Document.py 59.1 KB
Newer Older
1
# -*- coding: utf-8 -*-
Jean-Paul Smets's avatar
Jean-Paul Smets committed
2 3
##############################################################################
#
4
# Copyright (c) 2009 Nexedi SA and Contributors. All Rights Reserved.
5
#                    Jean-Paul Smets-Solanes <jp@nexedi.com>
Jean-Paul Smets's avatar
Jean-Paul Smets committed
6 7
#
# WARNING: This program as such is intended to be used by professional
8
# programmers who take the whole responsibility of assessing all potential
Jean-Paul Smets's avatar
Jean-Paul Smets committed
9 10
# consequences resulting from its eventual inadequacies and bugs
# End users who are looking for a ready-to-use solution with commercial
11
# guarantees and support are strongly adviced to contract a Free Software
Jean-Paul Smets's avatar
Jean-Paul Smets committed
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
# 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.
#
##############################################################################

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

57 58 59
# Mixin Import
from Products.ERP5.mixin.cached_convertable import CachedConvertableMixin

Bartek Górny's avatar
Bartek Górny committed
60
_MARKER = []
61
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
62
# these property ids are unchangable
63
FIXED_PROPERTY_IDS =  ('id', 'uid', 'rid', 'sid')
Jean-Paul Smets's avatar
Jean-Paul Smets committed
64

65 66 67 68 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
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

106
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
107

108 109
class DocumentProxyError(Exception):pass

110
class NotConvertedError(Exception):pass
111
allow_class(NotConvertedError)
112

113 114 115 116 117 118 119
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
120

121 122 123
  # Declarative security
  security = ClassSecurityInfo()

124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
  ### 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':
142
        user_folder = getattr(self.getPortalObject(), 'acl_users', None)
143 144 145 146
        if user_folder is not None:
          try:
            if request.get('PUBLISHED', _MARKER) is _MARKER:
              # request['PUBLISHED'] is required by validate
147
              request['PUBLISHED'] = self
148 149 150
              has_published = False
            else:
              has_published = True
151 152 153 154 155 156 157
            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
158
            if not has_published:
159 160 161 162 163 164
              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']
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
          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)
191

192 193 194
    # 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
195 196
    isAuthorizationForced = getattr(self, 'isAuthorizationForced', None)
    if isAuthorizationForced is not None and isAuthorizationForced():
197
      if unrestricted_apply(self.getDocumentValue, (name, portal)) is not None:
198 199
        # force user to login as specified in Web Section
        raise Unauthorized
200

201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222
  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()

223
    method = self._getTypeBasedMethod('getDocumentValue',
224
              fallback_script_id='WebSection_getDocumentValue')
225

226 227
    document = method(name, portal=portal, **kw)
    if document is not None:
228
      return document.__of__(self)
229

230 231 232 233 234 235 236 237 238 239
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' )
240
  def index_html(self, REQUEST, RESPONSE, **kw):
241
    """ Only a proxy method """
242
    return self.getProxiedDocument().index_html(REQUEST, RESPONSE, **kw)
243 244 245 246 247 248 249 250 251

  security.declareProtected(Permissions.AccessContentsInformation,
                            'getProxiedDocument' )
  def getProxiedDocument(self):
    """
    Try to retrieve the original document
    """
    proxied_document = self.getDocumentProxyValue()
    if proxied_document is None:
252
      raise DocumentProxyError("Unable to find a proxied document")
253 254
    return proxied_document

255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
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):
    """
281
    Returns the document Creation Date Index which is the creation
282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
    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.
    """
304
    method = self._getTypeBasedMethod('isUpdatable',
305 306 307 308
        fallback_script_id = 'Document_isUpdatable')
    return method()


309
class Document(PermanentURLMixIn, XMLObject, UrlMixIn, CachedConvertableMixin, SnapshotMixin, UpdateMixIn):
310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
  """Document is an abstract class with all methods related to document
  management in ERP5. This includes searchable text, explicit relations,
  implicit relations, metadata, versions, languages, etc.

  Documents may either store their content directly or cache content
  which is retrieved from a specified URL. The second case if often
  referred as "External Document". Standalone "External Documents" may
  be created by specifying a URL to the contribution tool which is in
  charge of initiating the download process and selecting the appropriate
  document type. Groups of "External Documents" may also be generated from
  so-called "External Source" (refer to ExternalSource class for more
  information).

  External Documents may be downloaded once or updated at regular interval.
  The later can be useful to update the content of an external source.
  Previous versions may be stored in place or kept in a separate file.
  This feature is known as the crawling API. It is mostly implemented
  in ContributionTool with wrappers in the Document class. It can be useful
  for create a small search engine.

  There are currently two types of Document subclasses:

  * File for binary file based documents. File has subclasses such as Image,
    OOoDocument, PDFDocument, etc. to implement specific conversion methods.

  * TextDocument for text based documents. TextDocument has subclasses such
    as Wiki to implement specific methods. 
    TextDocument itself has a subclass (XSLTDocument) which provides
    XSLT based analysis and transformation of XML content based on XSLT
    templates. 

  Conversion should be achieved through the convert method and other methods
  of the conversion API (convertToBaseFormat, etc.).
  Moreover, any Document subclass must ne able to convert documents to text
  (asText method) and HTML (asHTML method). Text is required for full text
  indexing. HTML is required for crawling.

  Instances can be created directly, or via portal_contributions tool which
  manages document ingestion process whereby a file can be uploaded by http
  or sent in by email or dropped in by webdav or in some other way as yet
  unknown. The ingestion process has the following steps:

  (1) portal type detection
  (2) object creation and upload of data
  (3) metadata discovery (optionally with conversion of data to another format)
  (4) other possible actions to finalise the ingestion (ex. by assigning
      a reference)

  This class handles (3) and calls a ZMI script to do (4).

  Metadata can be drawn from various sources:

  input      -   data supplied with http request or set on the object during (2) (e.g.
                 discovered from email text)
  file_name  -   data which might be encoded in file name
  user_login -   information about user who is contributing the file
  content    -   data which might be derived from document content

  If a certain property is defined in more than one source, it is set according to
  preference order returned by a script 
     Document_getPreferredDocumentMetadataDiscoveryOrderList
     (or any type-based version since discovery is type dependent)

  Methods for discovering metadata are:

    getPropertyDictFromInput
    getPropertyDictFromFileName
    getPropertyDictFromUserLogin
    getPropertyDictFromContent

  Methods for processing content are implemented either in Document class
  or in Base class:

    getSearchableReferenceList (Base)
    getSearchableText (Base)
    index_html (overriden in Document subclasses)

  Methods for handling relations are implemented either in Document class
  or in Base class:

    getImplicitSuccessorValueList (Base)
    getImplicitPredecessorValueList (Base)
    getImplicitSimilarValueList (Base)
    getSimilarCloudValueList (Document)

  Implicit relations consist in finding document references inside
  searchable text (ex. INV-23456) and deducting relations from that.
  Two customisable methods required. One to find a list of implicit references
  inside the content (getSearchableReferenceList) and one to convert a given
  document reference into a list of reference strings which could be present
  in other content (asSearchableReferenceList).

  document.getSearchableReferenceList() returns
    [
     {'reference':' INV-12367'},
     {'reference': 'INV-1112', 'version':'012}', 
     {'reference': 'AB-CC-DRK', 'version':'011', 'language': 'en'}
    ]

  The Document class behaviour can be extended / customized through scripts
  (which are type-based so can be adjusted per portal type).

  * Document_getPropertyDictFromUserLogin - finds a user (by user_login or
    from session) and returns properties which should be set on the document

  * Document_getPropertyDictFromContent - analyzes document content and returns
    properties which should be set on the document

  * Base_getImplicitSuccessorValueList - finds appropriate all documents
    referenced in the current content

  * Base_getImplicitPredecessorValueList - finds document predecessors based on
    the document coordinates (can use only complete coordinates, or also partial)

  * Document_getPreferredDocumentMetadataDiscoveryOrderList - returns an order
    in which metadata should be set/overwritten

  * Document_finishIngestion - called by portal_activities after all the ingestion
    is completed (and after document has been converted, so text_content
    is available if the document has it)

  * Document_getNewRevision - calculates revision number which should be set
    on this document. Implementation depends on revision numbering policy which
    can be very different. Interaction workflow should call setNewRevision method.

  * Document_populateContent - analyses the document content and produces
    subcontent based on it (ex. images, news, etc.). This scripts can
    involve for example an XSLT transformation to process XML.

  Subcontent: documents may include subcontent (files, images, etc.)
  so that publication of rich content can be path independent. Subcontent
  can also be used to help the rendering in HTML of complex documents
  such as ODF documents.

  Consistency checking:
    Default implementation uses DocumentReferenceConstraint to check if the 
    reference/language/version triplet is unique. Additional constraints
    can be added if necessary.

  NOTE: Document.py supports a notion of revision which is very specific.
  The underlying concept is that, as soon as a document has a reference,
  the association of (reference, version, language) must be unique accross
  the whole system. This means that a given document in a given version in a
  given language is unique. The underlying idea is similar to the one in a Wiki
  system in which each page is unique and acts the the atom of collaboration.
  In the case of ERP5, if a team collaborates on a Text document written with
  an offline word processor, all updates should be placed inside the same object.
  A Contribution will thus modify an existing document, if allowed from security
  point of view, and increase the revision number. Same goes for properties
  (title). Each change generates a new revision.

    conversion API - not same as document - XXX BAD
    XXX make multiple interfaces

  TODO:
    - move all implementation bits to MixIn classes
    - in the end, Document class should have zero code
      and only serve as a quick and easy way to create 
      new types of documents (one could even consider 
      that this class should be trashed)
    - 
Bartek Górny's avatar
Bartek Górny committed
471 472 473 474 475
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
476
  isDocument = ConstantGetter('isDocument', value=True)
477
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
478

479 480 481 482 483 484 485
  zope.interface.implements(interfaces.IConvertable,
                            interfaces.ITextConvertable,
                            interfaces.IHtmlConvertable,
                            interfaces.ICachedConvertable,
                            interfaces.IVersionable,
                            interfaces.IDownloadable,
                            interfaces.ICrawlable,
486
                            interfaces.IDocument
487
                           )
488

Jean-Paul Smets's avatar
Jean-Paul Smets committed
489 490
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
491 492
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
493
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
494
  charset_parser = re.compile('charset="?([a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
495

Bartek Górny's avatar
Bartek Górny committed
496 497 498 499 500 501 502 503 504 505 506
  # 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
507 508
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
509
                    , PropertySheet.Periodicity
510
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
511 512
                    )

513 514
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
515
                              'subject', 'source_reference',)
Bartek Górny's avatar
Bartek Górny committed
516

517
  ### Content processing methods
518
  security.declareProtected(Permissions.View, 'index_html')
519
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
520 521
    """
      We follow here the standard Zope API for files and images
Jean-Paul Smets's avatar
Jean-Paul Smets committed
522 523 524 525 526 527 528 529 530 531 532
      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.

533 534 535
      Should return appropriate format (calling convert
      if necessary) and set headers.

536
      format -- the format specied in the form of an extension
537
      string (ex. jpeg, html, text, txt, etc.)
538 539

      **kw -- can be various things - e.g. resolution
540

Jean-Paul Smets's avatar
Jean-Paul Smets committed
541 542 543
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
544
    """
545
    raise NotImplementedError
546

Bartek Górny's avatar
Bartek Górny committed
547 548
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
549
    """
550 551 552
      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.
553

554 555
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
556

557
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
558 559
    """
    def getPropertyListOrValue(property):
560
      """
561
        we try to get a list, else we get value and convert to list
562
      """
563 564 565 566
      method = getattr(self, property, None)
      if method is not None:
        if callable(method):
          val = method()
Kazuhiko Shiozaki's avatar
Kazuhiko Shiozaki committed
567
          if isinstance(val, (list, tuple)):
568 569
            return list(val)
          return [str(val)]
Bartek Górny's avatar
Bartek Górny committed
570 571 572 573
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
574
          val = [str(val)]
575 576 577
        else:
          val = []
      else:
578
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
579
      return val
580

581
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
582
                                                self.searchable_property_list))
583
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
584 585 586
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
587
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
588

589 590 591 592 593 594 595
  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
596
  ### Relation getters
597
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
598
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
599
    """
600 601 602 603 604
      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.
    """
605
    text = self.getSearchableText() # XXX getSearchableText or asText ?
606
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
607
    try:
608
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
609
    except TypeError: # no regexp in preference
610 611 612
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
613 614 615 616 617 618 619 620 621 622 623 624
    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

625
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
626 627
  def getImplicitSuccessorValueList(self):
    """
628 629 630
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
631

632
      The implementation goes in 2 steps:
633 634

    - Step 1: extract with a regular expression
635
      a list of distionaries with various parameters such as
636 637 638 639 640
      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
641
      with appropriate parameters (and if possible build
642 643
      a complex query whenever this becomes available in
      portal catalog)
644

645 646
      The script is reponsible for calling getSearchableReferenceList
      so that it can use another approach if needed.
647

648 649
      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
650
    """
651
    tv = getTransactionalVariable(self) # XXX Performance improvement required
652 653 654 655 656 657
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

658
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
659
    result = self.Base_getImplicitSuccessorValueList(reference_list)
660 661
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
662

663
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
664 665 666
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
667
      by reference/language etc. Implementation is passed to
668
        Base_getImplicitPredecessorValueList
669 670 671 672 673 674 675 676

      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)
677
      ( if possible build  a complex query whenever
678 679 680 681
      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
682
    """
683
    return self._getTypeBasedMethod('getImplicitPredecessorValueList')()
Bartek Górny's avatar
Bartek Górny committed
684

685
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
686 687 688
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
689
      are similar. Not implemented yet.
Bartek Górny's avatar
Bartek Górny committed
690 691 692 693 694

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

695
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
696
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
697 698
    """
      Returns all documents which are similar to us, directly or indirectly, and
699
      in both directions. In other words, it is a transitive closure of similar
Bartek Górny's avatar
Bartek Górny committed
700 701 702 703 704
      relation. Every document is returned in the latest version available.
    """
    lista = {}
    depth = int(depth)

705 706
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
707

708
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
709
      level += 1
710 711 712 713
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
714 715 716
      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
717 718
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
719

720
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
721 722 723
    lista_latest = {}
    for o in lista.keys():
      lista_latest[o.getLatestVersionValue()] = True # get latest versions avoiding duplicates again
724
    if lista_latest.has_key(self):
Bartek Górny's avatar
Bartek Górny committed
725
      lista_latest.pop(self) # remove this document
726 727
    if lista_latest.has_key(self.getLatestVersionValue()):
      # remove last version of document itself from related documents
728
      lista_latest.pop(self.getLatestVersionValue())
Bartek Górny's avatar
Bartek Górny committed
729 730 731

    return lista_latest.keys()

732
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
733 734 735
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
736 737
      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
738

739 740
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
741

742 743 744
      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
745
    """
Bartek Górny's avatar
Bartek Górny committed
746 747
    if not self.getReference():
      return self
748
    catalog = getToolByName(self, 'portal_catalog', None)
749
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
750 751 752 753 754 755
    if language is not None: kw['language'] = language
    res = catalog(**kw)

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

756
    # if language was given return it (if there are any docs in this language)
757
    if language is not None:
758 759 760 761
      try:
        return res[0].getObject()
      except IndexError:
        return None
762 763 764 765 766 767 768
    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
769
            return in_original.getObject()
770
          else:
Bartek Górny's avatar
Bartek Górny committed
771
            return first.getObject() # this shouldn't happen in real life
772 773
        if ob.getLanguage() == user_language:
          # we found it in the user language
Bartek Górny's avatar
Bartek Górny committed
774
          return ob.getObject()
775 776 777
        if ob.getLanguage() == original_language:
          # this is in original language
          in_original = ob
778 779
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
780 781 782 783 784 785 786

  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
787
    catalog = getToolByName(self, 'portal_catalog', None)
788
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
789
                   reference=self.getReference(),
790
                   sort_on=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
791
                  )
792 793 794
    if version: kw['version'] = version
    if language: kw['language'] = language
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
795

796
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
797 798
  def isVersionUnique(self):
    """
799 800 801
      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
802
    """
803 804
    if not self.getReference():
      return 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
805
    catalog = getToolByName(self, 'portal_catalog', None)
806
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
807 808 809
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
810
                                            uid=self.getUid(),
811
                                            validation_state="!=cancelled"
812 813 814 815 816
                                            )[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
817
                                            validation_state="!=cancelled"
818 819 820 821
                                            )[0][0]
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

822 823 824 825
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
826 827 828
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
829
    """
830 831
    getInfoFor = getToolByName(self, 'portal_workflow').getInfoFor
    revision = len(getInfoFor(self, 'history', (), 'edit_workflow'))
832
    # XXX Also look at processing_status_workflow for compatibility.
833 834
    for history_item in getInfoFor(self, 'history', (),
                                   'processing_status_workflow'):
835 836 837
      if history_item.get('action') == 'edit':
        revision += 1
    return str(revision)
838 839 840 841 842 843 844

  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.
    """
845
    return map(str, range(1, 1 + int(self.getRevision())))
846

847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
  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(),
873
                where_expression=SQLQuery("validation_state NOT IN ('cancelled', 'deleted')"))
874 875 876 877 878 879
      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
880
      document_list = list(document_list)
881
      document_list.sort(key=lambda x: x.getId())
Ivan Tyagov's avatar
Ivan Tyagov committed
882
      #LOG('[DMS] Existing documents for %s' %self.getSourceReference(), INFO, len(document_list))
883 884 885 886 887 888 889 890 891 892
      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():
893
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
894
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
895
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
896 897 898
        else:
          update_kw = {}
          for k in self.propertyIds():
899
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
900 901 902 903 904 905
              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

906
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
907 908 909 910 911
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
912
    if not self.getReference(): return []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
913
    catalog = getToolByName(self, 'portal_catalog', None)
914
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
915
                           reference=self.getReference(),
916 917 918 919
                           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
920

921
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
922 923 924
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
925 926

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
927 928 929 930 931
    """
    # 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?
932 933
    reference = self.getReference()
    if not reference:
934
      return
935 936 937 938
    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
939 940 941 942 943 944

  ### 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')
945
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
946 947 948 949 950
    """
      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:
951 952
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
953
        fallback_script_id='Document_getPropertyDictFromUserLogin')
954
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
955 956 957 958 959 960 961

  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
    """
962
    if getattr(self, 'hasData', None) is not None and not self.hasData():
Jérome Perrin's avatar
Jérome Perrin committed
963 964
      # if document is empty, we will not find anything in its content
      return dict()
965
    if not self.hasBaseData():
966
      raise NotConvertedError
967
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
968
        fallback_script_id='Document_getPropertyDictFromContent')
969
    return method()
Bartek Górny's avatar
Bartek Górny committed
970 971 972 973 974 975 976 977 978 979 980 981 982 983

  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).
984 985 986 987 988

      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
989 990 991 992 993 994
    """
    if hasattr(self, '_backup_input'):
      return getattr(self, '_backup_input')
    kw = {}
    for id in self.propertyIds():
      # We should not consider file data
995 996
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
997
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
998 999 1000 1001 1002 1003
        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

1004 1005 1006 1007 1008 1009
  security.declareProtected(Permissions.AccessContentsInformation, 'getStandardFileName')
  def getStandardFileName(self):
    """
    Returns the document coordinates as a standard file name. This
    method is the reverse of getPropertyDictFromFileName.
    """
1010
    method = self._getTypeBasedMethod('getStandardFileName',
1011 1012
        fallback_script_id = 'Document_getStandardFileName')
    return method()
1013

Bartek Górny's avatar
Bartek Górny committed
1014 1015
  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
1016
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
1017
    """
1018 1019
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
1020 1021 1022
      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
1023

1024
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
1025

1026
      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
1027
                   currently logged in, then we'll get him from session
1028
    """
Bartek Górny's avatar
Bartek Górny committed
1029
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
1030
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList',
Bartek Górny's avatar
Bartek Górny committed
1031
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
1032
    order_list = list(method())
1033
    order_list.reverse()
1034
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
1035
    kw = {}
1036
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1037
      result = None
Bartek Górny's avatar
Bartek Górny committed
1038 1039
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
1040
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
1041 1042 1043
      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
1044 1045
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
1046
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1047 1048
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
1049 1050
      else:
        result = method()
1051 1052
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1053

1054 1055 1056
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
1057
    # Prepare the content edit parameters - portal_type should not be changed
1058 1059 1060 1061
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
1062 1063
    self.finishIngestion() # XXX - is this really the right place ?
    self.reindexObject() # XXX - is this really the right place ?
1064 1065
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
1066 1067 1068
    merged_doc = self.mergeRevision() # XXX - is this really the right place ?
    merged_doc.reindexObject() # XXX - is this really the right place ?
    return merged_doc # XXX - is this really the right place ?
Bartek Górny's avatar
Bartek Górny committed
1069 1070 1071 1072

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1073 1074 1075
      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
1076
    """
1077 1078
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1079

1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140
  # Conversion methods
  security.declareProtected(Permissions.AccessContentsInformation, '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

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

      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
    """
    if format == 'html':
      return 'text/html', '' # XXX - Why ?
    if format in ('text', 'txt'):
      return 'text/plain', '' # XXX - Why ?
    raise NotImplementedError

  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)

  security.declareProtected(Permissions.View, 'asText')
  def asText(self, **kw):
    """
      Converts the content of the document to a textual representation.
    """
    kw['format'] = 'txt'
    mime, data = self.convert(**kw)
    return str(data)

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

      Actual conversion is delegated to _asHTML
    """
    html = self._asHTML(**kw)
    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:
Nicolas Delaby's avatar
Nicolas Delaby committed
1141 1142
        base = '<base href="%s"/>' % self.getContentBaseURL()
        html = html.replace('<head>', '<head>%s' % base, 1)
1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 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 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
      self.setConversion(html, mime='text/html', format='base-html')
    return html

  security.declarePrivate('_asHTML')
  def _asHTML(self, **kw):
    """
      A private method which converts to HTML. This method
      is the one to override in subclasses.
    """
    if not self.hasBaseData():
      raise ConversionError('This document has not been processed yet.')
    try:
      # FIXME: no substitution may occur in this case.
      mime, data = self.getConversion(format='base-html')
      return data
    except KeyError:
      kw['format'] = 'html'
      mime, html = self.convert(**kw)
      return html

  security.declareProtected(Permissions.View, 'asStrippedHTML')
  def asStrippedHTML(self, **kw):
    """
      Returns a stripped HTML representation of the document
      (without html and body tags, etc.) which can be used to inline
      a preview of the document.
    """
    if not self.hasBaseData():
      return ''
    try:
      # FIXME: no substitution may occur in this case.
      mime, data = self.getConversion(format='stripped-html')
      return data
    except KeyError:
      kw['format'] = 'html'
      mime, html = self.convert(**kw)
      return self._stripHTML(str(html))

  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)

  def _stripHTML(self, html, charset=None):
    """
      A private method which can be reused by subclasses
      to strip HTML content
    """
    body_list = re.findall(self.body_parser, str(html))
    if len(body_list):
      stripped_html = body_list[0]
    else:
      stripped_html = html
    # find charset and convert to utf-8
    charset_list = self.charset_parser.findall(str(html)) # XXX - Not efficient if this
                                         # is datastream instance but hard to do better
    if charset and not charset_list:
      # Use optional parameter is we can not find encoding in HTML
      charset_list = [charset]
    if charset_list and charset_list[0] not in ('utf-8', 'UTF-8'):
      try:
        stripped_html = unicode(str(stripped_html),
                                charset_list[0]).encode('utf-8')
      except (UnicodeDecodeError, LookupError):
        return str(stripped_html)
    return stripped_html

  def _safeHTML(self, html, format='text/x-html-safe', charset=None):
    """
      A private method to strip HTML content in safe mode,
      w/o emmbed javascript, forms and any external plugins imports.
      This should be used when we do not trust the user (Anonymous)
      who push data into database.
      - html: content to strip
      - format: destination format
      - charset: charset used to encode string. Take precedence
      on charset values found in html string
    """
    portal = self.getPortalObject()
    if charset is None:
      # find charset
      charset_list = self.charset_parser.findall(html)
      if charset_list:
        charset = charset_list[0]
    if charset and charset not in ('utf-8', 'UTF-8'):
      try:
        safe_html_string = html.decode(charset).encode('utf-8')
      except (UnicodeDecodeError, LookupError):
        pass
      else:
        charset = 'utf-8' # Override charset if convertion succeeds
    transform_tool = getToolByName(portal, 'portal_transforms')
    safe_html_string = transform_tool.convertToData(format, html,
                                                    encoding=charset)
    return safe_html_string

  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 = {}
    html = self.asEntireHTML()
    if not html: return result
    title_list = re.findall(self.title_parser, str(html))
    if title_list:
      result['title'] = title_list[0]
    return result

  # Base format support
  security.declareProtected(Permissions.ModifyPortalContent, 'convertToBaseFormat')
  def convertToBaseFormat(self, **kw):
    """
      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.)
    """
    if getattr(self, 'hasData', None) is not None and not self.hasData():
      # Empty document cannot be converted
      return
    try:
      message = self._convertToBaseFormat() # Call implemetation method
      self.clearConversionCache() # Conversion cache is now invalid
      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

  def _convertToBaseFormat(self):
    """
    """
    raise NotImplementedError

  security.declareProtected(Permissions.AccessContentsInformation,
                            'isSupportBaseDataConversion')
  def isSupportBaseDataConversion(self):
    """
    """
    return False

  def convertFile(self, **kw): # XXX - It it really useful to explicitly define ?
    """
    Workflow transition invoked when conversion occurs.
    """
  convertFile = WorkflowMethod(convertFile)

  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 {}

  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

1341
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1342
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354
  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
1355 1356 1357 1358 1359
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1360 1361

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1362
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1363 1364 1365
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1366 1367 1368 1369 1370 1371
      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.
1372
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1373 1374
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1375

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1376
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1377
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1378 1379
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1380 1381
      Implementation is handled by ContributionTool.
    """
1382
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1383

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1384 1385
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1386
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1387 1388 1389 1390
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1391 1392 1393 1394 1395 1396
  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
1397
      a difference between URLs which return an index (ex. the
1398 1399 1400 1401 1402 1403 1404 1405 1406
      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
1407 1408 1409 1410 1411 1412 1413 1414
  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
1415
    if len(base_url_list):
1416
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1417
        # Cut the trailing part in http://www.some.site/at/trailing.html
1418
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1419
        base_url = '/'.join(base_url_list[:-1])
1420
    return base_url
1421 1422 1423

  security.declareProtected(Permissions.ModifyPortalContent, '_setBaseData')
  def _setBaseData(self, data):
1424 1425 1426 1427 1428 1429 1430 1431 1432
    """
      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)
1433
    self._baseSetBaseData(data)
1434 1435 1436 1437 1438

  security.declareProtected(Permissions.AccessContentsInformation,
                            'getBaseData')
  def getBaseData(self, default=None):
    """return BaseData as str."""
1439 1440 1441 1442 1443
    base_data = self._baseGetBaseData()
    if base_data is None:
      return None
    else:
      return str(base_data)
1444 1445 1446

  security.declareProtected(Permissions.ModifyPortalContent, '_setData')
  def _setData(self, data):
1447 1448 1449 1450 1451 1452
    """
      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
1453 1454 1455 1456 1457
    # update_data use len(data) when size is None, which breaks this method.
    # define size = 0 will prevent len be use and keep the consistency of 
    # getData() and setData()
    if data is None:
      size = 0
1458 1459 1460
    if not isinstance(data, Pdata) and data is not None:
      file = cStringIO.StringIO(data)
      data, size = self._read_data(file)
Julien Muchembled's avatar
typo  
Julien Muchembled committed
1461
    if getattr(self, 'update_data', None) is not None:
1462 1463
      # We call this method to make sure size is set and caches reset
      self.update_data(data, size=size)
1464 1465 1466 1467 1468 1469
    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()
1470 1471 1472 1473

  security.declareProtected(Permissions.AccessContentsInformation, 'getData')
  def getData(self, default=None):
    """return Data as str."""
1474 1475 1476 1477 1478
    data = self._baseGetData()
    if data is None:
      return None
    else:
      return str(data)