Document.py 32.5 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.
#
##############################################################################

Aurel's avatar
Aurel committed
30
import re
Rafael Monnerat's avatar
Rafael Monnerat committed
31
from zLOG import LOG
Aurel's avatar
Aurel committed
32
from AccessControl import ClassSecurityInfo
Romain Courteaud's avatar
Romain Courteaud committed
33
from Acquisition import aq_base
34
from Products.ERP5Type.Accessor.Constant import PropertyGetter as ConstantGetter
Nicolas Delaby's avatar
Nicolas Delaby committed
35
from Products.CMFCore.utils import _checkPermission
Nicolas Dumazet's avatar
Nicolas Dumazet committed
36
from Products.ERP5Type import Permissions, PropertySheet, interfaces
Jean-Paul Smets's avatar
Jean-Paul Smets committed
37
from Products.ERP5Type.XMLObject import XMLObject
Aurel's avatar
Aurel committed
38
from Products.ERP5Type.Utils import deprecated, guessEncodingFromText
39
from Products.ERP5Type.TransactionalVariable import getTransactionalVariable
40
from Products.ERP5.Tool.ContributionTool import MAX_REPEAT
41
from Products.ZSQLCatalog.SQLCatalog import Query, NegatedQuery
42
from AccessControl import Unauthorized
43
import zope.interface
44
from Products.PythonScripts.Utility import allow_class
45

46 47
# Mixin Import
from Products.ERP5.mixin.cached_convertable import CachedConvertableMixin
Nicolas Delaby's avatar
Nicolas Delaby committed
48 49
from Products.ERP5.mixin.text_convertable import TextConvertableMixin
from Products.ERP5.mixin.downloadable import DownloadableMixin
50
from Products.ERP5.mixin.document import DocumentMixin
51
from Products.ERP5.mixin.extensible_traversable import DocumentExtensibleTraversableMixin
52
from Products.ERP5.mixin.crawlable import CrawlableMixin
Nicolas Delaby's avatar
Nicolas Delaby committed
53 54
from Products.ERP5.mixin.discoverable import DiscoverableMixin
from Products.ERP5.mixin.url import UrlMixin
55

Bartek Górny's avatar
Bartek Górny committed
56
_MARKER = []
57

58
# these property ids are unchangable
Nicolas Delaby's avatar
Nicolas Delaby committed
59
FIXED_PROPERTY_IDS = ('id', 'uid', 'rid', 'sid')
Jean-Paul Smets's avatar
Jean-Paul Smets committed
60

61
# XXX: move to an easier to configure place (System Preference ?)
62 63 64
VALID_TEXT_FORMAT_LIST = ('text', 'txt', 'html', 'base_html',
                          'stripped-html')

65 66
VALID_IMAGE_FORMAT_LIST = ('jpg', 'jpeg', 'png', 'gif', 'pnm', 'ppm', 'tiff', 'svg')
VALID_TRANSPARENT_IMAGE_FORMAT_LIST = ('png', 'gif', 'tiff', 'svg')
67

68 69
DEFAULT_DISPLAY_ID_LIST = ('nano', 'micro', 'thumbnail',
                            'xsmall', 'small', 'medium',
Kazuhiko Shiozaki's avatar
Kazuhiko Shiozaki committed
70
                            'large', 'xlarge',)
71 72
# default image qualitay (obsoleted use getPreferredImageQuality on a portal_preferences tool)
DEFAULT_IMAGE_QUALITY = 75.0
73

Ivan Tyagov's avatar
Ivan Tyagov committed
74 75
DEFAULT_CONTENT_TYPE = 'text/html'

76
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
77

78 79
class DocumentProxyError(Exception):pass

80
class NotConvertedError(Exception):pass
81
allow_class(NotConvertedError)
82

Nicolas Delaby's avatar
Nicolas Delaby committed
83 84 85
class Document(DocumentExtensibleTraversableMixin, XMLObject, UrlMixin,
               CachedConvertableMixin, CrawlableMixin, TextConvertableMixin,
               DownloadableMixin, DocumentMixin, DiscoverableMixin):
86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
  """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
112
    as Wiki to implement specific methods.
113 114
    TextDocument itself has a subclass (XSLTDocument) which provides
    XSLT based analysis and transformation of XML content based on XSLT
115
    templates.
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139

  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)
Nicolas Delaby's avatar
Nicolas Delaby committed
140
  filename   -   data which might be encoded in filename
141 142 143 144
  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
145
  preference order returned by a script
146 147 148 149 150 151
     Document_getPreferredDocumentMetadataDiscoveryOrderList
     (or any type-based version since discovery is type dependent)

  Methods for discovering metadata are:

    getPropertyDictFromInput
Nicolas Delaby's avatar
Nicolas Delaby committed
152
    getPropertyDictFromFilename
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
    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'},
181
     {'reference': 'INV-1112', 'version':'012}',
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
     {'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:
221
    Default implementation uses DocumentReferenceConstraint to check if the
222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242
    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
243 244
      and only serve as a quick and easy way to create
      new types of documents (one could even consider
245
      that this class should be trashed)
246
    -
Bartek Górny's avatar
Bartek Górny committed
247 248 249 250 251
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
252
  isDocument = ConstantGetter('isDocument', value=True)
253
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
254

255 256 257 258 259 260 261
  zope.interface.implements(interfaces.IConvertable,
                            interfaces.ITextConvertable,
                            interfaces.IHtmlConvertable,
                            interfaces.ICachedConvertable,
                            interfaces.IVersionable,
                            interfaces.IDownloadable,
                            interfaces.ICrawlable,
Nicolas Delaby's avatar
Nicolas Delaby committed
262 263 264
                            interfaces.IDocument,
                            interfaces.IDiscoverable,
                            interfaces.IUrl,
265
                           )
266

Jean-Paul Smets's avatar
Jean-Paul Smets committed
267
  # Regular expressions
Nicolas Delaby's avatar
Nicolas Delaby committed
268 269 270
  # XXX those regex are weak, fast but not reliable.
  # this is a valid url than regex are not able to parse
  # http://www.example.com//I don't care i put what/ i want/
Jean-Paul Smets's avatar
Jean-Paul Smets committed
271
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
272 273
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
274
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
275
  charset_parser = re.compile('(?P<keyword>charset="?)(?P<charset>[a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
276

Bartek Górny's avatar
Bartek Górny committed
277 278 279 280 281 282 283 284 285 286
  # Declarative security
  security = ClassSecurityInfo()
  security.declareObjectProtected(Permissions.AccessContentsInformation)

  # Declarative properties
  property_sheets = ( PropertySheet.Base
                    , PropertySheet.XMLObject
                    , PropertySheet.CategoryCore
                    , PropertySheet.DublinCore
                    , PropertySheet.Version
287
                    , PropertySheet.Reference
Bartek Górny's avatar
Bartek Górny committed
288
                    , PropertySheet.Document
Jean-Paul Smets's avatar
Jean-Paul Smets committed
289 290
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
291
                    , PropertySheet.Periodicity
Bartek Górny's avatar
Bartek Górny committed
292 293
                    )

Nicolas Delaby's avatar
Nicolas Delaby committed
294 295
  index_html = DownloadableMixin.index_html

296 297 298 299 300 301 302
  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
303
  ### Relation getters
304
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
305
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
306
    """
307 308 309 310 311
      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.
    """
312
    text = self.getSearchableText() # XXX getSearchableText or asText ?
313 314 315 316 317 318 319 320 321
    return self._getSearchableReferenceList(text)

  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
  def isSearchableReference(self):
    """
      Determine if current document's reference can be used for searching - i.e. follows
      certain defined at system level preferences format.
    """
    reference = self.getReference()
322 323 324
    if reference is not None:
      return len(self._getSearchableReferenceList(reference))
    return False
325 326 327

  def _getSearchableReferenceList(self, text):
    """
328
      Extract all reference alike strings from text using for that a
329 330
      regular expression defined at system level preferences.
    """
331
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
332
    try:
333
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
334
    except TypeError: # no regexp in preference
335 336 337
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
338 339 340 341 342 343 344 345 346 347
    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)))
348
    return result
349

350
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
351 352
  def getImplicitSuccessorValueList(self):
    """
353 354 355
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
356

357
      The implementation goes in 2 steps:
358 359

    - Step 1: extract with a regular expression
360
      a list of distionaries with various parameters such as
361 362 363 364 365
      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
366
      with appropriate parameters (and if possible build
367 368
      a complex query whenever this becomes available in
      portal catalog)
369

370 371
      The script is reponsible for calling getSearchableReferenceList
      so that it can use another approach if needed.
372

373 374
      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
375
    """
376
    tv = getTransactionalVariable() # XXX Performance improvement required
377 378 379 380 381 382
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

383
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
384
    result = self.Base_getImplicitSuccessorValueList(reference_list)
385 386
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
387

388
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
389 390 391
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
392
      by reference/language etc. Implementation is passed to
393
        Base_getImplicitPredecessorValueList
394 395 396 397 398 399 400 401

      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)
402
      ( if possible build  a complex query whenever
403 404 405 406
      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
407
    """
408
    return self._getTypeBasedMethod('getImplicitPredecessorValueList')()
Bartek Górny's avatar
Bartek Górny committed
409

410
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
411 412 413
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
414
      are similar. Not implemented yet.
Bartek Górny's avatar
Bartek Górny committed
415 416 417 418 419

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

420
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
421
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
422 423
    """
      Returns all documents which are similar to us, directly or indirectly, and
424
      in both directions. In other words, it is a transitive closure of similar
Bartek Górny's avatar
Bartek Górny committed
425 426 427 428 429
      relation. Every document is returned in the latest version available.
    """
    lista = {}
    depth = int(depth)

430 431
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
432

433
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
434
      level += 1
435 436 437 438
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
439 440 441
      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
442 443
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
444

445
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
446 447 448
    lista_latest = {}
    for o in lista.keys():
      lista_latest[o.getLatestVersionValue()] = True # get latest versions avoiding duplicates again
449 450 451 452 453

    # remove this document
    lista_latest.pop(self, None)
    # remove last version of document itself from related documents
    lista_latest.pop(self.getLatestVersionValue(), None)
Bartek Górny's avatar
Bartek Górny committed
454 455 456

    return lista_latest.keys()

457
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
458 459 460
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
461 462
      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
463

464 465
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
466

467 468 469
      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
470
    """
Bartek Górny's avatar
Bartek Górny committed
471 472
    if not self.getReference():
      return self
Nicolas Delaby's avatar
Nicolas Delaby committed
473
    catalog = self.getPortalObject().portal_catalog
474 475
    kw = dict(reference=self.getReference(),
              sort_on=(('version', 'descending', 'SIGNED'),))
476 477
    if language is not None:
      kw['language'] = language
478
    result_list = catalog(**kw)
479 480 481 482

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

483
    # if language was given return it (if there are any docs in this language)
484
    if language is not None:
485
      try:
Nicolas Delaby's avatar
Nicolas Delaby committed
486
        return result_list[0].getObject()
487 488
      except IndexError:
        return None
489 490
    elif result_list:
      first =  result_list[0].getObject()
491
      in_original = None
Nicolas Delaby's avatar
Nicolas Delaby committed
492 493 494
      for record in result_list:
        document = record.getObject()
        if document.getVersion() != first.getVersion():
495 496
          # we are out of the latest version - return in_original or first
          if in_original is not None:
Nicolas Delaby's avatar
Nicolas Delaby committed
497
            return in_original
498
          else:
Nicolas Delaby's avatar
Nicolas Delaby committed
499 500
            return first # this shouldn't happen in real life
        if document.getLanguage() == user_language:
501
          # we found it in the user language
Nicolas Delaby's avatar
Nicolas Delaby committed
502 503
          return document
        if document.getLanguage() == original_language:
504
          # this is in original language
Nicolas Delaby's avatar
Nicolas Delaby committed
505
          in_original = document
506 507
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
508 509 510 511 512 513 514

  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.
    """
Nicolas Delaby's avatar
Nicolas Delaby committed
515
    catalog = self.getPortalObject().portal_catalog
516
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
517
                   reference=self.getReference(),
518
                   sort_on=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
519
                  )
Jérome Perrin's avatar
Jérome Perrin committed
520 521 522 523
    if version:
      kw['version'] = version
    if language:
      kw['language'] = language
524
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
525

526
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
527 528
  def isVersionUnique(self):
    """
529 530 531
      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
532
    """
533
    if not self.getReference():
Jérome Perrin's avatar
Jérome Perrin committed
534
      return True
535 536 537 538
    kw = dict(portal_type=self.getPortalDocumentTypeList(),
              reference=self.getReference(),
              version=self.getVersion(),
              language=self.getLanguage(),
539
              query=NegatedQuery(Query(validation_state=('cancelled', 'deleted'))))
Nicolas Delaby's avatar
Nicolas Delaby committed
540
    catalog = self.getPortalObject().portal_catalog
541 542
    self_count = catalog.unrestrictedCountResults(uid=self.getUid(), **kw)[0][0]
    count = catalog.unrestrictedCountResults(**kw)[0][0]
543 544 545
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

546 547 548 549
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
550 551 552
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
553
    """
Nicolas Delaby's avatar
Nicolas Delaby committed
554
    getInfoFor = self.getPortalObject().portal_workflow.getInfoFor
555
    revision = len(getInfoFor(self, 'history', (), 'edit_workflow'))
556
    # XXX Also look at processing_status_workflow for compatibility.
Nicolas Delaby's avatar
Nicolas Delaby committed
557 558 559
    revision += len([history_item for history_item in\
                 getInfoFor(self, 'history', (), 'processing_status_workflow')\
                 if history_item.get('action') == 'edit'])
560
    return str(revision)
561 562 563 564 565 566 567

  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.
    """
568
    return map(str, range(1, 1 + int(self.getRevision())))
569

570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591
  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
    if self.getReference():
592
      invalid_validation_state_list = ('archived', 'cancelled', 'deleted')
593
      catalog = self.getPortalObject().portal_catalog
594
      # Find all document with same (reference, version, language)
595
      kw = dict(portal_type=self.getPortalType(),
596
                reference=self.getReference(),
597
                query=NegatedQuery(Query(validation_state=invalid_validation_state_list)),
Nicolas Delaby's avatar
Nicolas Delaby committed
598
                sort_on='creation_date')
599

Jérome Perrin's avatar
Jérome Perrin committed
600 601 602 603
      if self.getVersion():
        kw['version'] = self.getVersion()
      if self.getLanguage():
        kw['language'] = self.getLanguage()
604 605 606 607 608 609 610 611 612
      document_list = catalog.unrestrictedSearchResults(**kw)
      existing_document = None
      # Select the first one which is not self and which
      # shares the same coordinates
      for o in document_list:
        if o.getRelativeUrl() != self.getRelativeUrl() and\
           o.getVersion() == self.getVersion() and\
           o.getLanguage() == self.getLanguage():
          existing_document = o.getObject()
613 614 615 616 617 618
          if existing_document.getValidationState() not in \
            invalid_validation_state_list:
            break
      else:
        existing_document = None

619 620 621
      # We found an existing document to update
      if existing_document is not None:
        document = existing_document
622
        if not _checkPermission(Permissions.ModifyPortalContent, existing_document):
623
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
624 625 626
        else:
          update_kw = {}
          for k in self.propertyIds():
627
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
628 629 630
              update_kw[k] = self.getProperty(k)
          existing_document.edit(**update_kw)
          # Erase self
631
          self.getParentValue().manage_delObjects([self.getId(),])
632 633
    return document

634
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
635 636 637 638 639
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
640
    if not self.getReference(): return []
Nicolas Delaby's avatar
Nicolas Delaby committed
641
    catalog = self.getPortalObject().portal_catalog
642
    kw = dict(portal_type=self.getPortalType(),
Jérome Perrin's avatar
Jérome Perrin committed
643 644
              reference=self.getReference(),
              group_by=('language',))
645 646 647
    if version is not None:
      kw['version'] = version
    return map(lambda o:o.getLanguage(), catalog(**kw))
Bartek Górny's avatar
Bartek Górny committed
648

649
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
650 651 652
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
653 654

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
655 656 657 658 659
    """
    # 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?
660 661
    reference = self.getReference()
    if not reference:
662
      return
Nicolas Delaby's avatar
Nicolas Delaby committed
663
    catalog = self.getPortalObject().portal_catalog
Nicolas Delaby's avatar
Nicolas Delaby committed
664 665
    result_list = catalog.unrestrictedSearchResults(
                                      reference=self.getReference(),
666
                                      sort_on=(('creation_date',
Nicolas Delaby's avatar
Nicolas Delaby committed
667 668 669 670
                                                'ascending'),))
    if result_list:
      return result_list[0].getLanguage()
    return
Bartek Górny's avatar
Bartek Górny committed
671

672 673 674 675 676
  security.declareProtected(Permissions.View, 'asSubjectText')
  def asSubjectText(self, **kw):
    """
      Converts the subject of the document to a textual representation.
    """
Nicolas Delaby's avatar
Nicolas Delaby committed
677
    subject = self.getSubject('')
678 679
    if not subject:
      # XXX not sure if this fallback is a good idea.
Nicolas Delaby's avatar
Nicolas Delaby committed
680
      subject = self.getTitle('')
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
    return str(subject)

  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
699 700
        base = '<base href="%s"/>' % self.getContentBaseURL()
        html = html.replace('<head>', '<head>%s' % base, 1)
701 702 703 704 705 706 707 708 709
      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.
    """
710 711 712
    kw['format'] = 'html'
    mime, html = self.convert(**kw)
    return html
713 714 715 716 717 718 719 720

  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.
    """
721
    return self._stripHTML(self._asHTML(**kw))
722

723
  security.declarePrivate('_guessEncoding')
Nicolas Delaby's avatar
Nicolas Delaby committed
724
  @deprecated
725
  def _guessEncoding(self, string, mime='text/html'):
726
    """
Nicolas Delaby's avatar
Nicolas Delaby committed
727
      Deprecated method
728
    """
729
    return guessEncodingFromText(string, content_type=mime)
730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751

  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
    return stripped_html

  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')
752
    except (KeyError, AttributeError):
753 754 755 756 757 758
      method = None
    if method is not None:
      return method()
    else:
      return {}

759
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
760
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
761 762 763 764 765 766 767 768 769 770 771 772
  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
773 774
    try:
      method = self._getTypeBasedMethod('populateContent')
775
    except (KeyError, AttributeError):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
776 777
      method = None
    if method is not None: method()
778

Jean-Paul Smets's avatar
Jean-Paul Smets committed
779
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
780 781
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0,
                           repeat_interval=1, batch_mode=True):
782 783
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
784 785
      Implementation is handled by ContributionTool.
    """
786
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
787

Jean-Paul Smets's avatar
Jean-Paul Smets committed
788 789
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
790
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
791 792 793 794
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

795 796 797 798 799 800
  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
801
      a difference between URLs which return an index (ex. the
802 803 804 805 806 807 808 809
      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
810