Daniel Veillard | 3371ff8 | 2002-10-01 13:37:48 +0000 | [diff] [blame^] | 1 | #!/usr/bin/python -u |
| 2 | # |
| 3 | # imports the API description and fills up a database with |
| 4 | # name relevance to modules, functions or web pages |
| 5 | # |
| 6 | import MySQLdb |
| 7 | import libxml2 |
| 8 | import sys |
| 9 | import string |
| 10 | import os |
| 11 | |
| 12 | # |
| 13 | # The dictionnary of tables required and the SQL command needed |
| 14 | # to create them |
| 15 | # |
| 16 | TABLES={ |
| 17 | "symbols" : """CREATE TABLE symbols ( |
| 18 | name varchar(255) NOT NULL, |
| 19 | module varchar(255) NOT NULL, |
| 20 | type varchar(25) NOT NULL, |
| 21 | descr varchar(255), |
| 22 | UNIQUE KEY name (name), |
| 23 | KEY module (module))""", |
| 24 | "words" : """CREATE TABLE words ( |
| 25 | name varchar(50) NOT NULL, |
| 26 | symbol varchar(255) NOT NULL, |
| 27 | relevance int, |
| 28 | KEY name (name), |
| 29 | KEY symbol (symbol), |
| 30 | UNIQUE KEY ID (name, symbol))""", |
| 31 | } |
| 32 | |
| 33 | # |
| 34 | # The XML API description file to parse |
| 35 | # |
| 36 | API="libxml2-api.xml" |
| 37 | DB=None |
| 38 | |
| 39 | ######################################################################### |
| 40 | # # |
| 41 | # MySQL database interfaces # |
| 42 | # # |
| 43 | ######################################################################### |
| 44 | def createTable(db, name): |
| 45 | global TABLES |
| 46 | |
| 47 | if db == None: |
| 48 | return -1 |
| 49 | if name == None: |
| 50 | return -1 |
| 51 | c = db.cursor() |
| 52 | |
| 53 | ret = c.execute("DROP TABLE IF EXISTS %s" % (name)) |
| 54 | if ret == 1: |
| 55 | print "Removed table %s" % (name) |
| 56 | print "Creating table %s" % (name) |
| 57 | try: |
| 58 | ret = c.execute(TABLES[name]) |
| 59 | except: |
| 60 | print "Failed to create table %s" % (name) |
| 61 | return -1 |
| 62 | return ret |
| 63 | |
| 64 | def checkTables(db): |
| 65 | global TABLES |
| 66 | |
| 67 | if db == None: |
| 68 | return -1 |
| 69 | c = db.cursor() |
| 70 | nbtables = c.execute("show tables") |
| 71 | print "Found %d tables" % (nbtables) |
| 72 | tables = {} |
| 73 | i = 0 |
| 74 | while i < nbtables: |
| 75 | l = c.fetchone() |
| 76 | name = l[0] |
| 77 | tables[name] = {} |
| 78 | i = i + 1 |
| 79 | |
| 80 | for table in TABLES.keys(): |
| 81 | if not tables.has_key(table): |
| 82 | print "table %s missing" % (table) |
| 83 | createTable(db, table) |
| 84 | print "checkTables finished" |
| 85 | |
| 86 | # make sure apache can access the tables read-only |
| 87 | try: |
| 88 | ret = c.execute("GRANT SELECT ON xmlsoft.* TO nobody@localhost") |
| 89 | except: |
| 90 | pass |
| 91 | return 0 |
| 92 | |
| 93 | def openMySQL(db="xmlsoft", passwd=None): |
| 94 | global DB |
| 95 | |
| 96 | if passwd == None: |
| 97 | passwd = "ducon" |
| 98 | DB = MySQLdb.connect(passwd=passwd, db=db) |
| 99 | if DB == None: |
| 100 | return -1 |
| 101 | ret = checkTables(DB) |
| 102 | return ret |
| 103 | |
| 104 | def updateWord(name, symbol, relevance): |
| 105 | global DB |
| 106 | |
| 107 | if DB == None: |
| 108 | openMySQL() |
| 109 | if DB == None: |
| 110 | return -1 |
| 111 | if name == None: |
| 112 | return -1 |
| 113 | if symbol == None: |
| 114 | return -1 |
| 115 | |
| 116 | c = DB.cursor() |
| 117 | try: |
| 118 | ret = c.execute( |
| 119 | """INSERT INTO words (name, symbol, relevance) VALUES ('%s','%s', %d)""" % |
| 120 | (name, symbol, relevance)) |
| 121 | except: |
| 122 | try: |
| 123 | ret = c.execute( |
| 124 | """UPDATE words SET relevance = %d where name = '%s' and symbol = '%s'""" % |
| 125 | (relevance, name, symbol)) |
| 126 | except: |
| 127 | print "Update word (%s, %s, %s) failed command" % (name, symbol, relevance) |
| 128 | print "UPDATE words SET relevance = %d where name = '%s' and symbol = '%s'" % (relevance, name, symbol) |
| 129 | print sys.exc_type, sys.exc_value |
| 130 | return -1 |
| 131 | |
| 132 | return ret |
| 133 | |
| 134 | def updateSymbol(name, module, type, desc): |
| 135 | global DB |
| 136 | |
| 137 | updateWord(name, name, 50) |
| 138 | if DB == None: |
| 139 | openMySQL() |
| 140 | if DB == None: |
| 141 | return -1 |
| 142 | if name == None: |
| 143 | return -1 |
| 144 | if module == None: |
| 145 | return -1 |
| 146 | if type == None: |
| 147 | return -1 |
| 148 | |
| 149 | try: |
| 150 | desc = string.replace(desc, "'", " ") |
| 151 | l = string.split(desc, ".") |
| 152 | desc = l[0] |
| 153 | desc = desc[0:99] |
| 154 | except: |
| 155 | desc = "" |
| 156 | |
| 157 | c = DB.cursor() |
| 158 | try: |
| 159 | ret = c.execute( |
| 160 | """INSERT INTO symbols (name, module, type, descr) VALUES ('%s','%s', '%s', '%s')""" % |
| 161 | (name, module, type, desc)) |
| 162 | except: |
| 163 | try: |
| 164 | ret = c.execute( |
| 165 | """UPDATE symbols SET module='%s', type='%s', descr='%s' where name='%s'""" % |
| 166 | (module, type, desc, name)) |
| 167 | except: |
| 168 | print "Update symbol (%s, %s, %s) failed command" % (name, module, type) |
| 169 | print """UPDATE symbols SET module='%s', type='%s', descr='%s' where name='%s'""" % (module, type, desc, name) |
| 170 | print sys.exc_type, sys.exc_value |
| 171 | return -1 |
| 172 | |
| 173 | return ret |
| 174 | |
| 175 | def addFunction(name, module, desc = ""): |
| 176 | return updateSymbol(name, module, 'function', desc) |
| 177 | |
| 178 | def addMacro(name, module, desc = ""): |
| 179 | return updateSymbol(name, module, 'macro', desc) |
| 180 | |
| 181 | def addEnum(name, module, desc = ""): |
| 182 | return updateSymbol(name, module, 'enum', desc) |
| 183 | |
| 184 | def addStruct(name, module, desc = ""): |
| 185 | return updateSymbol(name, module, 'struct', desc) |
| 186 | |
| 187 | def addConst(name, module, desc = ""): |
| 188 | return updateSymbol(name, module, 'const', desc) |
| 189 | |
| 190 | def addType(name, module, desc = ""): |
| 191 | return updateSymbol(name, module, 'type', desc) |
| 192 | |
| 193 | def addFunctype(name, module, desc = ""): |
| 194 | return updateSymbol(name, module, 'functype', desc) |
| 195 | |
| 196 | ######################################################################### |
| 197 | # # |
| 198 | # Word dictionnary and analysis routines # |
| 199 | # # |
| 200 | ######################################################################### |
| 201 | |
| 202 | wordsDict = {} |
| 203 | |
| 204 | def splitIdentifier(str): |
| 205 | ret = [] |
| 206 | while str != "": |
| 207 | cur = string.lower(str[0]) |
| 208 | str = str[1:] |
| 209 | if ((cur < 'a') or (cur > 'z')): |
| 210 | continue |
| 211 | while (str != "") and (str[0] >= 'A') and (str[0] <= 'Z'): |
| 212 | cur = cur + string.lower(str[0]) |
| 213 | str = str[1:] |
| 214 | while (str != "") and (str[0] >= 'a') and (str[0] <= 'z'): |
| 215 | cur = cur + str[0] |
| 216 | str = str[1:] |
| 217 | while (str != "") and (str[0] >= '0') and (str[0] <= '9'): |
| 218 | str = str[1:] |
| 219 | ret.append(cur) |
| 220 | return ret |
| 221 | |
| 222 | def addWord(word, module, symbol, relevance): |
| 223 | global wordsDict |
| 224 | |
| 225 | if word == None or len(word) < 3: |
| 226 | return -1 |
| 227 | if module == None or symbol == None: |
| 228 | return -1 |
| 229 | if wordsDict.has_key(word): |
| 230 | d = wordsDict[word] |
| 231 | if d == None: |
| 232 | return 0 |
| 233 | if len(d) > 500: |
| 234 | wordsDict[word] = None |
| 235 | return 0 |
| 236 | try: |
| 237 | relevance = relevance + d[(module, symbol)] |
| 238 | except: |
| 239 | pass |
| 240 | else: |
| 241 | wordsDict[word] = {} |
| 242 | wordsDict[word][(module, symbol)] = relevance |
| 243 | return relevance |
| 244 | |
| 245 | def addString(str, module, symbol, relevance): |
| 246 | if str == None or len(str) < 3: |
| 247 | return -1 |
| 248 | ret = 0 |
| 249 | str = string.replace(str, ".", " ") |
| 250 | str = string.replace(str, ",", " ") |
| 251 | str = string.replace(str, "'", " ") |
| 252 | str = string.replace(str, '"', " ") |
| 253 | str = string.replace(str, ";", " ") |
| 254 | str = string.replace(str, "-", " ") |
| 255 | l = string.split(str) |
| 256 | for word in l: |
| 257 | if len(word) > 2: |
| 258 | ret = ret + addWord(word, module, symbol, 5) |
| 259 | |
| 260 | return ret |
| 261 | |
| 262 | |
| 263 | ######################################################################### |
| 264 | # # |
| 265 | # XML API description analysis # |
| 266 | # # |
| 267 | ######################################################################### |
| 268 | |
| 269 | def loadAPI(filename): |
| 270 | doc = libxml2.parseFile(filename) |
| 271 | print "loaded %s" % (filename) |
| 272 | return doc |
| 273 | |
| 274 | def foundExport(file, symbol): |
| 275 | if file == None: |
| 276 | return 0 |
| 277 | if symbol == None: |
| 278 | return 0 |
| 279 | addFunction(symbol, file) |
| 280 | l = splitIdentifier(symbol) |
| 281 | for word in l: |
| 282 | addWord(word, file, symbol, 10) |
| 283 | return 1 |
| 284 | |
| 285 | def analyzeAPIFile(top): |
| 286 | count = 0 |
| 287 | name = top.prop("name") |
| 288 | cur = top.children |
| 289 | while cur != None: |
| 290 | if cur.type == 'text': |
| 291 | cur = cur.next |
| 292 | continue |
| 293 | if cur.name == "exports": |
| 294 | count = count + foundExport(name, cur.prop("symbol")) |
| 295 | else: |
| 296 | print "unexpected element %s in API doc <file name='%s'>" % (name) |
| 297 | cur = cur.next |
| 298 | return count |
| 299 | |
| 300 | def analyzeAPIFiles(top): |
| 301 | count = 0 |
| 302 | cur = top.children |
| 303 | |
| 304 | while cur != None: |
| 305 | if cur.type == 'text': |
| 306 | cur = cur.next |
| 307 | continue |
| 308 | if cur.name == "file": |
| 309 | count = count + analyzeAPIFile(cur) |
| 310 | else: |
| 311 | print "unexpected element %s in API doc <files>" % (cur.name) |
| 312 | cur = cur.next |
| 313 | return count |
| 314 | |
| 315 | def analyzeAPIEnum(top): |
| 316 | file = top.prop("file") |
| 317 | if file == None: |
| 318 | return 0 |
| 319 | symbol = top.prop("name") |
| 320 | if symbol == None: |
| 321 | return 0 |
| 322 | |
| 323 | addEnum(symbol, file) |
| 324 | l = splitIdentifier(symbol) |
| 325 | for word in l: |
| 326 | addWord(word, file, symbol, 10) |
| 327 | |
| 328 | return 1 |
| 329 | |
| 330 | def analyzeAPIConst(top): |
| 331 | file = top.prop("file") |
| 332 | if file == None: |
| 333 | return 0 |
| 334 | symbol = top.prop("name") |
| 335 | if symbol == None: |
| 336 | return 0 |
| 337 | |
| 338 | addConst(symbol, file) |
| 339 | l = splitIdentifier(symbol) |
| 340 | for word in l: |
| 341 | addWord(word, file, symbol, 10) |
| 342 | |
| 343 | return 1 |
| 344 | |
| 345 | def analyzeAPIType(top): |
| 346 | file = top.prop("file") |
| 347 | if file == None: |
| 348 | return 0 |
| 349 | symbol = top.prop("name") |
| 350 | if symbol == None: |
| 351 | return 0 |
| 352 | |
| 353 | addType(symbol, file) |
| 354 | l = splitIdentifier(symbol) |
| 355 | for word in l: |
| 356 | addWord(word, file, symbol, 10) |
| 357 | return 1 |
| 358 | |
| 359 | def analyzeAPIFunctype(top): |
| 360 | file = top.prop("file") |
| 361 | if file == None: |
| 362 | return 0 |
| 363 | symbol = top.prop("name") |
| 364 | if symbol == None: |
| 365 | return 0 |
| 366 | |
| 367 | addFunctype(symbol, file) |
| 368 | l = splitIdentifier(symbol) |
| 369 | for word in l: |
| 370 | addWord(word, file, symbol, 10) |
| 371 | return 1 |
| 372 | |
| 373 | def analyzeAPIStruct(top): |
| 374 | file = top.prop("file") |
| 375 | if file == None: |
| 376 | return 0 |
| 377 | symbol = top.prop("name") |
| 378 | if symbol == None: |
| 379 | return 0 |
| 380 | |
| 381 | addStruct(symbol, file) |
| 382 | l = splitIdentifier(symbol) |
| 383 | for word in l: |
| 384 | addWord(word, file, symbol, 10) |
| 385 | |
| 386 | info = top.prop("info") |
| 387 | if info != None: |
| 388 | l = string.split(info) |
| 389 | for word in l: |
| 390 | if len(word) > 2: |
| 391 | addWord(word, file, symbol, 5) |
| 392 | return 1 |
| 393 | |
| 394 | def analyzeAPIMacro(top): |
| 395 | file = top.prop("file") |
| 396 | if file == None: |
| 397 | return 0 |
| 398 | symbol = top.prop("name") |
| 399 | if symbol == None: |
| 400 | return 0 |
| 401 | |
| 402 | info = None |
| 403 | cur = top.children |
| 404 | while cur != None: |
| 405 | if cur.type == 'text': |
| 406 | cur = cur.next |
| 407 | continue |
| 408 | if cur.name == "info": |
| 409 | info = cur.content |
| 410 | break |
| 411 | cur = cur.next |
| 412 | |
| 413 | l = splitIdentifier(symbol) |
| 414 | for word in l: |
| 415 | addWord(word, file, symbol, 10) |
| 416 | |
| 417 | if info == None: |
| 418 | addMacro(symbol, file) |
| 419 | print "Macro %s description has no <info>" % (symbol) |
| 420 | return 0 |
| 421 | |
| 422 | addMacro(symbol, file, info) |
| 423 | l = string.split(info) |
| 424 | for word in l: |
| 425 | if len(word) > 2: |
| 426 | addWord(word, file, symbol, 5) |
| 427 | return 1 |
| 428 | |
| 429 | def analyzeAPIFunction(top): |
| 430 | file = top.prop("file") |
| 431 | if file == None: |
| 432 | return 0 |
| 433 | symbol = top.prop("name") |
| 434 | if symbol == None: |
| 435 | return 0 |
| 436 | |
| 437 | info = None |
| 438 | cur = top.children |
| 439 | while cur != None: |
| 440 | if cur.type == 'text': |
| 441 | cur = cur.next |
| 442 | continue |
| 443 | if cur.name == "info": |
| 444 | info = cur.content |
| 445 | elif cur.name == "return": |
| 446 | rinfo = cur.prop("info") |
| 447 | if rinfo != None: |
| 448 | addString(rinfo, file, symbol, 7) |
| 449 | elif cur.name == "arg": |
| 450 | ainfo = cur.prop("info") |
| 451 | if rinfo != None: |
| 452 | addString(ainfo, file, symbol, 5) |
| 453 | name = cur.prop("name") |
| 454 | if name != None: |
| 455 | addWord(name, file, symbol, 7) |
| 456 | cur = cur.next |
| 457 | if info == None: |
| 458 | print "Function %s description has no <info>" % (symbol) |
| 459 | addFunction(symbol, file, "") |
| 460 | else: |
| 461 | addFunction(symbol, file, info) |
| 462 | addString(info, file, symbol, 5) |
| 463 | |
| 464 | l = splitIdentifier(symbol) |
| 465 | for word in l: |
| 466 | addWord(word, file, symbol, 10) |
| 467 | |
| 468 | return 1 |
| 469 | |
| 470 | def analyzeAPISymbols(top): |
| 471 | count = 0 |
| 472 | cur = top.children |
| 473 | |
| 474 | while cur != None: |
| 475 | if cur.type == 'text': |
| 476 | cur = cur.next |
| 477 | continue |
| 478 | if cur.name == "macro": |
| 479 | count = count + analyzeAPIMacro(cur) |
| 480 | elif cur.name == "function": |
| 481 | count = count + analyzeAPIFunction(cur) |
| 482 | elif cur.name == "const": |
| 483 | count = count + analyzeAPIConst(cur) |
| 484 | elif cur.name == "typedef": |
| 485 | count = count + analyzeAPIType(cur) |
| 486 | elif cur.name == "struct": |
| 487 | count = count + analyzeAPIStruct(cur) |
| 488 | elif cur.name == "enum": |
| 489 | count = count + analyzeAPIEnum(cur) |
| 490 | elif cur.name == "functype": |
| 491 | count = count + analyzeAPIFunctype(cur) |
| 492 | else: |
| 493 | print "unexpected element %s in API doc <files>" % (cur.name) |
| 494 | cur = cur.next |
| 495 | return count |
| 496 | |
| 497 | def analyzeAPI(doc): |
| 498 | count = 0 |
| 499 | if doc == None: |
| 500 | return -1 |
| 501 | root = doc.getRootElement() |
| 502 | if root.name != "api": |
| 503 | print "Unexpected root name" |
| 504 | return -1 |
| 505 | cur = root.children |
| 506 | while cur != None: |
| 507 | if cur.type == 'text': |
| 508 | cur = cur.next |
| 509 | continue |
| 510 | if cur.name == "files": |
| 511 | pass |
| 512 | # count = count + analyzeAPIFiles(cur) |
| 513 | elif cur.name == "symbols": |
| 514 | count = count + analyzeAPISymbols(cur) |
| 515 | else: |
| 516 | print "unexpected element %s in API doc" % (cur.name) |
| 517 | cur = cur.next |
| 518 | return count |
| 519 | |
| 520 | ######################################################################### |
| 521 | # # |
| 522 | # Main code: open the DB, the API XML and analyze it # |
| 523 | # # |
| 524 | ######################################################################### |
| 525 | try: |
| 526 | openMySQL() |
| 527 | except: |
| 528 | print "Failed to open the database" |
| 529 | print sys.exc_type, sys.exc_value |
| 530 | sys.exit(1) |
| 531 | |
| 532 | try: |
| 533 | doc = loadAPI(API) |
| 534 | ret = analyzeAPI(doc) |
| 535 | print "Analyzed %d blocs" % (ret) |
| 536 | doc.freeDoc() |
| 537 | except: |
| 538 | print "Failed to parse and analyze %s" % (API) |
| 539 | print sys.exc_type, sys.exc_value |
| 540 | sys.exit(1) |
| 541 | |
| 542 | print "Indexed %d words" % (len(wordsDict)) |
| 543 | i = 0 |
| 544 | skipped = 0 |
| 545 | for word in wordsDict.keys(): |
| 546 | refs = wordsDict[word] |
| 547 | if refs == None: |
| 548 | skipped = skipped + 1 |
| 549 | continue; |
| 550 | for (module, symbol) in refs.keys(): |
| 551 | updateWord(word, symbol, refs[(module, symbol)]) |
| 552 | i = i + 1 |
| 553 | |
| 554 | print "Found %d associations, skipped %d words" % (i, skipped) |