Document.py 53.4 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.
#
##############################################################################

Nicolas Delaby's avatar
Nicolas Delaby committed
30
import re, sys, os
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
Nicolas Delaby's avatar
Nicolas Delaby committed
41 42
from Products.ERP5Type.DateUtils import convertDateToHour,\
                                number_of_hours_in_day, number_of_hours_in_year
43
from Products.ERP5Type.Utils import convertToUpperCase
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
from Products.PythonScripts.Utility import allow_class
Nicolas Delaby's avatar
Nicolas Delaby committed
54 55
import tempfile
from subprocess import Popen, PIPE
56

57 58
# Mixin Import
from Products.ERP5.mixin.cached_convertable import CachedConvertableMixin
Nicolas Delaby's avatar
Nicolas Delaby committed
59 60
from Products.ERP5.mixin.text_convertable import TextConvertableMixin
from Products.ERP5.mixin.downloadable import DownloadableMixin
61

Bartek Górny's avatar
Bartek Górny committed
62
_MARKER = []
63
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
64
# these property ids are unchangable
Nicolas Delaby's avatar
Nicolas Delaby committed
65
FIXED_PROPERTY_IDS = ('id', 'uid', 'rid', 'sid')
Jean-Paul Smets's avatar
Jean-Paul Smets committed
66

67 68 69 70 71
VALID_TEXT_FORMAT_LIST = ('text', 'txt', 'html', 'base_html',
                          'stripped-html')

VALID_IMAGE_FORMAT_LIST = ('jpg', 'jpeg', 'png', 'gif', 'pnm', 'ppm', 'tiff')

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
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

113
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
114

115 116
class DocumentProxyError(Exception):pass

117
class NotConvertedError(Exception):pass
118
allow_class(NotConvertedError)
119

120 121 122 123 124 125 126
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
127

128 129 130
  # Declarative security
  security = ClassSecurityInfo()

131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148
  ### 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':
149
        user_folder = getattr(self.getPortalObject(), 'acl_users', None)
150 151 152 153
        if user_folder is not None:
          try:
            if request.get('PUBLISHED', _MARKER) is _MARKER:
              # request['PUBLISHED'] is required by validate
154
              request['PUBLISHED'] = self
155 156 157
              has_published = False
            else:
              has_published = True
158 159 160 161 162 163 164
            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
165
            if not has_published:
166 167 168 169 170 171
              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']
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
          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)
198

199 200 201
    # 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
202 203
    isAuthorizationForced = getattr(self, 'isAuthorizationForced', None)
    if isAuthorizationForced is not None and isAuthorizationForced():
204
      if unrestricted_apply(self.getDocumentValue, (name, portal)) is not None:
205 206
        # force user to login as specified in Web Section
        raise Unauthorized
207

208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
  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()

230
    method = self._getTypeBasedMethod('getDocumentValue',
231
              fallback_script_id='WebSection_getDocumentValue')
232

233 234
    document = method(name, portal=portal, **kw)
    if document is not None:
235
      return document.__of__(self)
236

237 238 239 240 241 242 243 244 245 246
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' )
247 248
  def index_html(self, REQUEST, RESPONSE, display=None, format='', quality=75,
                                resolution=None, frame=0, **kw):
249
    """ Only a proxy method """
250 251 252
    return self.getProxiedDocument().index_html(REQUEST, RESPONSE, 
        display=display, format=format, quality=quality, resolution=resolution,
        frame=frame, **kw)
253 254 255 256 257 258 259 260 261

  security.declareProtected(Permissions.AccessContentsInformation,
                            'getProxiedDocument' )
  def getProxiedDocument(self):
    """
    Try to retrieve the original document
    """
    proxied_document = self.getDocumentProxyValue()
    if proxied_document is None:
262
      raise DocumentProxyError("Unable to find a proxied document")
263 264
    return proxied_document

265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
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):
    """
291
    Returns the document Creation Date Index which is the creation
292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
    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.
    """
314
    method = self._getTypeBasedMethod('isUpdatable',
315 316 317 318
        fallback_script_id = 'Document_isUpdatable')
    return method()


Nicolas Delaby's avatar
Nicolas Delaby committed
319 320 321
class Document(PermanentURLMixIn, XMLObject, UrlMixIn, CachedConvertableMixin,
               SnapshotMixin, UpdateMixIn, TextConvertableMixin,
               DownloadableMixin):
322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482
  """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
483 484 485 486 487
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
488
  isDocument = ConstantGetter('isDocument', value=True)
489
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
490

491 492 493 494 495 496 497
  zope.interface.implements(interfaces.IConvertable,
                            interfaces.ITextConvertable,
                            interfaces.IHtmlConvertable,
                            interfaces.ICachedConvertable,
                            interfaces.IVersionable,
                            interfaces.IDownloadable,
                            interfaces.ICrawlable,
498
                            interfaces.IDocument
499
                           )
500

Jean-Paul Smets's avatar
Jean-Paul Smets committed
501 502
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
503 504
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
505
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
506
  charset_parser = re.compile('(?P<keyword>charset="?)(?P<charset>[a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
507

Bartek Górny's avatar
Bartek Górny committed
508 509 510 511 512 513 514 515 516 517 518
  # 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
519 520
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
521
                    , PropertySheet.Periodicity
522
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
523 524
                    )

525
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
Nicolas Delaby's avatar
Nicolas Delaby committed
526 527
                              'version', 'short_title', 'subject',
                              'source_reference', 'source_project_title')
528

529

Bartek Górny's avatar
Bartek Górny committed
530 531
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
532
    """
533 534 535
      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.
536

537 538
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
539

540
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
541 542
    """
    def getPropertyListOrValue(property):
543
      """
544
        we try to get a list, else we get value and convert to list
545
      """
546 547 548 549
      method = getattr(self, property, None)
      if method is not None:
        if callable(method):
          val = method()
Kazuhiko Shiozaki's avatar
Kazuhiko Shiozaki committed
550
          if isinstance(val, (list, tuple)):
551 552
            return list(val)
          return [str(val)]
Bartek Górny's avatar
Bartek Górny committed
553 554 555 556
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
557
          val = [str(val)]
558 559 560
        else:
          val = []
      else:
561
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
562
      return val
563

564
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
565
                                                self.searchable_property_list))
566
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
567 568 569
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
570
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
571

Nicolas Delaby's avatar
Nicolas Delaby committed
572 573
  index_html = DownloadableMixin.index_html

574 575 576 577 578 579 580
  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
581
  ### Relation getters
582
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
583
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
584
    """
585 586 587 588 589
      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.
    """
590
    text = self.getSearchableText() # XXX getSearchableText or asText ?
591
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
592
    try:
593
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
594
    except TypeError: # no regexp in preference
595 596 597
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
598 599 600 601 602 603 604 605 606 607 608 609
    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

610
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
611 612
  def getImplicitSuccessorValueList(self):
    """
613 614 615
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
616

617
      The implementation goes in 2 steps:
618 619

    - Step 1: extract with a regular expression
620
      a list of distionaries with various parameters such as
621 622 623 624 625
      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
626
      with appropriate parameters (and if possible build
627 628
      a complex query whenever this becomes available in
      portal catalog)
629

630 631
      The script is reponsible for calling getSearchableReferenceList
      so that it can use another approach if needed.
632

633 634
      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
635
    """
636
    tv = getTransactionalVariable(self) # XXX Performance improvement required
637 638 639 640 641 642
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

643
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
644
    result = self.Base_getImplicitSuccessorValueList(reference_list)
645 646
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
647

648
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
649 650 651
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
652
      by reference/language etc. Implementation is passed to
653
        Base_getImplicitPredecessorValueList
654 655 656 657 658 659 660 661

      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)
662
      ( if possible build  a complex query whenever
663 664 665 666
      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
667
    """
668
    return self._getTypeBasedMethod('getImplicitPredecessorValueList')()
Bartek Górny's avatar
Bartek Górny committed
669

670
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
671 672 673
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
674
      are similar. Not implemented yet.
Bartek Górny's avatar
Bartek Górny committed
675 676 677 678 679

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

680
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
681
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
682 683
    """
      Returns all documents which are similar to us, directly or indirectly, and
684
      in both directions. In other words, it is a transitive closure of similar
Bartek Górny's avatar
Bartek Górny committed
685 686 687 688 689
      relation. Every document is returned in the latest version available.
    """
    lista = {}
    depth = int(depth)

690 691
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
692

693
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
694
      level += 1
695 696 697 698
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
699 700 701
      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
702 703
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
704

705
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
706 707 708
    lista_latest = {}
    for o in lista.keys():
      lista_latest[o.getLatestVersionValue()] = True # get latest versions avoiding duplicates again
709
    if lista_latest.has_key(self):
Bartek Górny's avatar
Bartek Górny committed
710
      lista_latest.pop(self) # remove this document
711 712
    if lista_latest.has_key(self.getLatestVersionValue()):
      # remove last version of document itself from related documents
713
      lista_latest.pop(self.getLatestVersionValue())
Bartek Górny's avatar
Bartek Górny committed
714 715 716

    return lista_latest.keys()

717
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
718 719 720
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
721 722
      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
723

724 725
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
726

727 728 729
      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
730
    """
Bartek Górny's avatar
Bartek Górny committed
731 732
    if not self.getReference():
      return self
Nicolas Delaby's avatar
Nicolas Delaby committed
733
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
734
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
735 736
    if language is not None:
      kw['language'] = language
737
    result_list = catalog(**kw)
738 739 740 741

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

742
    # if language was given return it (if there are any docs in this language)
743
    if language is not None:
744
      try:
Nicolas Delaby's avatar
Nicolas Delaby committed
745
        return result_list[0].getObject()
746 747
      except IndexError:
        return None
748
    else:
Nicolas Delaby's avatar
Nicolas Delaby committed
749
      first = result_list[0].getObject()
750
      in_original = None
Nicolas Delaby's avatar
Nicolas Delaby committed
751 752 753
      for record in result_list:
        document = record.getObject()
        if document.getVersion() != first.getVersion():
754 755
          # 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
756
            return in_original
757
          else:
Nicolas Delaby's avatar
Nicolas Delaby committed
758 759
            return first # this shouldn't happen in real life
        if document.getLanguage() == user_language:
760
          # we found it in the user language
Nicolas Delaby's avatar
Nicolas Delaby committed
761 762
          return document
        if document.getLanguage() == original_language:
763
          # this is in original language
Nicolas Delaby's avatar
Nicolas Delaby committed
764
          in_original = document
765 766
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
767 768 769 770 771 772 773

  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
774
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
775
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
776
                   reference=self.getReference(),
777
                   sort_on=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
778
                  )
Jérome Perrin's avatar
Jérome Perrin committed
779 780 781 782
    if version:
      kw['version'] = version
    if language:
      kw['language'] = language
783
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
784

785
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
786 787
  def isVersionUnique(self):
    """
788 789 790
      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
791
    """
792
    if not self.getReference():
Jérome Perrin's avatar
Jérome Perrin committed
793
      return True
Nicolas Delaby's avatar
Nicolas Delaby committed
794
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
795
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
796 797 798
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
799
                                            uid=self.getUid(),
800
                                            validation_state="!=cancelled"
801 802 803 804 805
                                            )[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
806
                                            validation_state="!=cancelled"
807 808 809 810
                                            )[0][0]
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

811 812 813 814
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
815 816 817
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
818
    """
Nicolas Delaby's avatar
Nicolas Delaby committed
819 820
    getInfoFor = getToolByName(self.getPortalObject(),
                                                  'portal_workflow').getInfoFor
821
    revision = len(getInfoFor(self, 'history', (), 'edit_workflow'))
822
    # XXX Also look at processing_status_workflow for compatibility.
Nicolas Delaby's avatar
Nicolas Delaby committed
823 824 825
    revision += len([history_item for history_item in\
                 getInfoFor(self, 'history', (), 'processing_status_workflow')\
                 if history_item.get('action') == 'edit'])
826
    return str(revision)
827 828 829 830 831 832 833

  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.
    """
834
    return map(str, range(1, 1 + int(self.getRevision())))
835

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

      NOTE: revision support is implemented in the Document
      class rather than within the ContributionTool
      because the ingestion process requires to analyse the content
      of the document first. Hence, it is not possible to
      do any kind of update operation until the whole ingestion
      process is completed, since update requires to know
      reference, version, language, etc. In addition,
      we have chosen to try to merge revisions after each
      metadata discovery as a way to make sure that any
      content added in the system through the ContributionTool
      (ex. through webdav) will be merged if necessary.
      It may be posssible though to split disoverMetadata and
      finishIngestion.
    """
    document = self
Nicolas Delaby's avatar
Nicolas Delaby committed
857
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
858 859 860 861
    if self.getReference():
      # Find all document with same (reference, version, language)
      kw = dict(portal_type=self.getPortalDocumentTypeList(),
                reference=self.getReference(),
862
                where_expression=SQLQuery("validation_state NOT IN ('cancelled', 'deleted')"))
Jérome Perrin's avatar
Jérome Perrin committed
863 864 865 866
      if self.getVersion():
        kw['version'] = self.getVersion()
      if self.getLanguage():
        kw['language'] = self.getLanguage()
867 868 869 870
      document_list = catalog.unrestrictedSearchResults(**kw)
      existing_document = None
      # Select the first one which is not self and which
      # shares the same coordinates
871
      document_list = list(document_list)
872
      document_list.sort(key=lambda x: x.getId())
873 874 875 876 877 878 879 880 881 882
      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():
883
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
884
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
885
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
886 887 888
        else:
          update_kw = {}
          for k in self.propertyIds():
889
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
890 891 892 893 894 895
              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

896
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
897 898 899 900 901
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
902
    if not self.getReference(): return []
Nicolas Delaby's avatar
Nicolas Delaby committed
903
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
904
    kw = dict(portal_type=self.getPortalType(),
Jérome Perrin's avatar
Jérome Perrin committed
905 906
              reference=self.getReference(),
              group_by=('language',))
907 908 909
    if version is not None:
      kw['version'] = version
    return map(lambda o:o.getLanguage(), catalog(**kw))
Bartek Górny's avatar
Bartek Górny committed
910

911
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
912 913 914
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
915 916

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
917 918 919 920 921
    """
    # 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?
922 923
    reference = self.getReference()
    if not reference:
924
      return
Nicolas Delaby's avatar
Nicolas Delaby committed
925
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
926 927 928
    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
929 930 931 932 933 934

  ### 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')
935
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
936 937 938 939 940
    """
      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:
941 942
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
943
        fallback_script_id='Document_getPropertyDictFromUserLogin')
944
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
945 946 947 948 949 950 951

  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
    """
952 953 954 955 956
    # accesss data through convert
    mime, content = self.convert(None)
    if not content:
       # if document is empty, we will not find anything in its content
      return {}
957
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
958
        fallback_script_id='Document_getPropertyDictFromContent')
959
    return method()
Bartek Górny's avatar
Bartek Górny committed
960 961 962 963 964 965 966 967 968 969 970 971 972 973

  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).
974 975 976 977 978

      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
979
    """
980 981 982
    kw = getattr(self, '_backup_input', {})
    if kw:
      return kw
Bartek Górny's avatar
Bartek Górny committed
983 984
    for id in self.propertyIds():
      # We should not consider file data
985 986
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
987
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
988 989 990 991 992 993 994 995
        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

  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
996
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
997
    """
998 999
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
1000 1001 1002
      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
1003

1004
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
1005

1006
      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
1007
                   currently logged in, then we'll get him from session
1008
    """
Bartek Górny's avatar
Bartek Górny committed
1009
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
1010
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList',
Bartek Górny's avatar
Bartek Górny committed
1011
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
1012
    order_list = list(method())
1013
    order_list.reverse()
1014
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
1015
    kw = {}
1016
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1017
      result = None
Bartek Górny's avatar
Bartek Górny committed
1018 1019
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
1020
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
1021 1022 1023
      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
1024 1025
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
1026
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1027 1028
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
1029 1030
      else:
        result = method()
1031 1032
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1033

1034 1035 1036
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
1037
    # Prepare the content edit parameters - portal_type should not be changed
1038 1039 1040 1041
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
1042 1043
    self.finishIngestion() # XXX - is this really the right place ?
    self.reindexObject() # XXX - is this really the right place ?
1044 1045
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
1046 1047 1048
    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
1049 1050 1051 1052

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1053 1054 1055
      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
1056
    """
1057 1058
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1059

1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
  # 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
    """
    raise NotImplementedError

  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
1084
    subject = self.getSubject('')
1085 1086
    if not subject:
      # XXX not sure if this fallback is a good idea.
Nicolas Delaby's avatar
Nicolas Delaby committed
1087
      subject = self.getTitle('')
1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
    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
1106 1107
        base = '<base href="%s"/>' % self.getContentBaseURL()
        html = html.replace('<head>', '<head>%s' % base, 1)
1108 1109 1110 1111 1112 1113 1114 1115 1116
      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.
    """
1117 1118 1119
    kw['format'] = 'html'
    mime, html = self.convert(**kw)
    return html
1120 1121 1122 1123 1124 1125 1126 1127

  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.
    """
1128
    return self._stripHTML(self._asHTML(**kw))
1129

1130 1131
  security.declarePrivate('_guessEncoding')
  def _guessEncoding(self, string, mime='text/html'):
1132 1133 1134 1135 1136 1137 1138
    """
      Try to guess the encoding for this string.
      Returns None if no encoding can be guessed.
    """
    try:
      import chardet
    except ImportError:
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156
      chardet = None
    if chardet is not None and (mime == 'text/html'\
                                               or os.sys.platform != 'linux2'):
      # chardet works fine on html document and its platform independent
      return chardet.detect(string).get('encoding', None)
    else:
      # file command provide better result
      # for text/plain documents
      # store the content into tempfile
      file_descriptor, path = tempfile.mkstemp()
      file_object = os.fdopen(file_descriptor, 'w')
      file_object.write(string)
      file_object.close()
      # run file command against tempfile to and read encoded
      command_result = Popen(['file', '-b', '--mime-encoding', path],
                                                  stdout=PIPE).communicate()[0]
      # return detected encoding
      return command_result.strip()
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

  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, '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
      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

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

  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 {}

1242
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1243
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
  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
1256 1257 1258 1259 1260
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1261 1262

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1263
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1264 1265 1266
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1267 1268 1269 1270 1271 1272
      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.
1273
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1274 1275
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1276

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1277
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1278
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1279 1280
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1281 1282
      Implementation is handled by ContributionTool.
    """
1283
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1284

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1285 1286
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1287
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1288 1289 1290 1291
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1292 1293 1294 1295 1296 1297
  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
1298
      a difference between URLs which return an index (ex. the
1299 1300 1301 1302 1303 1304 1305 1306 1307
      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
1308 1309 1310 1311 1312 1313 1314 1315
  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
1316
    if len(base_url_list):
1317
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1318
        # Cut the trailing part in http://www.some.site/at/trailing.html
1319
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1320
        base_url = '/'.join(base_url_list[:-1])
1321
    return base_url