SQLCatalog.py 92.7 KB
Newer Older
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1 2
##############################################################################
#
3
# Copyright (c) 2002-2009 Nexedi SARL. All Rights Reserved.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
4 5 6 7 8 9 10 11 12 13 14 15 16 17
# Copyright (c) 2001 Zope Corporation and Contributors. All Rights Reserved.
#
# This software is subject to the provisions of the Zope Public License,
# Version 2.0 (ZPL).  A copy of the ZPL should accompany this distribution.
# THIS SOFTWARE IS PROVIDED "AS IS" AND ANY AND ALL EXPRESS OR IMPLIED
# WARRANTIES ARE DISCLAIMED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
# WARRANTIES OF TITLE, MERCHANTABILITY, AGAINST INFRINGEMENT, AND FITNESS
# FOR A PARTICULAR PURPOSE
#
##############################################################################

from Persistence import Persistent
import Acquisition
import ExtensionClass
18
import Globals
19
import OFS.History
20
from Globals import DTMLFile, PersistentMapping
21
from thread import allocate_lock, get_ident
22
from OFS.Folder import Folder
23
from AccessControl import ClassSecurityInfo
24
from BTrees.OIBTree import OIBTree
25 26
from App.config import getConfiguration
from BTrees.Length import Length
27
from Shared.DC.ZRDB.TM import TM
Jean-Paul Smets's avatar
Jean-Paul Smets committed
28

29 30
from Acquisition import aq_parent, aq_inner, aq_base
from zLOG import LOG, WARNING, INFO, TRACE
31
from ZODB.POSException import ConflictError
32
from Products.PythonScripts.Utility import allow_class
Jean-Paul Smets's avatar
Jean-Paul Smets committed
33 34

import time
35
import sys
36 37
import urllib
import string
38
import pprint
39
from cStringIO import StringIO
40
from xml.dom.minidom import parse
41
from xml.sax.saxutils import escape, quoteattr
42 43
import os
import md5
44 45 46 47

from Interface.IQueryCatalog import ISearchKeyCatalog
from Interface.Verify import verifyClass

48 49
from SearchText import isAdvancedSearchText

50 51 52 53 54 55
PROFILING_ENABLED = False
if PROFILING_ENABLED:
  from tiny_profiler import profiler_decorator, profiler_report, profiler_reset
else:
  def profiler_decorator(func):
    return func
56 57 58 59 60 61 62 63 64

try:
  from Products.CMFCore.Expression import Expression
  from Products.PageTemplates.Expressions import getEngine
  from Products.CMFCore.utils import getToolByName
  withCMF = 1
except ImportError:
  withCMF = 0

65 66 67 68
try:
  import psyco
except ImportError:
  psyco = None
Jean-Paul Smets's avatar
Jean-Paul Smets committed
69

70
try:
71 72
  from Products.ERP5Type.Cache import enableReadOnlyTransactionCache
  from Products.ERP5Type.Cache import disableReadOnlyTransactionCache, CachingMethod
73
except ImportError:
74
  LOG('SQLCatalog', WARNING, 'Count not import CachingMethod, expect slowness.')
75 76
  def doNothing(context):
    pass
77 78 79 80 81 82
  class CachingMethod:
    """
      Dummy CachingMethod class.
    """
    def __init__(self, callable, **kw):
      self.function = callable
Yoshinori Okuji's avatar
Yoshinori Okuji committed
83
    def __call__(self, *opts, **kw):
84
      return self.function(*opts, **kw)
85 86
  enableReadOnlyTransactionCache = doNothing
  disableReadOnlyTransactionCache = doNothing
87 88 89 90 91 92 93 94 95 96 97 98 99

class caching_class_method_decorator:
  def __init__(self, *args, **kw):
    self.args = args
    self.kw = kw

  def __call__(self, method):
    caching_method = CachingMethod(method, *self.args, **self.kw)
    return lambda *args, **kw: caching_method(*args, **kw)

try:
  from Products.ERP5Type.TransactionalVariable import getTransactionalVariable
except ImportError:
100
  LOG('SQLCatalog', WARNING, 'Count not import getTransactionalVariable, expect slowness.')
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
  def getTransactionalVariable(context):
    return {}

class transactional_cache_decorator:
  """
    Implements singleton-style caching.
    Wrapped method must have no parameters (besides "self").
  """
  def __init__(self, cache_id):
    self.cache_id = cache_id

  def __call__(self, method):
    def wrapper(wrapped_self):
      transactional_cache = getTransactionalVariable(None)
      try:
        return transactional_cache[self.cache_id]
      except KeyError:
        result = transactional_cache[self.cache_id] = method(wrapped_self)
        return result
    return wrapper

122 123 124 125
try:
  from ZPublisher.HTTPRequest import record
except ImportError:
  dict_type_list = (dict, )
126 127
else:
  dict_type_list = (dict, record)
128 129


130
UID_BUFFER_SIZE = 300
131 132
OBJECT_LIST_SIZE = 300
MAX_PATH_LEN = 255
133
RESERVED_KEY_LIST = ('where_expression', 'sort-on', 'sort_on', 'sort-order', 'sort_order', 'limit',
134 135
                     'format', 'search_mode', 'operator', 'selection_domain', 'selection_report',
                     'extra_key_list', )
136

137
valid_method_meta_type_list = ('Z SQL Method', 'LDIF Method', 'Script (Python)') # Nicer
138

139
manage_addSQLCatalogForm = DTMLFile('dtml/addSQLCatalog',globals())
140

141 142 143 144 145
# Here go uid buffers
# Structure:
#  global_uid_buffer_dict[catalog_path][thread_id] = UidBuffer
global_uid_buffer_dict = {}

146
def manage_addSQLCatalog(self, id, title,
147
             vocab_id='create_default_catalog_', # vocab_id is a strange name - not abbreviation
148 149 150
             REQUEST=None):
  """Add a Catalog object
  """
151 152 153
  id = str(id)
  title = str(title)
  vocab_id = str(vocab_id)
154 155 156
  if vocab_id == 'create_default_catalog_':
    vocab_id = None

157
  c = Catalog(id, title, self)
158 159 160 161
  self._setObject(id, c)
  if REQUEST is not None:
    return self.manage_main(self, REQUEST,update_menu=1)

162 163
class UidBuffer(TM):
  """Uid Buffer class caches a list of reserved uids in a transaction-safe way."""
164

Yoshinori Okuji's avatar
Yoshinori Okuji committed
165
  def __init__(self):
166
    """Initialize some variables.
167

168
      temporary_buffer is used to hold reserved uids created by non-committed transactions.
169

170
      finished_buffer is used to hold reserved uids created by committed-transactions.
171

172
      This distinction is important, because uids by non-committed transactions might become
Yoshinori Okuji's avatar
Yoshinori Okuji committed
173
      invalid afterwards, so they may not be used by other transactions."""
174 175
    self.temporary_buffer = {}
    self.finished_buffer = []
176

177 178 179 180 181 182 183 184
  def _finish(self):
    """Move the uids in the temporary buffer to the finished buffer."""
    tid = get_ident()
    try:
      self.finished_buffer.extend(self.temporary_buffer[tid])
      del self.temporary_buffer[tid]
    except KeyError:
      pass
185

186 187 188 189 190 191 192
  def _abort(self):
    """Erase the uids in the temporary buffer."""
    tid = get_ident()
    try:
      del self.temporary_buffer[tid]
    except KeyError:
      pass
193

194 195 196 197 198 199 200 201
  def __len__(self):
    tid = get_ident()
    l = len(self.finished_buffer)
    try:
      l += len(self.temporary_buffer[tid])
    except KeyError:
      pass
    return l
202

203 204 205 206 207 208 209 210 211 212 213
  def remove(self, value):
    self._register()
    for uid_list in self.temporary_buffer.values():
      try:
        uid_list.remove(value)
      except ValueError:
        pass
    try:
      self.finished_buffer.remove(value)
    except ValueError:
      pass
214

215 216 217 218
  def pop(self):
    self._register()
    tid = get_ident()
    try:
219
      uid = self.temporary_buffer[tid].pop()
220
    except (KeyError, IndexError):
221 222
      uid = self.finished_buffer.pop()
    return uid
223

224 225 226
  def extend(self, iterable):
    self._register()
    tid = get_ident()
Yoshinori Okuji's avatar
Yoshinori Okuji committed
227
    self.temporary_buffer.setdefault(tid, []).extend(iterable)
228

229
related_key_definition_cache = {}
230
related_key_warned_column_set = set()
231

232 233 234 235 236
class Catalog(Folder,
              Persistent,
              Acquisition.Implicit,
              ExtensionClass.Base,
              OFS.History.Historical):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
237 238 239 240
  """ An Object Catalog

  An Object Catalog maintains a table of object metadata, and a
  series of manageable indexes to quickly search for objects
241
  (references in the metadata) that satisfy a search where_expression.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
242 243 244 245 246 247 248 249

  This class is not Zope specific, and can be used in any python
  program to build catalogs of objects.  Note that it does require
  the objects to be Persistent, and thus must be used with ZODB3.

  uid -> the (local) universal ID of objects
  path -> the (local) path of objects

250 251
  If you pass it a keyword argument which is present in sql_catalog_full_text_search_keys
  (in catalog properties), it does a MySQL full-text search.
252 253 254
  Additionally you can pass it a search_mode argument ('natural', 'in_boolean_mode'
  or 'with_query_expansion') to use an advanced search mode ('natural'
  is the default).
255 256 257
  search_mode arg can be given for all full_text keys, or for a specific key by naming
  the argument search_mode_KeyName, or even more specifically, search_mode_Table.Key
  or search_mode_Table_Key
258

Jean-Paul Smets's avatar
Jean-Paul Smets committed
259

260
  brain defined in methods...
Jean-Paul Smets's avatar
Jean-Paul Smets committed
261 262 263 264 265 266

  TODO:

    - optmization: indexing objects should be deferred
      until timeout value or end of transaction
  """
267 268 269

  __implements__ = ISearchKeyCatalog

270
  meta_type = "SQLCatalog"
271
  icon = 'misc_/ZCatalog/ZCatalog.gif' # FIXME: use a different icon
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
  security = ClassSecurityInfo()

  manage_options = (
    {'label': 'Contents',       # TAB: Contents
     'action': 'manage_main',
     'help': ('OFSP','ObjectManager_Contents.stx')},
    {'label': 'Catalog',      # TAB: Catalogged Objects
     'action': 'manage_catalogView',
     'help':('ZCatalog','ZCatalog_Cataloged-Objects.stx')},
    {'label': 'Properties',     # TAB: Properties
     'action': 'manage_propertiesForm',
     'help': ('OFSP','Properties.stx')},
    {'label': 'Filter',     # TAB: Filter
     'action': 'manage_catalogFilter',},
    {'label': 'Find Objects',     # TAB: Find Objects
     'action': 'manage_catalogFind',
     'help':('ZCatalog','ZCatalog_Find-Items-to-ZCatalog.stx')},
    {'label': 'Advanced',       # TAB: Advanced
     'action': 'manage_catalogAdvanced',
     'help':('ZCatalog','ZCatalog_Advanced.stx')},
    {'label': 'Undo',         # TAB: Undo
     'action': 'manage_UndoForm',
     'help': ('OFSP','Undo.stx')},
    {'label': 'Security',       # TAB: Security
     'action': 'manage_access',
     'help': ('OFSP','Security.stx')},
    {'label': 'Ownership',      # TAB: Ownership
     'action': 'manage_owner',
     'help': ('OFSP','Ownership.stx'),}
301
    ) + OFS.History.Historical.manage_options
302

303
  __ac_permissions__= (
304 305 306 307 308

    ('Manage ZCatalog Entries',
     ['manage_catalogObject', 'manage_uncatalogObject',

      'manage_catalogView', 'manage_catalogFind',
309 310
      'manage_catalogFilter',
      'manage_catalogAdvanced',
311 312

      'manage_catalogReindex', 'manage_catalogFoundItems',
313 314
      'manage_catalogClear',
      'manage_main',
315 316 317 318 319 320
      'manage_editFilter',
      ],
     ['Manager']),

    ('Search ZCatalog',
     ['searchResults', '__call__', 'uniqueValuesFor',
321 322
      'all_meta_types', 'valid_roles',
      'getCatalogSearchTableIds',
323
      'getFilterableMethodList',],
324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
     ['Anonymous', 'Manager']),

    ('Import/Export objects',
     ['manage_exportProperties', 'manage_importProperties', ],
     ['Manager']),

    )

  _properties = (
    { 'id'      : 'title',
      'description' : 'The title of this catalog',
      'type'    : 'string',
      'mode'    : 'w' },

    # Z SQL Methods
    { 'id'      : 'sql_catalog_produce_reserved',
      'description' : 'A method to produce new uid values in advance',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_clear_reserved',
      'description' : 'A method to clear reserved uid values',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
349 350 351 352 353
    { 'id'      : 'sql_catalog_reserve_uid',
      'description' : 'A method to reserve a uid value',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
354 355 356 357 358
    { 'id'      : 'sql_catalog_delete_uid',
      'description' : 'A method to delete a uid value',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
359 360
    { 'id'      : 'sql_catalog_object_list',
      'description' : 'Methods to be called to catalog the list of objects',
361 362 363 364 365 366 367 368
      'type'    : 'multiple selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_uncatalog_object',
      'description' : 'Methods to be called to uncatalog an object',
      'type'    : 'multiple selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
369 370 371
    { 'id'      : 'sql_catalog_translation_list',
      'description' : 'Methods to be called to catalog the list of translation objects',
      'type'    : 'selection',
372 373
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
374 375 376
    { 'id'      : 'sql_delete_translation_list',
      'description' : 'Methods to be called to delete translations',
      'type'    : 'selection',
377 378
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
379 380
    { 'id'      : 'sql_clear_catalog',
      'description' : 'The methods which should be called to clear the catalog',
381 382 383
      'type'    : 'multiple selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
384
    { 'id'      : 'sql_record_object_list',
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403
      'description' : 'Method to record catalog information',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_read_recorded_object_list',
      'description' : 'Method to get recorded information',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_delete_recorded_object_list',
      'description' : 'Method to delete recorded information',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_search_results',
      'description' : 'Main method to search the catalog',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
404 405 406 407 408
    { 'id'      : 'sql_search_security',
      'description' : 'Main method to search security',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443
    { 'id'      : 'sql_search_tables',
      'description' : 'Tables to join in the result',
      'type'    : 'multiple selection',
      'select_variable' : 'getTableIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_search_result_keys',
      'description' : 'Keys to display in the result',
      'type'    : 'multiple selection',
      'select_variable' : 'getResultColumnIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_count_results',
      'description' : 'Main method to search the catalog',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_getitem_by_path',
      'description' : 'Get a catalog brain by path',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_getitem_by_uid',
      'description' : 'Get a catalog brain by uid',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_tables',
      'description' : 'Method to get the main catalog tables',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_schema',
      'description' : 'Method to get the main catalog schema',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
444 445 446 447 448
    { 'id'      : 'sql_catalog_index',
      'description' : 'Method to get the main catalog index',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
    { 'id'      : 'sql_unique_values',
      'description' : 'Find unique disctinct values in a column',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_paths',
      'description' : 'List all object paths in catalog',
      'type'    : 'selection',
      'select_variable' : 'getCatalogMethodIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_keyword_search_keys',
      'description' : 'Columns which should be considered as full text search',
      'type'    : 'multiple selection',
      'select_variable' : 'getColumnIds',
      'mode'    : 'w' },
464 465 466 467 468
    { 'id'      : 'sql_catalog_datetime_search_keys',
      'description' : 'Columns which should be considered as full text search',
      'type'    : 'multiple selection',
      'select_variable' : 'getColumnIds',
      'mode'    : 'w' },
469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
    { 'id'      : 'sql_catalog_full_text_search_keys',
      'description' : 'Columns which should be considered as full text search',
      'type'    : 'multiple selection',
      'select_variable' : 'getColumnIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_request_keys',
      'description' : 'Columns which should be ignore in the REQUEST in order to accelerate caching',
      'type'    : 'multiple selection',
      'select_variable' : 'getColumnIds',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_multivalue_keys',
      'description' : 'Keys which hold multiple values',
      'type'    : 'multiple selection',
      'select_variable' : 'getColumnIds',
      'mode'    : 'w' },
484 485 486 487 488
    { 'id'      : 'sql_catalog_index_on_order_keys',
      'description' : 'Columns which should be used by specifying an index when sorting on them',
      'type'    : 'multiple selection',
      'select_variable' : 'getSortColumnIds',
      'mode'    : 'w' },
489 490 491 492 493 494 495 496 497
    { 'id'      : 'sql_catalog_topic_search_keys',
      'description' : 'Columns which should be considered as topic index',
      'type'    : 'lines',
      'mode'    : 'w' },
    { 'id'      : 'sql_catalog_related_keys',
      'title'   : 'Related keys',
      'description' : 'Additional columns obtained through joins',
      'type'    : 'lines',
      'mode'    : 'w' },
498 499 500 501 502
    { 'id'      : 'sql_catalog_scriptable_keys',
      'title'   : 'Related keys',
      'description' : 'Virtual columns to generate scriptable scriptable queries',
      'type'    : 'lines',
      'mode'    : 'w' },
503 504 505 506 507 508 509 510 511 512 513
    { 'id': 'sql_catalog_role_keys',
      'title': 'Role keys',
      'description': 'Columns which should be used to map a monovalued role',
      'type': 'lines',
      'mode': 'w' },
    { 'id': 'sql_catalog_local_role_keys',
      'title': 'Local Role keys',
      'description': 'Columns which should be used to map' \
                      'a monovalued local role',
      'type': 'lines',
      'mode': 'w' },
514 515 516 517 518 519 520
    { 'id': 'sql_catalog_table_vote_scripts',
      'title': 'Table vote scripts',
      'description': 'Scripts helping column mapping resolution',
      'type': 'multiple selection',
      'select_variable' : 'getPythonMethodIds',
      'mode': 'w' },

521 522
  )

523
  sql_catalog_produce_reserved = ''
524
  sql_catalog_delete_uid = ''
525 526 527 528 529 530 531 532 533 534 535
  sql_catalog_clear_reserved = ''
  sql_catalog_reserve_uid = ''
  sql_catalog_object_list = ()
  sql_uncatalog_object = ()
  sql_clear_catalog = ()
  sql_catalog_translation_list = ''
  sql_delete_translation_list = ''
  sql_record_object_list = ''
  sql_read_recorded_object_list = ''
  sql_delete_recorded_object_list = ''
  sql_search_results = ''
536
  sql_search_security = ''
537 538 539 540 541 542
  sql_count_results = ''
  sql_getitem_by_path = ''
  sql_getitem_by_uid = ''
  sql_catalog_tables = ''
  sql_search_tables = ()
  sql_catalog_schema = ''
543
  sql_catalog_index = ''
544 545 546
  sql_unique_values = ''
  sql_catalog_paths = ''
  sql_catalog_keyword_search_keys =  ()
547
  sql_catalog_datetime_search_keys = ()
548 549 550 551 552
  sql_catalog_full_text_search_keys = ()
  sql_catalog_request_keys = ()
  sql_search_result_keys = ()
  sql_catalog_topic_search_keys = ()
  sql_catalog_multivalue_keys = ()
553
  sql_catalog_index_on_order_keys = ()
554
  sql_catalog_related_keys = ()
555
  sql_catalog_scriptable_keys = ()
556 557
  sql_catalog_role_keys = ()
  sql_catalog_local_role_keys = ()
558
  sql_catalog_table_vote_scripts = ()
559

560 561 562 563 564 565
  # These are ZODB variables, so shared by multiple Zope instances.
  # This is set to the last logical time when clearReserved is called.
  _last_clear_reserved_time = 0
  # This is to record the maximum value of uids. Because this uses the class Length
  # in BTrees.Length, this does not generate conflict errors.
  _max_uid = None
566

567 568 569 570 571 572 573
  # These are class variable on memory, so shared only by threads in the same Zope instance.
  # This is set to the time when reserved uids are cleared in this Zope instance.
  _local_clear_reserved_time = None
  # This is used for exclusive access to the list of reserved uids.
  _reserved_uid_lock = allocate_lock()
  # This is an instance id which specifies who owns which reserved uids.
  _instance_id = getattr(getConfiguration(), 'instance_id', None)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
574

575 576 577 578 579 580 581 582 583 584
  manage_catalogView = DTMLFile('dtml/catalogView',globals())
  manage_catalogFilter = DTMLFile('dtml/catalogFilter',globals())
  manage_catalogFind = DTMLFile('dtml/catalogFind',globals())
  manage_catalogAdvanced = DTMLFile('dtml/catalogAdvanced', globals())

  def __init__(self, id, title='', container=None):
    if container is not None:
      self=self.__of__(container)
    self.id=id
    self.title=title
Jean-Paul Smets's avatar
Jean-Paul Smets committed
585 586 587
    self.schema = {}  # mapping from attribute name to column
    self.names = {}   # mapping from column to attribute name
    self.indexes = {}   # empty mapping
588
    self.filter_dict = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
589

590 591 592 593 594 595 596 597 598 599 600 601 602 603
  def getSQLCatalogRoleKeysList(self):
    """
    Return the list of role keys.
    """
    return [tuple([y.strip() for y in x.split('|')]) \
              for x in self.sql_catalog_role_keys]

  def getSQLCatalogLocalRoleKeysList(self):
    """
    Return the list of local role keys.
    """
    return [tuple([y.strip() for y in x.split('|')]) \
              for x in self.sql_catalog_local_role_keys]

604 605 606 607 608 609
  def manage_exportProperties(self, REQUEST=None, RESPONSE=None):
    """
      Export properties to an XML file.
    """
    f = StringIO()
    f.write('<?xml version="1.0"?>\n<SQLCatalogData>\n')
610 611 612 613 614 615 616 617
    property_id_list = self.propertyIds()
    # Get properties and values
    property_list = []
    for property_id in property_id_list:
      value = self.getProperty(property_id)
      if value is not None:
        property_list.append((property_id, value))
    # Sort for easy diff
618
    property_list.sort(key=lambda x: x[0])
619 620 621
    for property in property_list:
      property_id = property[0]
      value       = property[1]
622
      if isinstance(value, basestring):
623
        f.write('  <property id=%s type="str">%s</property>\n' % (quoteattr(property_id), escape(value)))
624
      elif isinstance(value, (tuple, list)):
625 626 627
        f.write('  <property id=%s type="tuple">\n' % quoteattr(property_id))
        # Sort for easy diff
        item_list = []
628
        for item in value:
629
          if isinstance(item, basestring):
630 631 632 633
            item_list.append(item)
        item_list.sort()
        for item in item_list:
          f.write('    <item type="str">%s</item>\n' % escape(str(item)))
634
        f.write('  </property>\n')
635 636 637
    # XXX Although filters are not properties, output filters here.
    # XXX Ideally, filters should be properties in Z SQL Methods, shouldn't they?
    if hasattr(self, 'filter_dict'):
638 639
      filter_list = []
      for filter_id in self.filter_dict.keys():
640
        filter_definition = self.filter_dict[filter_id]
641 642
        filter_list.append((filter_id, filter_definition))
      # Sort for easy diff
643
      filter_list.sort(key=lambda x: x[0])
644 645 646 647
      for filter_item in filter_list:
        filter_id  = filter_item[0]
        filter_def = filter_item[1]
        if not filter_def['filtered']:
648 649
          # If a filter is not activated, no need to output it.
          continue
650
        if not filter_def['expression']:
651 652
          # If the expression is not specified, meaningless to specify it.
          continue
653
        f.write('  <filter id=%s expression=%s />\n' % (quoteattr(filter_id), quoteattr(filter_def['expression'])))
654
        # For now, portal types are not exported, because portal types are too specific to each site.
655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675
    f.write('</SQLCatalogData>\n')

    if RESPONSE is not None:
      RESPONSE.setHeader('Content-type','application/data')
      RESPONSE.setHeader('Content-Disposition',
                          'inline;filename=properties.xml')
    return f.getvalue()

  def manage_importProperties(self, file):
    """
      Import properties from an XML file.
    """
    f = open(file)
    try:
      doc = parse(f)
      root = doc.documentElement
      try:
        for prop in root.getElementsByTagName("property"):
          id = prop.getAttribute("id")
          type = prop.getAttribute("type")
          if not id or not hasattr(self, id):
676
            raise CatalogError, 'unknown property id %r' % (id,)
677
          if type not in ('str', 'tuple'):
678
            raise CatalogError, 'unknown property type %r' % (type,)
679 680 681 682
          if type == 'str':
            value = ''
            for text in prop.childNodes:
              if text.nodeType == text.TEXT_NODE:
683
                value = str(text.data)
684 685 686 687 688 689
                break
          else:
            value = []
            for item in prop.getElementsByTagName("item"):
              item_type = item.getAttribute("type")
              if item_type != 'str':
690
                raise CatalogError, 'unknown item type %r' % (item_type,)
691 692
              for text in item.childNodes:
                if text.nodeType == text.TEXT_NODE:
693
                  value.append(str(text.data))
694 695 696 697
                  break
            value = tuple(value)

          setattr(self, id, value)
698

699 700 701
        if not hasattr(self, 'filter_dict'):
          self.filter_dict = PersistentMapping()
        for filt in root.getElementsByTagName("filter"):
702
          id = str(filt.getAttribute("id"))
703 704 705 706 707 708 709 710 711 712 713 714
          expression = filt.getAttribute("expression")
          if not self.filter_dict.has_key(id):
            self.filter_dict[id] = PersistentMapping()
          self.filter_dict[id]['filtered'] = 1
          self.filter_dict[id]['type'] = []
          if expression:
            expr_instance = Expression(expression)
            self.filter_dict[id]['expression'] = expression
            self.filter_dict[id]['expression_instance'] = expr_instance
          else:
            self.filter_dict[id]['expression'] = ""
            self.filter_dict[id]['expression_instance'] = None
715 716 717 718
      finally:
        doc.unlink()
    finally:
      f.close()
719

720 721 722 723 724 725 726
  def manage_historyCompare(self, rev1, rev2, REQUEST,
                            historyComparisonResults=''):
    return Catalog.inheritedAttribute('manage_historyCompare')(
          self, rev1, rev2, REQUEST,
          historyComparisonResults=OFS.History.html_diff(
              pprint.pformat(rev1.__dict__),
              pprint.pformat(rev2.__dict__)))
727 728 729

  def _clearSecurityCache(self):
    self.security_uid_dict = OIBTree()
730
    self.security_uid_index = None
731 732

  security.declarePrivate('getSecurityUid')
733
  def getSecurityUid(self, wrapped_object):
734 735 736 737 738 739 740
    """
      Cache a uid for each security permission

      We try to create a unique security (to reduce number of lines)
      and to assign security only to root document
    """
    # Get security information
741
    allowed_roles_and_users = wrapped_object.allowedRolesAndUsers()
742 743 744 745
    # Sort it
    allowed_roles_and_users = list(allowed_roles_and_users)
    allowed_roles_and_users.sort()
    allowed_roles_and_users = tuple(allowed_roles_and_users)
746 747
    # Make sure no duplicates
    if getattr(aq_base(self), 'security_uid_dict', None) is None:
748 749 750
      self._clearSecurityCache()
    if self.security_uid_dict.has_key(allowed_roles_and_users):
      return (self.security_uid_dict[allowed_roles_and_users], None)
751 752 753 754 755 756 757 758
    # If the id_tool is there, it is better to use it, it allows
    # to create many new security uids by the same time
    # because with this tool we are sure that we will have 2 different
    # uids if two instances are doing this code in the same time
    id_tool = getattr(self.getPortalObject(), 'portal_ids', None)
    if id_tool is not None:
      default = 1
      # We must keep compatibility with existing sites
759 760
      previous_security_uid = getattr(self, 'security_uid_index', None)
      if previous_security_uid is not None:
761 762 763 764
        # At some point, it was a Length
        if isinstance(previous_security_uid, Length):
          default = previous_security_uid() + 1
        else:
765
          default = previous_security_uid
766 767 768 769 770 771 772 773 774 775 776
      security_uid = id_tool.generateNewLengthId(id_group='security_uid_index',
                                        default=default)
    else:
      previous_security_uid = getattr(self, 'security_uid_index', None)
      if previous_security_uid is None:
        previous_security_uid = 0
      # At some point, it was a Length
      if isinstance(previous_security_uid, Length):
        previous_security_uid = previous_security_uid()
      security_uid = previous_security_uid + 1
      self.security_uid_index = security_uid
777 778
    self.security_uid_dict[allowed_roles_and_users] = security_uid
    return (security_uid, allowed_roles_and_users)
779

780 781 782 783 784 785 786 787 788 789 790
  def getRoleAndSecurityUidList(self):
    """
      Return a list of 2-tuples, suitable for direct use in a zsqlmethod.
      Goal: make it possible to regenerate a table containing this data.
    """
    result = []
    extend = result.extend
    for role_list, security_uid in getattr(aq_base(self), 'security_uid_dict', {}).iteritems():
      extend([(role, security_uid) for role in role_list])
    return result

Jean-Paul Smets's avatar
Jean-Paul Smets committed
791 792 793 794 795 796
  def clear(self):
    """
    Clears the catalog by calling a list of methods
    """
    methods = self.sql_clear_catalog
    for method_name in methods:
797
      method = getattr(self, method_name)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
798 799
      try:
        method()
800 801
      except ConflictError:
        raise
Jean-Paul Smets's avatar
Jean-Paul Smets committed
802
      except:
803
        LOG('SQLCatalog', WARNING,
Romain Courteaud's avatar
Romain Courteaud committed
804
            'could not clear catalog with %s' % method_name, error=sys.exc_info())
805

806
    # Reserved uids have been removed.
807
    self.clearReserved()
808

809
    # Add a dummy item so that SQLCatalog will not use existing uids again.
810
    self.insertMaxUid()
811

812
    # Remove the cache of catalog schema.
813 814
    if hasattr(self, '_v_catalog_schema_dict') :
      del self._v_catalog_schema_dict
815

816
    self._clearSecurityCache()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
817

818 819 820 821 822 823 824 825 826 827
  def insertMaxUid(self):
    """
      Add a dummy item so that SQLCatalog will not use existing uids again.
    """
    if self._max_uid is not None and self._max_uid() != 0:
      method_id = self.sql_catalog_reserve_uid
      method = getattr(self, method_id)
      self._max_uid.change(1)
      method(uid = [self._max_uid()])

828 829 830 831 832 833
  def clearReserved(self):
    """
    Clears reserved uids
    """
    method_id = self.sql_catalog_clear_reserved
    method = getattr(self, method_id)
Romain Courteaud's avatar
Romain Courteaud committed
834 835 836 837 838
    try:
      method()
    except ConflictError:
      raise
    except:
839
      LOG('SQLCatalog', WARNING,
Romain Courteaud's avatar
Romain Courteaud committed
840 841 842
          'could not clear reserved catalog with %s' % \
              method_id, error=sys.exc_info())
      raise
843
    self._last_clear_reserved_time += 1
844

Jean-Paul Smets's avatar
Jean-Paul Smets committed
845 846 847 848 849 850 851 852 853
  def __getitem__(self, uid):
    """
    Get an object by UID
    Note: brain is defined in Z SQL Method object
    """
    method = getattr(self,  self.sql_getitem_by_uid)
    search_result = method(uid = uid)
    if len(search_result) > 0:
      return search_result[0]
854
    raise KeyError, uid
Jean-Paul Smets's avatar
Jean-Paul Smets committed
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874

  def editSchema(self, names_list):
    """
    Builds a schema from a list of strings
    Splits each string to build a list of attribute names
    Columns on the database should not change during this operations
    """
    i = 0
    schema = {}
    names = {}
    for cid in self.getColumnIds():
      name_list = []
      for name in names_list[i].split():
        schema[name] = cid
        name_list += [name,]
      names[cid] = tuple(name_list)
      i += 1
    self.schema = schema
    self.names = names

875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891
  def getCatalogSearchTableIds(self):
    """Return selected tables of catalog which are used in JOIN.
       catalaog is always first
    """
    search_tables = self.sql_search_tables
    if len(search_tables) > 0:
      if search_tables[0] != 'catalog':
        result = ['catalog']
        for t in search_tables:
          if t != 'catalog':
            result.append(t)
        self.sql_search_tables = result
    else:
      self.sql_search_tables = ['catalog']

    return self.sql_search_tables

892
  security.declarePublic('getCatalogSearchResultKeys')
893 894 895 896
  def getCatalogSearchResultKeys(self):
    """Return search result keys.
    """
    return self.sql_search_result_keys
897

898
  def _getCatalogSchema(self, table=None):
899 900 901
    # XXX: Using a volatile as a cache makes it impossible to flush
    # consistently on all connections containing the volatile. Another
    # caching scheme must be used here.
902
    catalog_schema_dict = getattr(aq_base(self), '_v_catalog_schema_dict', {})
903

904 905 906 907 908 909 910 911
    if table not in catalog_schema_dict:
      result_list = []
      try:
        method_name = self.sql_catalog_schema
        method = getattr(self, method_name)
        search_result = method(table=table)
        for c in search_result:
          result_list.append(c.Field)
912 913
      except ConflictError:
        raise
914
      except:
915
        LOG('SQLCatalog', WARNING, '_getCatalogSchema failed with the method %s' % method_name, error=sys.exc_info())
916 917 918
        pass
      catalog_schema_dict[table] = tuple(result_list)
      self._v_catalog_schema_dict= catalog_schema_dict
919

920
    return catalog_schema_dict[table]
921

Jean-Paul Smets's avatar
Jean-Paul Smets committed
922 923
  def getColumnIds(self):
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
924 925 926
    Calls the show column method and returns dictionnary of
    Field Ids
    """
927
    def _getColumnIds():
928 929 930 931
      keys = {}
      for table in self.getCatalogSearchTableIds():
        field_list = self._getCatalogSchema(table=table)
        for field in field_list:
932 933
          keys[field] = None
          keys['%s.%s' % (table, field)] = None  # Is this inconsistent ?
934
      for related in self.getSQLCatalogRelatedKeyList():
935 936
        related_tuple = related.split('|')
        related_key = related_tuple[0].strip()
937
        keys[related_key] = None
938 939 940
      for scriptable in self.getSQLCatalogScriptableKeyList():
        scriptable_tuple = scriptable.split('|')
        scriptable = scriptable_tuple[0].strip()
941
        keys[scriptable] = None
942 943 944
      keys = keys.keys()
      keys.sort()
      return keys
945
    return CachingMethod(_getColumnIds, id='SQLCatalog.getColumnIds', cache_factory='erp5_content_long')()[:]
Jean-Paul Smets's avatar
Jean-Paul Smets committed
946

947 948 949 950 951
  @profiler_decorator
  @transactional_cache_decorator('SQLCatalog.getColumnMap')
  @profiler_decorator
  @caching_class_method_decorator(id='SQLCatalog.getColumnMap', cache_factory='erp5_content_long')
  @profiler_decorator
952 953 954 955 956
  def getColumnMap(self):
    """
    Calls the show column method and returns dictionnary of
    Field Ids
    """
957 958 959 960 961 962
    result = {}
    for table in self.getCatalogSearchTableIds():
      for field in self._getCatalogSchema(table=table):
        result.setdefault(field, []).append(table)
        result.setdefault('%s.%s' % (table, field), []).append(table) # Is this inconsistent ?
    return result
963

Jean-Paul Smets's avatar
Jean-Paul Smets committed
964 965 966 967 968 969 970
  def getResultColumnIds(self):
    """
    Calls the show column method and returns dictionnary of
    Field Ids
    """
    keys = {}
    for table in self.getCatalogSearchTableIds():
971 972 973
      field_list = self._getCatalogSchema(table=table)
      for field in field_list:
        keys['%s.%s' % (table, field)] = 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
974 975 976 977
    keys = keys.keys()
    keys.sort()
    return keys

978 979 980 981 982 983 984 985 986 987 988 989 990 991
  def getSortColumnIds(self):
    """
    Calls the show column method and returns dictionnary of
    Field Ids that can be used for a sort
    """
    keys = {}
    for table in self.getTableIds():
      field_list = self._getCatalogSchema(table=table)
      for field in field_list:
        keys['%s.%s' % (table, field)] = 1
    keys = keys.keys()
    keys.sort()
    return keys

Jean-Paul Smets's avatar
Jean-Paul Smets committed
992 993 994
  def getTableIds(self):
    """
    Calls the show table method and returns dictionnary of
Jean-Paul Smets's avatar
Jean-Paul Smets committed
995 996 997
    Field Ids
    """
    keys = []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
998 999
    method_name = self.sql_catalog_tables
    try:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1000 1001 1002
      method = getattr(self,  method_name)
      search_result = method()
      for c in search_result:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1003
        keys.append(c[0])
1004 1005
    except ConflictError:
      raise
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1006 1007
    except:
      pass
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1008 1009
    return keys

1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
  def getUIDBuffer(self, force_new_buffer=False):
    global global_uid_buffer_dict
    klass = self.__class__
    assert klass._reserved_uid_lock.locked()
    assert getattr(self, 'aq_base', None) is not None
    instance_key = self.getPhysicalPath()
    if instance_key not in global_uid_buffer_dict:
      global_uid_buffer_dict[instance_key] = {}
    uid_buffer_dict = global_uid_buffer_dict[instance_key]
    thread_key = get_ident()
    if force_new_buffer or thread_key not in uid_buffer_dict:
      uid_buffer_dict[thread_key] = UidBuffer()
    return uid_buffer_dict[thread_key]
  
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1024
  # the cataloging API
1025 1026 1027
  def produceUid(self):
    """
      Produces reserved uids in advance
1028
    """
1029 1030 1031 1032
    klass = self.__class__
    assert klass._reserved_uid_lock.locked()
    # This checks if the list of local reserved uids was cleared after clearReserved
    # had been called.
1033 1034 1035 1036
    force_new_buffer = (klass._local_clear_reserved_time != self._last_clear_reserved_time)
    uid_buffer = self.getUIDBuffer(force_new_buffer=force_new_buffer)
    klass._local_clear_reserved_time = self._last_clear_reserved_time
    if len(uid_buffer) == 0:
1037 1038 1039
      id_tool = getattr(self.getPortalObject(), 'portal_ids', None)
      if id_tool is not None:
        if self._max_uid is None:
1040
          self._max_uid = Length(1)
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
        uid_list = id_tool.generateNewLengthIdList(id_group='catalog_uid',
                     id_count=UID_BUFFER_SIZE, default=self._max_uid())
        # TODO: if this method is kept and former uid allocation code is
        # discarded, self._max_uid duplicates work done by portal_ids: it
        # already keeps track of the highest allocated number for all id
        # generator groups.
      else:
        method_id = self.sql_catalog_produce_reserved
        method = getattr(self, method_id)
        # Generate an instance id randomly. Note that there is a small possibility that this
        # would conflict with others.
        random_factor_list = [time.time(), os.getpid(), os.times()]
        try:
          random_factor_list.append(os.getloadavg())
        except (OSError, AttributeError): # AttributeError is required under cygwin
          pass
        instance_id = md5.new(str(random_factor_list)).hexdigest()
        uid_list = [x.uid for x in method(count = UID_BUFFER_SIZE, instance_id = instance_id) if x.uid != 0]
1059
      uid_buffer.extend(uid_list)
1060

1061 1062 1063 1064 1065 1066
  def isIndexable(self):
    """
    This is required to check in many methods that
    the site root and zope root are indexable
    """
    zope_root = self.getZopeRoot()
1067
    site_root = self.getSiteRoot() # XXX-JPS - Why don't we use getPortalObject here ?
1068 1069 1070 1071 1072 1073

    root_indexable = int(getattr(zope_root, 'isIndexable', 1))
    site_indexable = int(getattr(site_root, 'isIndexable', 1))
    if not (root_indexable and site_indexable):
      return False
    return True
1074

1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094
  def getSiteRoot(self):
    """
    Returns the root of the site
    """
    if withCMF:
      site_root = getToolByName(self, 'portal_url').getPortalObject()
    else:
      site_root = self.aq_parent
    return site_root

  def getZopeRoot(self):
    """
    Returns the root of the zope
    """
    if withCMF:
      zope_root = getToolByName(self, 'portal_url').getPortalObject().aq_parent
    else:
      zope_root = self.getPhysicalRoot()
    return zope_root

1095 1096 1097
  def newUid(self):
    """
      This is where uid generation takes place. We should consider a multi-threaded environment
1098 1099
      with multiple ZEO clients on a single ZEO server.

1100
      The main risk is the following:
1101

1102
      - objects a/b/c/d/e/f are created (a is parent of b which is parent of ... of f)
1103

1104
      - one reindexing node N1 starts reindexing f
1105

1106
      - another reindexing node N2 starts reindexing e
1107

1108 1109 1110
      - there is a strong risk that N1 and N2 start reindexing at the same time
        and provide different uid values for a/b/c/d/e

1111
      Similar problems may happen with relations and acquisition of uid values (ex. order_uid)
1112
      with the risk of graph loops
1113
    """
1114
    if not self.isIndexable():
1115 1116
      return None

1117 1118 1119 1120
    klass = self.__class__
    try:
      klass._reserved_uid_lock.acquire()
      self.produceUid()
1121 1122 1123
      uid_buffer = self.getUIDBuffer()
      if len(uid_buffer) > 0:
        uid = uid_buffer.pop()
1124
        if self._max_uid is None:
1125
          self._max_uid = Length(1)
1126 1127
        if uid > self._max_uid():
          self._max_uid.set(uid)
1128
        return long(uid)
1129 1130 1131 1132
      else:
        raise CatalogError("Could not retrieve new uid")
    finally:
      klass._reserved_uid_lock.release()
1133

1134 1135 1136
  def manage_catalogObject(self, REQUEST, RESPONSE, URL1, urls=None):
    """ index Zope object(s) that 'urls' point to """
    if urls:
1137
      if isinstance(urls, str):
1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
        urls=(urls,)

      for url in urls:
        obj = self.resolve_path(url)
        if not obj:
          obj = self.resolve_url(url, REQUEST)
        if obj is not None:
          self.aq_parent.catalog_object(obj, url, sql_catalog_id=self.id)

    RESPONSE.redirect(URL1 + '/manage_catalogView?manage_tabs_message=Object%20Cataloged')

  def manage_uncatalogObject(self, REQUEST, RESPONSE, URL1, urls=None):
    """ removes Zope object(s) 'urls' from catalog """

    if urls:
1153
      if isinstance(urls, str):
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
        urls=(urls,)

      for url in urls:
        self.aq_parent.uncatalog_object(url, sql_catalog_id=self.id)

    RESPONSE.redirect(URL1 + '/manage_catalogView?manage_tabs_message=Object%20Uncataloged')

  def manage_catalogReindex(self, REQUEST, RESPONSE, URL1):
    """ clear the catalog, then re-index everything """
    elapse = time.time()
    c_elapse = time.clock()

    self.aq_parent.refreshCatalog(clear=1, sql_catalog_id=self.id)

    elapse = time.time() - elapse
    c_elapse = time.clock() - c_elapse

    RESPONSE.redirect(URL1 +
              '/manage_catalogAdvanced?manage_tabs_message=' +
              urllib.quote('Catalog Updated<br>'
                     'Total time: %s<br>'
                     'Total CPU time: %s' % (`elapse`, `c_elapse`)))

1177
  def manage_catalogClear(self, REQUEST=None, RESPONSE=None,
Romain Courteaud's avatar
Romain Courteaud committed
1178
                          URL1=None, sql_catalog_id=None):
1179
    """ clears the whole enchilada """
1180 1181
    self.beforeCatalogClear()

1182 1183 1184
    self.clear()

    if RESPONSE and URL1:
Romain Courteaud's avatar
Romain Courteaud committed
1185
      RESPONSE.redirect('%s/manage_catalogAdvanced?' \
1186
                        'manage_tabs_message=Catalog%%20Cleared' % URL1)
1187 1188

  def manage_catalogClearReserved(self, REQUEST=None, RESPONSE=None, URL1=None):
1189
    """ clears reserved uids """
1190 1191 1192
    self.clearReserved()

    if RESPONSE and URL1:
Romain Courteaud's avatar
Romain Courteaud committed
1193
      RESPONSE.redirect('%s/manage_catalogAdvanced?' \
1194
                        'manage_tabs_message=Catalog%%20Cleared' % URL1)
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231

  def manage_catalogFoundItems(self, REQUEST, RESPONSE, URL2, URL1,
                 obj_metatypes=None,
                 obj_ids=None, obj_searchterm=None,
                 obj_expr=None, obj_mtime=None,
                 obj_mspec=None, obj_roles=None,
                 obj_permission=None):
    """ Find object according to search criteria and Catalog them
    """
    elapse = time.time()
    c_elapse = time.clock()

    words = 0
    obj = REQUEST.PARENTS[1]
    path = string.join(obj.getPhysicalPath(), '/')

    results = self.aq_parent.ZopeFindAndApply(obj,
                    obj_metatypes=obj_metatypes,
                    obj_ids=obj_ids,
                    obj_searchterm=obj_searchterm,
                    obj_expr=obj_expr,
                    obj_mtime=obj_mtime,
                    obj_mspec=obj_mspec,
                    obj_permission=obj_permission,
                    obj_roles=obj_roles,
                    search_sub=1,
                    REQUEST=REQUEST,
                    apply_func=self.aq_parent.catalog_object,
                    apply_path=path,
                    sql_catalog_id=self.id)

    elapse = time.time() - elapse
    c_elapse = time.clock() - c_elapse

    RESPONSE.redirect(URL1 + '/manage_catalogView?manage_tabs_message=' +
              urllib.quote('Catalog Updated<br>Total time: %s<br>Total CPU time: %s' % (`elapse`, `c_elapse`)))

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1232
  def catalogObject(self, object, path, is_object_moved=0):
1233 1234
    """Add an object to the Catalog by calling all SQL methods and
    providing needed arguments.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1235

1236 1237
    'object' is the object to be catalogged."""
    self._catalogObjectList([object])
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1238

1239 1240 1241 1242
  def catalogObjectList(self, object_list, method_id_list=None, 
                        disable_cache=0, check_uid=1, idxs=None):
    """Add objects to the Catalog by calling all SQL methods and
    providing needed arguments.
1243

1244 1245
      method_id_list : specify which methods should be used. If not
                       set, it will take the default value of portal_catalog.
1246 1247

      disable_cache : do not use cache, so values will be computed each time,
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
                      only useful in some particular cases, most of the time
                      you don't need to use it.

    Each element of 'object_list' is an object to be catalogged.

    'uid' is the unique Catalog identifier for this object.
    
    Note that this method calls _catalogObjectList with fragments of
    the object list, because calling _catalogObjectList with too many
    objects at a time bloats the process's memory consumption, due to
    caching."""
    # XXX 300 is arbitrary.
1260 1261
    for i in xrange(0, len(object_list), OBJECT_LIST_SIZE):
      self._catalogObjectList(object_list[i:i + OBJECT_LIST_SIZE],
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
                              method_id_list=method_id_list,
                              disable_cache=disable_cache,
                              check_uid=check_uid,
                              idxs=idxs)
    
  def _catalogObjectList(self, object_list, method_id_list=None, 
                         disable_cache=0, check_uid=1, idxs=None):
    """This is the real method to catalog objects.

    XXX: For now newUid is used to allocated UIDs. Is this good?
    Is it better to INSERT then SELECT?"""
1273
    LOG('SQLCatalog', TRACE, 'catalogging %d objects' % len(object_list))
1274
    #LOG('catalogObjectList', 0, 'called with %r' % (object_list,))
1275

1276
    if idxs not in (None, []):
1277 1278
      LOG('ZSLQCatalog.SQLCatalog:catalogObjectList', WARNING, 
          'idxs is ignored in this function and is only provided to be compatible with CMFCatalogAware.reindexObject.')
1279

1280
    if not self.isIndexable():
1281
      return
1282

1283 1284
    # Reminder about optimization: It might be possible to issue just one
    # query to get enought results to check uid & path consistency.
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
    path_uid_dict = {}
    uid_path_dict = {}

    if check_uid:
      path_list = []
      path_list_append = path_list.append
      uid_list = []
      uid_list_append = uid_list.append
      for object in object_list:
        path = object.getPath()
        if path is not None:
          path_list_append(path)
        uid = object.uid
        if uid is not None:
          uid_list_append(uid)
      path_uid_dict = self.getUidDictForPathList(path_list=path_list)
      uid_path_dict = self.getPathDictForUidList(uid_list=uid_list)

1303 1304 1305 1306 1307
    # This dict will store uids and objects which are verified below.
    # The purpose is to prevent multiple objects from having the same
    # uid inside object_list.
    assigned_uid_dict = {}

1308
    for object in object_list:
1309
      if getattr(aq_base(object), 'uid', None) is None:
1310
        try:
1311
          object.uid = self.newUid()
1312 1313
        except ConflictError:
          raise
1314
        except:
1315
          raise RuntimeError, 'could not set missing uid for %r' % (object,)
1316 1317 1318 1319
      elif isinstance(object.uid, int) and object.uid == 0:
        # Several Tool objects have uid=0 (not 0L) from the beginning, but
        # we need an unique uid for each object.
        object.uid = self.newUid()
1320
      elif check_uid:
1321
        uid = object.uid
1322 1323 1324 1325 1326 1327
        if uid in assigned_uid_dict:
          object.uid = self.newUid()
          LOG('SQLCatalog', WARNING,
              'uid of %r changed from %r to %r as old one is assigned to %r !!! This can be fatal. You should reindex the whole site immediately.' % (object, uid, object.uid, assigned_uid_dict[uid]))
          uid = object.uid

1328
        path = object.getPath()
1329
        index = path_uid_dict.get(path, None)
1330
        try:
1331
          index = long(index)
1332
        except TypeError:
1333
          index = None
1334 1335 1336
        if index is not None:
          if index < 0:
            raise CatalogError, 'A negative uid %d is used for %s. Your catalog is broken. Recreate your catalog.' % (index, path)
1337 1338
          if uid != index or isinstance(uid, int):
            # We want to make sure that uid becomes long if it is an int
1339
            LOG('SQLCatalog', WARNING, 'uid of %r changed from %r (property) to %r (catalog, by path) !!! This can be fatal. You should reindex the whole site immediately.' % (object, uid, index))
1340 1341 1342 1343 1344 1345
            uid = index
            object.uid = uid
        else:
          # Make sure no duplicates - ie. if an object with different path has same uid, we need a new uid
          # This can be very dangerous with relations stored in a category table (CMFCategory)
          # This is why we recommend completely reindexing subobjects after any change of id
1346
          if uid in uid_path_dict:
1347 1348 1349
            catalog_path = uid_path_dict.get(uid)
          else:
            catalog_path = self.getPathForUid(uid)
1350 1351 1352 1353 1354 1355
          #LOG('catalogObject', 0, 'uid = %r, catalog_path = %r' % (uid, catalog_path))
          if catalog_path == "reserved":
            # Reserved line in catalog table
            klass = self.__class__
            try:
              klass._reserved_uid_lock.acquire()
1356 1357
              uid_buffer = self.getUIDBuffer()
              if uid_buffer is not None:
1358 1359 1360 1361 1362 1363 1364 1365
                # This is the case where:
                #   1. The object got an uid.
                #   2. The catalog was cleared.
                #   3. The catalog produced the same reserved uid.
                #   4. The object was reindexed.
                # In this case, the uid is not reserved any longer, but
                # SQLCatalog believes that it is still reserved. So it is
                # necessary to remove the uid from the list explicitly.
1366
                try:
1367
                  uid_buffer.remove(uid)
1368 1369
                except ValueError:
                  pass
1370 1371 1372 1373
            finally:
              klass._reserved_uid_lock.release()
          elif catalog_path is not None:
            # An uid conflict happened... Why?
1374
            # can be due to path length
1375
            if len(path) > MAX_PATH_LEN:
1376 1377
              LOG('SQLCatalog', WARNING, 'path of object %r is too long for catalog. You should use a shorter path.' %(object,))

1378
            object.uid = self.newUid()
1379
            LOG('SQLCatalog', WARNING,
1380
                'uid of %r changed from %r to %r as old one is assigned to %s in catalog !!! This can be fatal. You should reindex the whole site immediately.' % (object, uid, object.uid, catalog_path))
1381 1382 1383
            uid = object.uid

        assigned_uid_dict[uid] = object
1384

1385 1386
    if method_id_list is None:
      method_id_list = self.sql_catalog_object_list
1387
    econtext_cache = {}
1388
    expression_result_cache = {}
1389 1390
    argument_cache = {}

1391
    try:
1392
      if not disable_cache:
1393
        enableReadOnlyTransactionCache(self)
1394

1395
      for method_name in method_id_list:
1396
        kw = {}
1397
        if self.isMethodFiltered(method_name):
1398
          catalogged_object_list = []
1399
          filter = self.filter_dict[method_name]
1400
          type_set = frozenset(filter['type']) or None
1401 1402
          expression = filter['expression_instance']
          expression_cache_key_list = filter.get('expression_cache_key', '').split()
1403 1404 1405 1406
          for object in object_list:
            # We will check if there is an filter on this
            # method, if so we may not call this zsqlMethod
            # for this object
1407
            if type_set is not None and object.getPortalType() not in type_set:
1408 1409
              continue
            elif expression is not None:
1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
              if expression_cache_key_list:
                # We try to save results of expressions by portal_type
                # or by anyother key which can prevent us from evaluating
                # expressions. This cache is built each time we reindex
                # objects but we could also use over multiple transactions
                # if this can improve performance significantly
                try:
                  cache_key = map(lambda key: object.getProperty(key, None), expression_cache_key_list)
                    # ZZZ - we could find a way to compute this once only
                  cache_key = (method_name, tuple(cache_key))
                  result = expression_result_cache[cache_key]
                  compute_result = 0
                except KeyError:
                  cache_result = 1
                  compute_result = 1
              else:
                cache_result = 0
                compute_result = 1
              if compute_result:
                try:
                  econtext = econtext_cache[object.uid]
                except KeyError:
                  econtext = self.getExpressionContext(object)
                  econtext_cache[object.uid] = econtext
                result = expression(econtext)
              if cache_result:
                expression_result_cache[cache_key] = result
1437 1438 1439 1440 1441
              if not result:
                continue
            catalogged_object_list.append(object)
        else:
          catalogged_object_list = object_list
1442

1443 1444
        if len(catalogged_object_list) == 0:
          continue
1445

1446 1447
        #LOG('catalogObjectList', 0, 'method_name = %s' % (method_name,))
        method = getattr(self, method_name)
1448
        if method.meta_type in ("Z SQL Method", "LDIF Method"):
1449
          # Build the dictionnary of values
1450
          arguments = method.arguments_src.split()
1451 1452 1453
        elif method.meta_type == "Script (Python)":
          arguments = \
            method.func_code.co_varnames[:method.func_code.co_argcount]
1454 1455 1456 1457 1458 1459 1460 1461 1462
        else:
          arguments = []
        for arg in arguments:
          value_list = []
          append = value_list.append
          for object in catalogged_object_list:
            try:
              value = argument_cache[(object.uid, arg)]
            except KeyError:
1463
              try:
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
                value = getattr(object, arg, None)
                if callable(value):
                  value = value()
              except ConflictError:
                raise
              except:
                value = None
              if not disable_cache:
                argument_cache[(object.uid, arg)] = value
            append(value)
          kw[arg] = value_list
1475

1476 1477 1478
        # Alter/Create row
        try:
          #start_time = DateTime()
1479
          #LOG('catalogObjectList', DEBUG, 'kw = %r, method_name = %r' % (kw, method_name))
1480 1481 1482 1483 1484 1485 1486
          method(**kw)
          #end_time = DateTime()
          #if method_name not in profile_dict:
          #  profile_dict[method_name] = end_time.timeTime() - start_time.timeTime()
          #else:
          #  profile_dict[method_name] += end_time.timeTime() - start_time.timeTime()
          #LOG('catalogObjectList', 0, '%s: %f seconds' % (method_name, profile_dict[method_name]))
1487

1488 1489 1490
        except ConflictError:
          raise
        except: