Document.py 54.6 KB
Newer Older
1
# -*- coding: utf-8 -*-
Jean-Paul Smets's avatar
Jean-Paul Smets committed
2 3 4
##############################################################################
#
# Copyright (c) 2002 Nexedi SARL and Contributors. All Rights Reserved.
5
#                    Jean-Paul Smets-Solanes <jp@nexedi.com>
Jean-Paul Smets's avatar
Jean-Paul Smets committed
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
#
# WARNING: This program as such is intended to be used by professional
# programmers who take the whole responsability of assessing all potential
# consequences resulting from its eventual inadequacies and bugs
# End users who are looking for a ready-to-use solution with commercial
# garantees and support are strongly adviced to contract a Free Software
# Service Company
#
# This program is Free Software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
#
##############################################################################

30
import re, socket, sys
Bartek Górny's avatar
Bartek Górny committed
31 32
from DateTime import DateTime
from operator import add
33
from xmlrpclib import Fault
34
from zLOG import LOG, INFO
35
from AccessControl import ClassSecurityInfo, getSecurityManager
36
from AccessControl.SecurityManagement import newSecurityManager, setSecurityManager
Romain Courteaud's avatar
Romain Courteaud committed
37 38
from Acquisition import aq_base
from Globals import PersistentMapping
39
from Globals import get_request
40
from Products.CMFCore.utils import getToolByName, _checkPermission
41
from Products.ERP5Type import Permissions, PropertySheet, Constraint, interfaces
Jean-Paul Smets's avatar
Jean-Paul Smets committed
42
from Products.ERP5Type.XMLObject import XMLObject
43
from Products.ERP5Type.DateUtils import convertDateToHour, number_of_hours_in_day, number_of_hours_in_year
44 45
from Products.ERP5Type.Utils import convertToUpperCase
from Products.ERP5Type.Base import WorkflowMethod
46
from Products.ERP5Type.TransactionalVariable import getTransactionalVariable
47
from Products.ERP5Type.ExtensibleTraversable import ExtensibleTraversableMixIn
48
from Products.ERP5Type.Cache import getReadOnlyTransactionCache, DEFAULT_CACHE_SCOPE
49
from Products.ERP5.Document.Url import UrlMixIn
50
from Products.ERP5.Tool.ContributionTool import MAX_REPEAT
51
from Products.ERP5Type.UnrestrictedMethod import UnrestrictedMethod
52
from Products.ZSQLCatalog.SQLCatalog import SQLQuery
53
from AccessControl import Unauthorized
54
import zope.interface
55
import cStringIO
56
import string
57
from OFS.Image import Pdata
58

Bartek Górny's avatar
Bartek Górny committed
59
_MARKER = []
60
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
61 62
# these property ids are unchangable
FIXED_PROPERTY_IDS =  ('id', 'uid', 'rid', 'sid') 
Jean-Paul Smets's avatar
Jean-Paul Smets committed
63

64 65 66 67 68
def makeSortedTuple(kw):
  items = kw.items()
  items.sort()
  return tuple(items)

69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109
class SnapshotMixin:
  """
    This class provides a generic API to store in the ZODB
    PDF snapshots of objects and documents with the
    goal to keep a facsimile copy of documents as they
    were at a given date.
  """

  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

  security.declareProtected(Permissions.ModifyPortalContent, 'createSnapshot')
  def createSnapshot(self):
    """
      Create a snapshot (PDF). This is the normal way to modifiy
      snapshot_data. Once a snapshot is taken, a new snapshot
      can not be taken.

      NOTE: use getSnapshotData and hasSnapshotData accessors
      to access a snapshot.

      NOTE2: implementation of createSnapshot should probably
      be delegated to a types base method since this it
      is configuration dependent.
    """
    if self.hasSnapshotData():
      raise ConversionError('This document already has a snapshot.')
    self._setSnapshotData(self.convert(format='pdf'))

  security.declareProtected(Permissions.ManagePortal, 'deleteSnapshot')
  def deleteSnapshot(self):
    """
      Deletes the snapshot - in theory this should never be done.
      It is there for programmers and system administrators.
    """
    try:
      del(self.snapshot_data)
    except AttributeError:
      pass

110
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
111

112 113
class DocumentProxyError(Exception):pass

114 115
class NotConvertedError(Exception):pass

116 117 118 119 120
class ConversionCacheMixin:
  """
    This class provides a generic API to store in the ZODB
    various converted versions of a file or of a string.

Bartek Górny's avatar
Bartek Górny committed
121 122 123
    Versions are stored in dictionaries; the class stores also
    generation time of every format and its mime-type string.
    Format can be a string or a tuple (e.g. format, resolution).
124 125 126 127 128 129
  """

  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

130 131 132
  def _getCacheFactory(self):
    """
    """
133 134
    if self.isTempObject():
      return
135 136
    cache_tool = getToolByName(self, 'portal_caches')
    preference_tool = getToolByName(self, 'portal_preferences')
137
    cache_factory_name = preference_tool.getPreferredConversionCacheFactory('document_cache_factory')
138
    cache_factory = cache_tool.getRamCacheRoot().get(cache_factory_name)
139 140 141
    #XXX This conditional statement should be remove as soon as
    #Broadcasting will be enable among all zeo clients.
    #Interaction which update portal_caches should interact with all nodes.
142 143 144
    if cache_factory is None and getattr(cache_tool, cache_factory_name, None) is not None:
      #ram_cache_root is not up to date for current node
      cache_tool.updateCache()
145 146
    return cache_tool.getRamCacheRoot().get(cache_factory_name)

147 148 149 150
  security.declareProtected(Permissions.ModifyPortalContent, 'clearConversionCache')
  def clearConversionCache(self):
    """
    """
151 152 153
    if self.isTempObject():
      self.temp_conversion_data = {}
      return
154 155
    for cache_plugin in self._getCacheFactory().getCachePluginList():
      cache_plugin.delete(self.getPath(), DEFAULT_CACHE_SCOPE)
Romain Courteaud's avatar
Romain Courteaud committed
156

157 158
  security.declareProtected(Permissions.View, 'hasConversion')
  def hasConversion(self, **kw):
159
    """
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
    If you want to get conversion cache value if exists, please write
    the code like:

      try:
        mime, data = getConversion(**kw)
      except KeyError:
        ...

    instead of:

      if self.hasConversion(**kw):
        mime, data = self.getConversion(**kw)
      else:
        ...

    for better performance.
176
    """
177 178 179 180 181
    try:
      self.getConversion(**kw)
      return True
    except KeyError:
      return False
182

Bartek Górny's avatar
Bartek Górny committed
183
  security.declareProtected(Permissions.ModifyPortalContent, 'setConversion')
184
  def setConversion(self, data, mime=None, calculation_time=None, **kw):
Bartek Górny's avatar
Bartek Górny committed
185 186
    """
    """
187
    cache_id = self.generateCacheId(**kw)
188
    if self.isTempObject():
189 190
      if getattr(aq_base(self), 'temp_conversion_data', None) is None:
        self.temp_conversion_data = {}
191
      self.temp_conversion_data[cache_id] = (mime, aq_base(data))
192
      return
193 194
    cache_factory = self._getCacheFactory()
    cache_duration = cache_factory.cache_duration
195
    if data is not None:
196
      for cache_plugin in cache_factory.getCachePluginList():
197
        try:
198 199
          cache_entry = cache_plugin.get(self.getPath(), DEFAULT_CACHE_SCOPE)
          cache_dict = cache_entry.getValue()
200
        except KeyError:
201
          cache_dict = {}
202 203 204 205
        cache_dict.update({cache_id: (mime, aq_base(data))})
        cache_plugin.set(self.getPath(), DEFAULT_CACHE_SCOPE,
                         cache_dict, calculation_time=calculation_time,
                         cache_duration=cache_duration)
206

Bartek Górny's avatar
Bartek Górny committed
207
  security.declareProtected(Permissions.View, 'getConversion')
208
  def getConversion(self, **kw):
Bartek Górny's avatar
Bartek Górny committed
209 210
    """
    """
211
    cache_id = self.generateCacheId(**kw)
212
    if self.isTempObject():
213
      return getattr(aq_base(self), 'temp_conversion_data', {})[cache_id]
214 215 216 217 218 219
    for cache_plugin in self._getCacheFactory().getCachePluginList():
      cache_entry = cache_plugin.get(self.getPath(), DEFAULT_CACHE_SCOPE)
      data = cache_entry.getValue().get(cache_id)
      if data:
        return data
    raise KeyError, 'Conversion cache key does not exists for %r' % cache_id
Jean-Paul Smets's avatar
Jean-Paul Smets committed
220 221

  security.declareProtected(Permissions.View, 'getConversionSize')
222
  def getConversionSize(self, **kw):
223
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
224
    """
225
    try:
226
      return len(self.getConversion(**kw))
227 228
    except KeyError:
      return 0
229

230 231 232 233 234 235
  def generateCacheId(self, **kw):
    """Generate proper cache id based on **kw.
    Function inspired from ERP5Type.Cache
    """
    return str(makeSortedTuple(kw)).translate(string.maketrans('', ''), """[]()<>'", """)

236 237 238 239 240 241 242
class PermanentURLMixIn(ExtensibleTraversableMixIn):
  """
    Provides access to documents through their permanent URL.
    This class must be inherited by all document classes so
    that documents displayed outside a Web Site / Web Section
    can also use the permanent URL principle.
  """
Bartek Górny's avatar
Bartek Górny committed
243

244 245 246
  # Declarative security
  security = ClassSecurityInfo()

247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
  ### Extensible content
  def _getExtensibleContent(self, request, name):
    # Permanent URL traversal
    # First we must get the logged user by forcing identification
    cache = getReadOnlyTransactionCache(self)
    # LOG('getReadOnlyTransactionCache', 0, repr(cache)) # Currently, it is always None
    if cache is not None:
      key = ('__bobo_traverse__', self, 'user')
      try:
        user = cache[key]
      except KeyError:
        user = _MARKER
    else:
      user = _MARKER
    old_user = getSecurityManager().getUser()
    if user is _MARKER:
      user = None # By default, do nothing
      if old_user is None or old_user.getUserName() == 'Anonymous User':
265
        user_folder = getattr(self.getPortalObject(), 'acl_users', None)
266 267 268 269
        if user_folder is not None:
          try:
            if request.get('PUBLISHED', _MARKER) is _MARKER:
              # request['PUBLISHED'] is required by validate
270
              request['PUBLISHED'] = self
271 272 273
              has_published = False
            else:
              has_published = True
274 275 276 277 278 279 280
            try:
              user = user_folder.validate(request)
            except AttributeError:
              # This kind of error happens with unrestrictedTraverse,
              # because the request object is a fake, and it is just
              # a dict object.
              user = None
281
            if not has_published:
282 283 284 285 286 287
              try:
                del request.other['PUBLISHED']
              except AttributeError:
                # The same here as above. unrestrictedTraverse provides
                # just a plain dict, so request.other does not exist.
                del request['PUBLISHED']
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
          except:
            LOG("ERP5 WARNING",0,
                "Failed to retrieve user in __bobo_traverse__ of WebSection %s" % self.getPath(),
                error=sys.exc_info())
            user = None
      if user is not None and user.getUserName() == 'Anonymous User':
        user = None # If the user which is connected is anonymous,
                    # do not try to change SecurityManager
      if cache is not None:
        cache[key] = user
    if user is not None:
      # We need to perform identification
      old_manager = getSecurityManager()
      newSecurityManager(get_request(), user)
    # Next get the document per name
    portal = self.getPortalObject()
    document = self.getDocumentValue(name=name, portal=portal)
    # Last, cleanup everything
    if user is not None:
      setSecurityManager(old_manager)
    if document is not None:
      document = aq_base(document.asContext(id=name, # Hide some properties to permit locating the original
                                            original_container=document.getParentValue(),
                                            original_id=document.getId(),
                                            editable_absolute_url=document.absolute_url()))
      return document.__of__(self)
314

315 316 317
    # no document found for current user, still such document may exists
    # in some cases user (like Anonymous) can not view document according to portal catalog
    # but we may ask him to login if such a document exists
318 319 320 321
    isAuthorizationForced = getattr(self, 'isAuthorizationForced', None)
    if isAuthorizationForced is not None and isAuthorizationForced():
      getDocumentValue = UnrestrictedMethod(self.getDocumentValue)
      if getDocumentValue(name=name, portal=portal) is not None:
322 323
        # force user to login as specified in Web Section
        raise Unauthorized
324

325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355
  security.declareProtected(Permissions.View, 'getDocumentValue')
  def getDocumentValue(self, name=None, portal=None, **kw):
    """
      Return the default document with the given
      name. The name parameter may represent anything
      such as a document reference, an identifier,
      etc.

      If name is not provided, the method defaults
      to returning the default document by calling
      getDefaultDocumentValue.

      kw parameters can be useful to filter content
      (ex. force a given validation state)

      This method must be implemented through a
      portal type dependent script:
        WebSection_getDocumentValue
    """
    if name is None:
      return self.getDefaultDocumentValue()

    cache = getReadOnlyTransactionCache(self)
    method = None
    if cache is not None:
      key = ('getDocumentValue', self)
      try:
        method = cache[key]
      except KeyError:
        pass

356 357 358
    if method is None:
      method = self._getTypeBasedMethod('getDocumentValue', 
              fallback_script_id='WebSection_getDocumentValue')
359 360

    if cache is not None:
361 362 363 364 365 366 367
      if cache.get(key, _MARKER) is _MARKER:
        cache[key] = method

    document = method(name, portal=portal, **kw)
    if document is not None:
      document = document.__of__(self)
    return document
368

369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
class DocumentProxyMixin:
  """
  Provides access to documents referenced by the follow_up field
  """
  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

  security.declareProtected(Permissions.AccessContentsInformation,
                            'index_html' )
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
    """ Only a proxy method """
    self.getProxiedDocument().index_html(REQUEST, RESPONSE, format, **kw)

  security.declareProtected(Permissions.AccessContentsInformation,
                            'getProxiedDocument' )
  def getProxiedDocument(self):
    """
    Try to retrieve the original document
    """
    proxied_document = self.getDocumentProxyValue()
    if proxied_document is None:
391
      raise DocumentProxyError("Unable to find a proxied document")
392 393
    return proxied_document

394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
class UpdateMixIn:
  """
    Provides an API to compute a date index based on the update
    frequency of the document.
  """

  # Declarative security
  security = ClassSecurityInfo()

  security.declareProtected(Permissions.AccessContentsInformation, 'getFrequencyIndex')
  def getFrequencyIndex(self):
    """
      Returns the document update frequency as an integer
      which is used by alamrs to decide which documents
      must be updates at which time. The index represents
      a time slot (ex. all days in a month, all hours in a week).
    """
    try:
      return self.getUpdateFrequencyValue().getIntIndex()
    except AttributeError:
      # Catch Attribute error or Key error - XXX not beautiful
      return 0

  security.declareProtected(Permissions.AccessContentsInformation, 'getCreationDateIndex')
  def getCreationDateIndex(self, at_date = None):
    """
    Returns the document Creation Date Index which is the creation 
    date converted into hours modulo the Frequency Index.
    """
    frequency_index = self.getFrequencyIndex()
    if not frequency_index: return -1 # If not update frequency is provided, make sure we never update
    hour = convertDateToHour(date=self.getCreationDate())
    creation_date_index = hour % frequency_index
    # in the case of bisextile year, we substract 24 hours from the creation date,
    # otherwise updating documents (frequency=yearly update) created the last
    # 24 hours of bissextile year will be launched once every 4 years.
    if creation_date_index >= number_of_hours_in_year:
      creation_date_index = creation_date_index - number_of_hours_in_day

    return creation_date_index

  security.declareProtected(Permissions.AccessContentsInformation, 'isUpdatable')
  def isUpdatable(self):
    """
      This method is used to decide which document can be updated
      in the crawling process. This can depend for example on
      workflow states (publication state,
      validation state) or on roles on the document.
    """
    method = self._getTypeBasedMethod('isUpdatable', 
        fallback_script_id = 'Document_isUpdatable')
    return method()


class Document(PermanentURLMixIn, XMLObject, UrlMixIn, ConversionCacheMixin, SnapshotMixin, UpdateMixIn):
Bartek Górny's avatar
Bartek Górny committed
449 450 451 452 453 454 455 456 457
  """
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
  isPortalContent = 1
  isRADContent = 1
  isDocument = 1
458
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
459

460 461
  zope.interface.implements( interfaces.IDocument, )

Jean-Paul Smets's avatar
Jean-Paul Smets committed
462 463
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
464 465
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
466
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
467
  charset_parser = re.compile('charset="?([a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
468

Bartek Górny's avatar
Bartek Górny committed
469 470 471 472 473 474 475 476 477 478 479
  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

  # Declarative properties
  property_sheets = ( PropertySheet.Base
                    , PropertySheet.XMLObject
                    , PropertySheet.CategoryCore
                    , PropertySheet.DublinCore
                    , PropertySheet.Version
                    , PropertySheet.Document
Jean-Paul Smets's avatar
Jean-Paul Smets committed
480 481
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
482
                    , PropertySheet.Periodicity
483
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
484 485
                    )

486 487 488
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
                              'subject', 'source_reference', 'source_project_title',)
Bartek Górny's avatar
Bartek Górny committed
489

490
  ### Content processing methods
491
  security.declareProtected(Permissions.View, 'index_html')
492
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
493 494
    """
      We follow here the standard Zope API for files and images
Jean-Paul Smets's avatar
Jean-Paul Smets committed
495 496 497 498 499 500 501 502 503 504 505
      and extend it to support format conversion. The idea
      is that an image which ID is "something.jpg" should
      ne directly accessible through the URL
      /a/b/something.jpg. The same is true for a file and
      for any document type which primary purpose is to
      be used by a helper application rather than displayed
      as HTML in a web browser. Exceptions to this approach
      include Web Pages which are intended to be primarily rendered
      withing the layout of a Web Site or withing a standard ERP5 page.
      Please refer to the index_html of TextDocument.

506 507 508
      Should return appropriate format (calling convert
      if necessary) and set headers.

509
      format -- the format specied in the form of an extension
510
      string (ex. jpeg, html, text, txt, etc.)
511 512

      **kw -- can be various things - e.g. resolution
513

Jean-Paul Smets's avatar
Jean-Paul Smets committed
514 515 516
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
517
    """
518
    raise NotImplementedError
519

Bartek Górny's avatar
Bartek Górny committed
520 521
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
522
    """
523 524 525
      Used by the catalog for basic full text indexing.
      Uses searchable_property_list attribute to put together various properties
      of the document into one searchable text string.
526

527 528
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
529

530
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
531 532
    """
    def getPropertyListOrValue(property):
533
      """
534
        we try to get a list, else we get value and convert to list
535
      """
536 537 538 539 540 541 542
      method = getattr(self, property, None)
      if method is not None:
        if callable(method):
          val = method()
          if isinstance(val, list) or isinstance(val, tuple):
            return list(val)
          return [str(val)]
Bartek Górny's avatar
Bartek Górny committed
543 544 545 546
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
547
          val = [str(val)]
548 549 550
        else:
          val = []
      else:
551
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
552
      return val
553

554
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
555
                                                self.searchable_property_list))
556
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
557 558 559
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
560
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
561

562 563 564 565 566 567 568
  security.declareProtected(Permissions.AccessContentsInformation, 'isExternalDocument')
  def isExternalDocument(self):
    """
    Return true if this document was obtained from an external source
    """
    return bool(self.getUrlString())

Bartek Górny's avatar
Bartek Górny committed
569
  ### Relation getters
570
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
571
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
572
    """
573 574 575 576 577
      This method returns a list of dictionaries which can
      be used to find objects by reference. It uses for
      that a regular expression defined at system level
      preferences.
    """
578
    text = self.getSearchableText() # XXX getSearchableText or asText ?
579
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
580
    try:
581
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
582
    except TypeError: # no regexp in preference
583 584 585
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
586 587 588 589 590 591 592 593 594 595 596 597
    result = []
    tmp = {}
    for match in rx_search.finditer(text):
      group = match.group()
      group_item_list = match.groupdict().items()
      group_item_list.sort()
      key = (group, tuple(group_item_list))
      tmp[key] = None
    for group, group_item_tuple in tmp.keys():
      result.append((group, dict(group_item_tuple)))
    return result

598
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
599 600
  def getImplicitSuccessorValueList(self):
    """
601 602 603
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
604

605
      The implementation goes in 2 steps:
606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622

    - Step 1: extract with a regular expression
      a list of distionaries with various parameters such as 
      reference, portal_type, language, version, user, etc. This
      part is configured through a portal preference.

    - Step 2: read the list of dictionaries
      and build a list of values by calling portal_catalog
      with appropriate parameters (and if possible build 
      a complex query whenever this becomes available in
      portal catalog)
      
      The script is reponsible for calling getSearchableReferenceList
      so that it can use another approach if needed.
      
      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
623
    """
624
    tv = getTransactionalVariable(self) # XXX Performance improvement required
625 626 627 628 629 630
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

631
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
632
    result = self.Base_getImplicitSuccessorValueList(reference_list)
633 634
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
635

636
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
637 638 639
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
640
      by reference/language etc. Implementation is passed to 
641
        Base_getImplicitPredecessorValueList
642 643 644 645 646 647 648 649 650 651 652 653 654

      The script should proceed in two steps:

      Step 1: build a list of references out of the context
      (ex. INV-123456, 123456, etc.)

      Step 2: search using the portal_catalog and use
      priorities (ex. INV-123456 before 123456)
      ( if possible build  a complex query whenever 
      this becomes available in portal catalog )

      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
655
    """
656
    tv = getTransactionalVariable(self) # XXX Performance improvement required
657 658 659 660 661 662
    cache_key = ('getImplicitPredecessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

663 664
    method = self._getTypeBasedMethod('getImplicitPredecessorValueList', 
        fallback_script_id = 'Base_getImplicitPredecessorValueList')
665
    result = method()
666 667
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
668

669
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
670 671 672
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
673
      are similar. Not implemented yet.
Bartek Górny's avatar
Bartek Górny committed
674 675 676 677 678

      No cloud needed because transitive process
    """
    return []

679
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
680
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
681 682 683 684 685 686 687 688
    """
      Returns all documents which are similar to us, directly or indirectly, and
      in both directions. In other words, it is a transitive closure of similar 
      relation. Every document is returned in the latest version available.
    """
    lista = {}
    depth = int(depth)

689 690
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
691

692
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
693
      level += 1
694 695 696 697
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
698 699 700
      for r in res:
        if lista.get(r) is None:
          lista[r] = True # we use dict keys to ensure uniqueness
Bartek Górny's avatar
Bartek Górny committed
701 702
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
703

704
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
705 706 707
    lista_latest = {}
    for o in lista.keys():
      lista_latest[o.getLatestVersionValue()] = True # get latest versions avoiding duplicates again
Bartek Górny's avatar
Bartek Górny committed
708 709
    if lista_latest.has_key(self): 
      lista_latest.pop(self) # remove this document
710 711 712
    if lista_latest.has_key(self.getLatestVersionValue()):
      # remove last version of document itself from related documents
      lista_latest.pop(self.getLatestVersionValue()) 
Bartek Górny's avatar
Bartek Górny committed
713 714 715

    return lista_latest.keys()

716
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
717 718 719
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
720 721
      Tries to find the latest version with the latest revision
      of self which the current user is allowed to access.
Bartek Górny's avatar
Bartek Górny committed
722

723 724
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
725

726 727 728
      If language is not provided, return the latest version
      in original language or in the user language if the version is
      the same.
Bartek Górny's avatar
Bartek Górny committed
729
    """
Bartek Górny's avatar
Bartek Górny committed
730 731
    if not self.getReference():
      return self
732
    catalog = getToolByName(self, 'portal_catalog', None)
733
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
734 735 736 737 738 739
    if language is not None: kw['language'] = language
    res = catalog(**kw)

    original_language = self.getOriginalLanguage()
    user_language = self.Localizer.get_selected_language()

740
    # if language was given return it (if there are any docs in this language)
741
    if language is not None:
742 743 744 745
      try:
        return res[0].getObject()
      except IndexError:
        return None
746 747 748 749 750 751 752
    else:
      first = res[0]
      in_original = None
      for ob in res:
        if ob.getVersion() != first.getVersion():
          # we are out of the latest version - return in_original or first
          if in_original is not None:
Bartek Górny's avatar
Bartek Górny committed
753
            return in_original.getObject()
754
          else:
Bartek Górny's avatar
Bartek Górny committed
755
            return first.getObject() # this shouldn't happen in real life
756 757
        if ob.getLanguage() == user_language:
          # we found it in the user language
Bartek Górny's avatar
Bartek Górny committed
758
          return ob.getObject()
759 760 761
        if ob.getLanguage() == original_language:
          # this is in original language
          in_original = ob
762 763
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
764 765 766 767 768 769 770

  security.declareProtected(Permissions.View, 'getVersionValueList')
  def getVersionValueList(self, version=None, language=None):
    """
      Returns a list of documents with same reference, same portal_type
      but different version and given language or any language if not given.
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
771
    catalog = getToolByName(self, 'portal_catalog', None)
772
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
773
                   reference=self.getReference(),
774
                   sort_on=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
775
                  )
776 777 778
    if version: kw['version'] = version
    if language: kw['language'] = language
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
779

780
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
781 782
  def isVersionUnique(self):
    """
783 784 785
      Returns true if no other document exists with the same
      reference, version and language, or if the current
      document has no reference.
Bartek Górny's avatar
Bartek Górny committed
786
    """
787 788
    if not self.getReference():
      return 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
789
    catalog = getToolByName(self, 'portal_catalog', None)
790
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
791 792 793
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
794
                                            uid=self.getUid(),
795
                                            validation_state="!=cancelled"
796 797 798 799 800
                                            )[0][0]
    count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
Ivan Tyagov's avatar
Ivan Tyagov committed
801
                                            validation_state="!=cancelled"
802 803 804 805
                                            )[0][0]
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

806 807 808 809
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
810 811 812
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
813
    """
814 815
    getInfoFor = getToolByName(self, 'portal_workflow').getInfoFor
    revision = len(getInfoFor(self, 'history', (), 'edit_workflow'))
816
    # XXX Also look at processing_status_workflow for compatibility.
817 818
    for history_item in getInfoFor(self, 'history', (),
                                   'processing_status_workflow'):
819 820 821
      if history_item.get('action') == 'edit':
        revision += 1
    return str(revision)
822 823 824 825 826 827 828

  security.declareProtected(Permissions.AccessContentsInformation, 'getRevisionList')
  def getRevisionList(self):
    """
      Returns the list of revision numbers of the current document
      by by analysing the change log of the current object.
    """
829
    return map(str, range(1, 1 + int(self.getRevision())))
830

831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
  security.declareProtected(Permissions.ModifyPortalContent, 'mergeRevision')
  def mergeRevision(self):
    """
      Merge the current document with any previous revision
      or change its version to make sure it is still unique.

      NOTE: revision support is implemented in the Document
      class rather than within the ContributionTool
      because the ingestion process requires to analyse the content
      of the document first. Hence, it is not possible to
      do any kind of update operation until the whole ingestion
      process is completed, since update requires to know
      reference, version, language, etc. In addition,
      we have chosen to try to merge revisions after each
      metadata discovery as a way to make sure that any
      content added in the system through the ContributionTool
      (ex. through webdav) will be merged if necessary.
      It may be posssible though to split disoverMetadata and
      finishIngestion.
    """
    document = self
    catalog = getToolByName(self, 'portal_catalog', None)
    if self.getReference():
      # Find all document with same (reference, version, language)
      kw = dict(portal_type=self.getPortalDocumentTypeList(),
                reference=self.getReference(),
857
                where_expression=SQLQuery("validation_state NOT IN ('cancelled', 'deleted')"))
858 859 860 861 862 863
      if self.getVersion(): kw['version'] = self.getVersion()
      if self.getLanguage(): kw['language'] = self.getLanguage()
      document_list = catalog.unrestrictedSearchResults(**kw)
      existing_document = None
      # Select the first one which is not self and which
      # shares the same coordinates
864
      document_list = list(document_list)
865
      document_list.sort(key=lambda x: x.getId())
Ivan Tyagov's avatar
Ivan Tyagov committed
866
      #LOG('[DMS] Existing documents for %s' %self.getSourceReference(), INFO, len(document_list))
867 868 869 870 871 872 873 874 875 876
      for o in document_list:
        if o.getRelativeUrl() != self.getRelativeUrl() and\
           o.getVersion() == self.getVersion() and\
           o.getLanguage() == self.getLanguage():
          existing_document = o.getObject()
          break
      # We found an existing document to update
      if existing_document is not None:
        document = existing_document
        if existing_document.getPortalType() != self.getPortalType():
877
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
878 879
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
          self.setUniqueReference(suffix='unauthorized')
880
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
881 882 883
        else:
          update_kw = {}
          for k in self.propertyIds():
884
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
885 886 887 888 889 890
              update_kw[k] = self.getProperty(k)
          existing_document.edit(**update_kw)
          # Erase self
          self.delete() # XXX Do we want to delete by workflow or for real ?
    return document

891
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
892 893 894 895 896
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
897
    if not self.getReference(): return []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
898
    catalog = getToolByName(self, 'portal_catalog', None)
899
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
900
                           reference=self.getReference(),
901 902 903 904
                           group_by=('language',))
    if version is not None:
      kw['version'] = version
    return map(lambda o:o.getLanguage(), catalog(**kw))
Bartek Górny's avatar
Bartek Górny committed
905

906
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
907 908 909
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
910 911

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
912 913 914 915 916
    """
    # Approach 1: use portal_catalog and creation dates
    # Approach 2: use workflow analysis (delegate to script if necessary)
    #             workflow analysis is the only way for multiple orginals
    # XXX - cache or set?
917 918 919 920 921 922 923
    reference = self.getReference()
    if not reference:
      return 
    catalog = getToolByName(self, 'portal_catalog', None)
    res = catalog(reference=self.getReference(), sort_on=(('creation_date','ascending'),))
    # XXX this should be security-unaware - delegate to script with proxy roles
    return res[0].getLanguage() # XXX what happens if it is empty?
Bartek Górny's avatar
Bartek Górny committed
924 925 926 927 928 929

  ### Property getters
  # Property Getters are document dependent so that we can
  # handle the weird cases in which needed properties change with the type of document
  # and the usual cases in which accessing content changes with the meta type
  security.declareProtected(Permissions.ModifyPortalContent,'getPropertyDictFromUserLogin')
930
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
931 932 933 934 935
    """
      Based on the user_login, find out as many properties as needed.
      returns properties which should be set on the document
    """
    if user_login is None:
936 937
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
938
        fallback_script_id='Document_getPropertyDictFromUserLogin')
939
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
940 941 942 943 944 945 946

  security.declareProtected(Permissions.ModifyPortalContent,'getPropertyDictFromContent')
  def getPropertyDictFromContent(self):
    """
      Based on the document content, find out as many properties as needed.
      returns properties which should be set on the document
    """
947
    if not self.hasBaseData():
948
      raise NotConvertedError
949
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
950
        fallback_script_id='Document_getPropertyDictFromContent')
951
    return method()
Bartek Górny's avatar
Bartek Górny committed
952 953 954 955 956 957 958 959 960 961 962 963 964 965

  security.declareProtected(Permissions.ModifyPortalContent,'getPropertyDictFromFileName')
  def getPropertyDictFromFileName(self, file_name):
    """
      Based on the file name, find out as many properties as needed.
      returns properties which should be set on the document
    """
    return self.portal_contributions.getPropertyDictFromFileName(file_name)

  security.declareProtected(Permissions.ModifyPortalContent,'getPropertyDictFromInput')
  def getPropertyDictFromInput(self):
    """
      Get properties which were supplied explicitly to the ingestion method
      (discovered or supplied before the document was created).
966 967 968 969 970

      The implementation consists in saving document properties
      into _backup_input by supposing that original input parameters were
      set on the document by ContributionTool.newContent as soon
      as the document was created.
Bartek Górny's avatar
Bartek Górny committed
971 972 973 974 975 976
    """
    if hasattr(self, '_backup_input'):
      return getattr(self, '_backup_input')
    kw = {}
    for id in self.propertyIds():
      # We should not consider file data
977 978
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
979
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
980 981 982 983 984 985
        kw[id] = self.getProperty(id)
    self._backup_input = kw # We could use volatile and pass kw in activate
                            # if we are garanteed that _backup_input does not
                            # disappear within a given transaction
    return kw

986 987 988 989 990 991
  security.declareProtected(Permissions.AccessContentsInformation, 'getStandardFileName')
  def getStandardFileName(self):
    """
    Returns the document coordinates as a standard file name. This
    method is the reverse of getPropertyDictFromFileName.
    """
992 993 994
    method = self._getTypeBasedMethod('getStandardFileName', 
        fallback_script_id = 'Document_getStandardFileName')
    return method()
995

Bartek Górny's avatar
Bartek Górny committed
996 997
  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
998
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
999
    """
1000 1001
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
1002 1003 1004
      delegated to scripts or uses preference-configurable regexps. The
      method returns either self or the document which has been
      merged in the discovery process.
Bartek Górny's avatar
Bartek Górny committed
1005

1006
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
1007

1008
      user_login - this is a login string of a person; can be None if the user is
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1009
                   currently logged in, then we'll get him from session
1010
    """
Bartek Górny's avatar
Bartek Górny committed
1011 1012 1013
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList', 
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
1014
    order_list = list(method())
1015
    order_list.reverse()
1016
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
1017
    kw = {}
1018
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1019
      result = None
Bartek Górny's avatar
Bartek Górny committed
1020 1021
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
1022
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
1023 1024 1025
      method_id = 'getPropertyDictFrom%s' % convertToUpperCase(order_id)
      method = getattr(self, method_id)
      if order_id == 'file_name':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1026 1027
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
1028
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1029 1030
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
1031 1032
      else:
        result = method()
1033 1034
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1035

1036 1037 1038
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
1039
    # Prepare the content edit parameters - portal_type should not be changed
1040 1041 1042 1043
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
1044
    self.finishIngestion()
1045
    self.reindexObject()
1046 1047
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
1048 1049
    merged_doc = self.mergeRevision()
    merged_doc.reindexObject()
1050
    return merged_doc
Bartek Górny's avatar
Bartek Górny committed
1051 1052 1053 1054

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1055 1056 1057
      Finish the ingestion process by calling the appropriate script. This
      script can for example allocate a reference number automatically if
      no reference was defined.
Bartek Górny's avatar
Bartek Górny committed
1058
    """
1059 1060
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1061

1062 1063 1064 1065 1066 1067 1068 1069 1070 1071
  # Conversion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'convert')
  def convert(self, format, **kw):
    """
      Main content conversion function, returns result which should
      be returned and stored in cache.
      format - the format specied in the form of an extension
      string (ex. jpeg, html, text, txt, etc.)
      **kw can be various things - e.g. resolution

1072 1073 1074
      Default implementation returns an empty string (html, text)
      or raises an error.

1075 1076 1077 1078
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
    """
1079
    if format == 'html':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1080
      return 'text/html', '' # XXX - Why ?
1081
    if format in ('text', 'txt'):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1082
      return 'text/plain', '' # XXX - Why ?
1083
    raise NotImplementedError
1084

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1085 1086 1087 1088 1089 1090 1091
  def getConvertedSize(self, format):
    """
      Returns the size of the converted document
    """
    format, data = self.convert(format)
    return len(data)

1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
  security.declareProtected(Permissions.View, 'asSubjectText')
  def asSubjectText(self, **kw):
    """
      Converts the subject of the document to a textual representation.
    """
    subject = self.getSubject()
    if not subject:
      # XXX not sure if this fallback is a good idea.
      subject = self.getTitle()
    if subject is None:
      subject = ''
    return str(subject)

1105
  security.declareProtected(Permissions.View, 'asText')
1106
  def asText(self, **kw):
1107 1108 1109
    """
      Converts the content of the document to a textual representation.
    """
1110 1111
    kw['format'] = 'txt'
    mime, data = self.convert(**kw)
1112
    return str(data)
1113

1114
  security.declareProtected(Permissions.View, 'asEntireHTML')
1115
  def asEntireHTML(self, **kw):
1116 1117
    """
      Returns a complete HTML representation of the document
1118 1119 1120
      (with body tags, etc.). Adds if necessary a base
      tag so that the document can be displayed in an iframe
      or standalone.
1121 1122

      Actual conversion is delegated to _asHTML
1123
    """
1124
    html = self._asHTML(**kw)
1125 1126 1127 1128 1129 1130 1131
    if self.getUrlString():
      # If a URL is defined, add the base tag
      # if base is defined yet.
      html = str(html)
      if not html.find('<base') >= 0:
        base = '<base href="%s">' % self.getContentBaseURL()
        html = html.replace('<head>', '<head>%s' % base)
1132
      self.setConversion(html, mime='text/html', format='base-html')
1133
    return html
1134

1135
  security.declarePrivate('_asHTML')
1136
  def _asHTML(self, **kw):
1137 1138 1139 1140
    """
      A private method which converts to HTML. This method
      is the one to override in subclasses.
    """
1141
    if not self.hasBaseData():
1142
      raise ConversionError('This document has not been processed yet.')
1143
    try:
1144
      # FIXME: no substitution may occur in this case.
1145 1146
      mime, data = self.getConversion(format='base-html')
      return data
1147 1148 1149 1150
    except KeyError:
      kw['format'] = 'html'
      mime, html = self.convert(**kw)
      return html
1151

1152
  security.declareProtected(Permissions.View, 'asStrippedHTML')
1153
  def asStrippedHTML(self, **kw):
1154
    """
1155 1156
      Returns a stripped HTML representation of the document
      (without html and body tags, etc.) which can be used to inline
1157 1158
      a preview of the document.
    """
1159
    if not self.hasBaseData():
1160
      return ''
1161
    try:
1162
      # FIXME: no substitution may occur in this case.
1163 1164
      mime, data = self.getConversion(format='stripped-html')
      return data
1165 1166 1167 1168
    except KeyError:
      kw['format'] = 'html'
      mime, html = self.convert(**kw)
      return self._stripHTML(str(html))
1169

1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
  def _guessEncoding(self, string):
    """
      Try to guess the encoding for this string.
      Returns None if no encoding can be guessed.
    """
    try:
      import chardet
    except ImportError:
      return None
    return chardet.detect(string).get('encoding', None)

1181 1182 1183 1184 1185
  def _stripHTML(self, html, charset=None):
    """
      A private method which can be reused by subclasses
      to strip HTML content
    """
1186 1187
    body_list = re.findall(self.body_parser, str(html))
    if len(body_list):
1188 1189 1190 1191
      stripped_html = body_list[0]
    else:
      stripped_html = html
    # find charset and convert to utf-8
1192 1193
    charset_list = self.charset_parser.findall(str(html)) # XXX - Not efficient if this 
                                         # is datastream instance but hard to do better
1194 1195 1196
    if charset and not charset_list:
      # Use optional parameter is we can not find encoding in HTML
      charset_list = [charset]
1197
    if charset_list and charset_list[0] not in ('utf-8', 'UTF-8'):
1198 1199 1200
      try:
        stripped_html = unicode(str(stripped_html), 
                                charset_list[0]).encode('utf-8')
Nicolas Delaby's avatar
Nicolas Delaby committed
1201
      except (UnicodeDecodeError, LookupError):
1202
        return str(stripped_html)
1203
    return stripped_html
1204

1205

1206 1207 1208 1209 1210 1211 1212 1213 1214
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentInformation')
  def getContentInformation(self):
    """
    Returns the content information from the HTML conversion.
    The default implementation tries to build a dictionnary
    from the HTML conversion of the document and extract
    the document title.
    """
    result = {}
Yusei Tahara's avatar
Yusei Tahara committed
1215
    html = self.asEntireHTML()
1216 1217 1218 1219 1220
    if not html: return result
    title_list = re.findall(self.title_parser, str(html))
    if title_list:
      result['title'] = title_list[0]
    return result
1221 1222

  # Base format support
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1223
  security.declareProtected(Permissions.ModifyPortalContent, 'convertToBaseFormat')
1224
  def convertToBaseFormat(self, **kw):
1225
    """
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
      Converts the content of the document to a base format
      which is later used for all conversions. This method
      is common to all kinds of documents and handles
      exceptions in a unified way.

      Implementation is delegated to _convertToBaseFormat which
      must be overloaded by subclasses of Document which
      need a base format.

      convertToBaseFormat is called upon file upload, document
      ingestion by the processing_status_workflow.

      NOTE: the data of the base format conversion should be stored
      using the base_data property. Refer to Document.py propertysheet.
      Use accessors (getBaseData, setBaseData, hasBaseData, etc.)
1241 1242
    """
    try:
1243
      message = self._convertToBaseFormat() # Call implemetation method
1244
      self.clearConversionCache() # Conversion cache is now invalid
1245 1246 1247 1248 1249 1250 1251
      if message is None:
        # XXX Need to translate.
        message = 'Converted to %s.' % self.getBaseContentType()
      self.convertFile(comment=message) # Invoke workflow method
    except NotImplementedError:
      message = ''
    return message
1252

1253
  def _convertToBaseFormat(self):
1254 1255
    """
    """
1256
    raise NotImplementedError
1257

1258 1259
  security.declareProtected(Permissions.AccessContentsInformation,
                            'isSupportBaseDataConversion')
1260 1261 1262 1263 1264
  def isSupportBaseDataConversion(self):
    """
    """
    return False

1265
  def convertFile(self, **kw):
1266 1267 1268 1269 1270
    """
    Workflow transition invoked when conversion occurs.
    """
  convertFile = WorkflowMethod(convertFile)

1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286
  security.declareProtected(Permissions.AccessContentsInformation,
                            'getMetadataMappingDict')
  def getMetadataMappingDict(self):
    """
    Return a dict of metadata mapping used to update base metadata of the
    document
    """
    try:
      method = self._getTypeBasedMethod('getMetadataMappingDict')
    except KeyError, AttributeError:
      method = None
    if method is not None: 
      return method()
    else:
      return {}

1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
  security.declareProtected(Permissions.ModifyPortalContent, 'updateBaseMetadata')
  def updateBaseMetadata(self, **kw):
    """
    Update the base format data with the latest properties entered
    by the user. For example, if title is changed in ERP5 interface,
    the base format file should be updated accordingly.

    Default implementation does nothing. Refer to OOoDocument class
    for an example of implementation.
    """
    pass

1299
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1300
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
  def populateContent(self):
    """
      Populates the Document with subcontent based on the
      document base data.

      This can be used for example to transform the XML
      of an RSS feed into a single piece per news or
      to transform an XML export from a database into
      individual records. Other application: populate
      an HTML text document with its images, used in
      conversion with convertToBaseFormat.
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1313 1314 1315 1316 1317
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1318 1319

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1320
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1321 1322 1323
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1324 1325 1326 1327 1328 1329
      Default implementation consists in analysing the document
      converted to HTML. Subclasses may overload this method
      if necessary. However, it is better to extend the conversion
      methods in order to produce valid HTML, which is useful to
      many people, rather than overload this method which is only
      useful for crawling.
1330
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1331 1332
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1333

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1334
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1335
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1336 1337
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1338 1339
      Implementation is handled by ContributionTool.
    """
1340
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1341

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1342 1343
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1344
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1345 1346 1347 1348
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364
  security.declareProtected(Permissions.View, 'isIndexContent')
  def isIndexContent(self, container=None):
    """
      Ask container if we are and index, or a content.
      In the vast majority of cases we are content.
      This method is required in a crawling process to make
      a difference between URLs which return an index (ex. the 
      list of files in remote server which is accessed through HTTP)
      and the files themselves.
    """
    if container is None:
      container = self.getParentValue()
    if hasattr(aq_base(container), 'isIndexContent'):
      return container.isIndexContent(self)
    return False

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1365 1366 1367 1368 1369 1370 1371 1372
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentBaseURL')
  def getContentBaseURL(self):
    """
      Returns the content base URL based on the actual content or
      on its URL.
    """
    base_url = self.asURL()
    base_url_list = base_url.split('/')
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1373
    if len(base_url_list):
1374
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1375
        # Cut the trailing part in http://www.some.site/at/trailing.html
1376
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1377
        base_url = '/'.join(base_url_list[:-1])
1378
    return base_url
1379 1380 1381

  security.declareProtected(Permissions.ModifyPortalContent, '_setBaseData')
  def _setBaseData(self, data):
1382 1383 1384 1385 1386 1387 1388 1389 1390
    """
      XXX - it is really wrong to put this method here since not
      all documents are subclasses of "File". Instead, there should
      be a interface for all classes which can convert their data
      to a base format.
    """
    if not isinstance(data, Pdata) and data is not None:
      file = cStringIO.StringIO(data)
      data, size = self._read_data(file)
1391
    self._baseSetBaseData(data)
1392 1393 1394 1395 1396

  security.declareProtected(Permissions.AccessContentsInformation,
                            'getBaseData')
  def getBaseData(self, default=None):
    """return BaseData as str."""
1397 1398 1399 1400 1401
    base_data = self._baseGetBaseData()
    if base_data is None:
      return None
    else:
      return str(base_data)
1402 1403 1404

  security.declareProtected(Permissions.ModifyPortalContent, '_setData')
  def _setData(self, data):
1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
    """
      XXX - it is really wrong to put this method here since not
      all documents are subclasses of "File". Instead, there should
      be a interface for all classes which can act as a File.
    """
    size = None
    if not isinstance(data, Pdata) and data is not None:
      file = cStringIO.StringIO(data)
      data, size = self._read_data(file)
    if getattr(self, 'update_date', None) is not None:
      self.update_data(data, size=size) # We call this method to make sure size is set and caches reset
    else:
      self._baseSetData(data) # XXX - It would be better to always use this accessor
      self.size=size # Using accessor or caching method would be better
      self.ZCacheable_invalidate()
      self.ZCacheable_set(None)
      self.http__refreshEtag()
1422 1423 1424 1425

  security.declareProtected(Permissions.AccessContentsInformation, 'getData')
  def getData(self, default=None):
    """return Data as str."""
1426 1427 1428 1429 1430
    data = self._baseGetData()
    if data is None:
      return None
    else:
      return str(data)