273 | rawdata = self.rawdata |
| 274 | if not res: |
| 275 | raise RuntimeError, 'unexpected call to parse_proc' |
| 276 | name = res.group('proc') |
| 277 | res = procclose.search(rawdata, res.end(0)) |
| 278 | if not res: |
| 279 | return -1 |
| 280 | self.handle_proc(name, rawdata[res.pos:res.start(0)]) |
| 281 | return res.end(0) |
| 282 | |
| 283 | # Internal -- handle starttag, return length or -1 if not terminated |
| 284 | def parse_starttag(self, i): |
| 285 | rawdata = self.rawdata |
| 286 | # i points to start of tag |
| 287 | end = endbracket.search(rawdata, i+1) |
| 288 | if not end: |
| 289 | return -1 |
| 290 | j = end.start(0) |
| 291 | # Now parse the data between i+1 and j into a tag and attrs |
| 292 | attrdict = {} |
| 293 | res = tagfind.match(rawdata, i+1) |
| 294 | if not res: |
| 295 | raise RuntimeError, 'unexpected call to parse_starttag' |
| 296 | k = res.end(0) |
| 297 | tag = res.group(0) |
| 298 | if hasattr(self, tag + '_attributes'): |
| 299 | attrlist = getattr(self, tag + '_attributes') |
| 300 | else: |
| 301 | attrlist = None |
| 302 | self.lasttag = tag |
| 303 | while k < j: |
| 304 | res = attrfind.match(rawdata, k) |
| 305 | if not res: break |
| 306 | attrname, attrvalue = res.group('name', 'value') |
| 307 | if attrvalue is None: |
| 308 | self.syntax_error(self.lineno, 'no attribute value specified') |
| 309 | attrvalue = attrname |
| 310 | elif attrvalue[:1] == "'" == attrvalue[-1:] or \ |
| 311 | attrvalue[:1] == '"' == attrvalue[-1:]: |
| 312 | attrvalue = attrvalue[1:-1] |
| 313 | else: |
| 314 | self.syntax_error(self.lineno, 'attribute value not quoted') |
Guido van Rossum | a219efa | 1997-11-18 15:09:54 +0000 | [diff] [blame] | 315 | if attrlist is not None and attrname not in attrlist: |
| 316 | self.syntax_error(self.lineno, |
| 317 | 'unknown attribute %s of element %s' % |
| 318 | (attrname, tag)) |
| 319 | if attrdict.has_key(attrname): |
| 320 | self.syntax_error(self.lineno, 'attribute specified twice') |
| 321 | attrdict[attrname] = self.translate_references(attrvalue) |
| 322 | k = res.end(0) |
| 323 | res = starttagend.match(rawdata, k) |
| 324 | if not res: |
| 325 | self.syntax_error(self.lineno, 'garbage in start tag') |
| 326 | self.finish_starttag(tag, attrdict) |
| 327 | if res and res.group('slash') == '/': |
| 328 | self.finish_endtag(tag) |
| 329 | return end.end(0) |
| 330 | |
| 331 | # Internal -- parse endtag |
| 332 | def parse_endtag(self, i): |
| 333 | rawdata = self.rawdata |
| 334 | end = endbracket.search(rawdata, i+1) |
| 335 | if not end: |
| 336 | return -1 |
| 337 | res = tagfind.match(rawdata, i+2) |
| 338 | if not res: |
| 339 | self.syntax_error(self.lineno, 'no name specified in end tag') |
| 340 | tag = '' |
| 341 | k = i+2 |
| 342 | else: |
| 343 | tag = res.group(0) |
| 344 | k = res.end(0) |
| 345 | if k != end.start(0): |
| 346 | # check that there is only white space at end of tag |
| 347 | res = space.match(rawdata, k) |
| 348 | if res is None or res.end(0) != end.start(0): |
| 349 | self.syntax_error(self.lineno, 'garbage in end tag') |
| 350 | self.finish_endtag(tag) |
| 351 | return end.end(0) |
| 352 | |
| 353 | # Internal -- finish processing of start tag |
| 354 | # Return -1 for unknown tag, 1 for balanced tag |
| 355 | def finish_starttag(self, tag, attrs): |
| 356 | self.stack.append(tag) |
| 357 | try: |
| 358 | method = getattr(self, 'start_' + tag) |
| 359 | except AttributeError: |
| 360 | self.unknown_starttag(tag, attrs) |
| 361 | return -1 |
| 362 | else: |
| 363 | self.handle_starttag(tag, method, attrs) |
| 364 | return 1 |
| 365 | |
| 366 | # Internal -- finish processing of end tag |
| 367 | def finish_endtag(self, tag): |
| 368 | if not tag: |
| 369 | found = len(self.stack) - 1 |
| 370 | if found < 0: |
| 371 | self.unknown_endtag(tag) |
| 372 | return |
| 373 | else: |
| 374 | if tag not in self.stack: |
| 375 | try: |
| 376 | method = getattr(self, 'end_' + tag) |
| 377 | except AttributeError: |
| 378 | self.unknown_endtag(tag) |
| 379 | return |
| 380 | found = len(self.stack) |
| 381 | for i in range(found): |
| 382 | if self.stack[i] == tag: found = i |
| 383 | while len(self.stack) > found: |
| 384 | tag = self.stack[-1] |
| 385 | try: |
| 386 | method = getattr(self, 'end_' + tag) |
| 387 | except AttributeError: |
| 388 | method = None |
| 389 | if method: |
| 390 | self.handle_endtag(tag, method) |
| 391 | else: |
| 392 | self.unknown_endtag(tag) |
| 393 | del self.stack[-1] |
| 394 | |
| 395 | # Overridable -- handle start tag |
| 396 | def handle_starttag(self, tag, method, attrs): |
| 397 | method(attrs) |
| 398 | |
| 399 | # Overridable -- handle end tag |
| 400 | def handle_endtag(self, tag, method): |
| 401 | method() |
| 402 | |
| 403 | # Example -- handle character reference, no need to override |
| 404 | def handle_charref(self, name): |
| 405 | try: |
| 406 | if name[0] == 'x': |
| 407 | n = string.atoi(name[1:], 16) |
| 408 | else: |
| 409 | n = string.atoi(name) |
| 410 | except string.atoi_error: |
| 411 | self.unknown_charref(name) |
| 412 | return |
| 413 | if not 0 <= n <= 255: |
| 414 | self.unknown_charref(name) |
| 415 | return |
| 416 | self.handle_data(chr(n)) |
| 417 | |
| 418 | # Definition of entities -- derived classes may override |
| 419 | entitydefs = \ |
| 420 | {'lt': '<', 'gt': '>', 'amp': '&', 'quot': '"', 'apos': '\''} |
| 421 | |
| 422 | # Example -- handle entity reference, no need to override |
| 423 | def handle_entityref(self, name): |
| 424 | table = self.entitydefs |
| 425 | if table.has_key(name): |
| 426 | self.handle_data(table[name]) |
| 427 | else: |
| 428 | self.unknown_entityref(name) |
| 429 | return |
| 430 | |
| 431 | # Example -- handle data, should be overridden |
| 432 | def handle_data(self, data): |
| 433 | pass |
| 434 | |
| 435 | # Example -- handle cdata, could be overridden |
| 436 | def handle_cdata(self, data): |
| 437 | pass |
| 438 | |
| 439 | # Example -- handle comment, could be overridden |
| 440 | def handle_comment(self, data): |
| 441 | pass |
| 442 | |
| 443 | # Example -- handle processing instructions, could be overridden |
| 444 | def handle_proc(self, name, data): |
| 445 | pass |
| 446 | |
| 447 | # Example -- handle special instructions, could be overridden |
| 448 | def handle_special(self, data): |
| 449 | pass |
| 450 | |
| 451 | # Example -- handle relatively harmless syntax errors, could be overridden |
| 452 | def syntax_error(self, lineno, message): |
| 453 | raise RuntimeError, 'Syntax error at line %d: %s' % (lineno, message) |
| 454 | |
| 455 | # To be overridden -- handlers for unknown objects |
| 456 | def unknown_starttag(self, tag, attrs): pass |
| 457 | def unknown_endtag(self, tag): pass |
| 458 | def unknown_charref(self, ref): pass |
| 459 | def unknown_entityref(self, ref): pass |
| 460 | |
| 461 | |
| 462 | class TestXMLParser(XMLParser): |
| 463 | |
| 464 | def __init__(self, verbose=0): |
| 465 | self.testdata = "" |
| 466 | XMLParser.__init__(self, verbose) |
| 467 | |
| 468 | def handle_data(self, data): |
| 469 | self.testdata = self.testdata + data |
| 470 | if len(`self.testdata`) >= 70: |
| 471 | self.flush() |
| 472 | |
| 473 | def flush(self): |
| 474 | data = self.testdata |
| 475 | if data: |
| 476 | self.testdata = "" |
| 477 | print 'data:', `data` |
| 478 | |
| 479 | def handle_cdata(self, data): |
| 480 | self.flush() |
| 481 | print 'cdata:', `data` |
| 482 | |
| 483 | def handle_proc(self, name, data): |
| 484 | self.flush() |
| 485 | print 'processing:',name,`data` |
| 486 | |
| 487 | def handle_special(self, data): |
| 488 | self.flush() |
| 489 | print 'special:',`data` |
| 490 | |
| 491 | def handle_comment(self, data): |
| 492 | self.flush() |
| 493 | r = `data` |
| 494 | if len(r) > 68: |
| 495 | r = r[:32] + '...' + r[-32:] |
| 496 | print 'comment:', r |
| 497 | |
| 498 | def syntax_error(self, lineno, message): |
| 499 | print 'error at line %d:' % lineno, message |
| 500 | |
| 501 | def unknown_starttag(self, tag, attrs): |
| 502 | self.flush() |
| 503 | if not attrs: |
| 504 | print 'start tag: <' + tag + '>' |
| 505 | else: |
| 506 | print 'start tag: <' + tag, |
| 507 | for name, value in attrs: |
| 508 | print name + '=' + '"' + value + '"', |
| 509 | print '>' |
| 510 | |
| 511 | def unknown_endtag(self, tag): |
| 512 | self.flush() |
| 513 | print 'end tag: </' + tag + '>' |
| 514 | |
| 515 | def unknown_entityref(self, ref): |
| 516 | self.flush() |
| 517 | print '*** unknown entity ref: &' + ref + ';' |
| 518 | |
| 519 | def unknown_charref(self, ref): |
| 520 | self.flush() |
| 521 | print '*** unknown char ref: &#' + ref + ';' |
| 522 | |
| 523 | def close(self): |
| 524 | XMLParser.close(self) |
| 525 | self.flush() |
| 526 | |
| 527 | def test(args = None): |
| 528 | import sys |
| 529 | |
| 530 | if not args: |
| 531 | args = sys.argv[1:] |
| 532 | |
| 533 | if args and args[0] == '-s': |
| 534 | args = args[1:] |
| 535 | klass = XMLParser |
| 536 | else: |
| 537 | klass = TestXMLParser |
| 538 | |
| 539 | if args: |
| 540 | file = args[0] |
| 541 | else: |
| 542 | file = 'test.xml' |
| 543 | |
| 544 | if file == '-': |
| 545 | f = sys.stdin |
| 546 | else: |
| 547 | try: |
| 548 | f = open(file, 'r') |
| 549 | except IOError, msg: |
| 550 | print file, ":", msg |
| 551 | sys.exit(1) |
| 552 | |
| 553 | data = f.read() |
| 554 | if f is not sys.stdin: |
| 555 | f.close() |
| 556 | |
| 557 | x = klass() |
| 558 | for c in data: |
| 559 | x.feed(c) |
| 560 | x.close() |
| 561 | |
| 562 | |
| 563 | if __name__ == '__main__': |
| 564 | test() |