Document.py 57 KB
Newer Older
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1 2 3
##############################################################################
#
# Copyright (c) 2002 Nexedi SARL and Contributors. All Rights Reserved.
4
#                    Jean-Paul Smets-Solanes <jp@nexedi.com>
Jean-Paul Smets's avatar
Jean-Paul Smets committed
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
#
# WARNING: This program as such is intended to be used by professional
# programmers who take the whole responsability of assessing all potential
# consequences resulting from its eventual inadequacies and bugs
# End users who are looking for a ready-to-use solution with commercial
# garantees and support are strongly adviced to contract a Free Software
# Service Company
#
# This program is Free Software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
#
##############################################################################

29
import re, socket, sys
Bartek Górny's avatar
Bartek Górny committed
30 31
from DateTime import DateTime
from operator import add
32
from xmlrpclib import Fault
33
from zLOG import LOG, INFO
34
from AccessControl import ClassSecurityInfo, getSecurityManager
35
from AccessControl.SecurityManagement import newSecurityManager, setSecurityManager
Romain Courteaud's avatar
Romain Courteaud committed
36 37
from Acquisition import aq_base
from Globals import PersistentMapping
38
from Globals import get_request
39
from Products.CMFCore.utils import getToolByName, _checkPermission
Jean-Paul Smets's avatar
Jean-Paul Smets committed
40 41
from Products.ERP5Type import Permissions, PropertySheet, Constraint, Interface
from Products.ERP5Type.XMLObject import XMLObject
42
from Products.ERP5Type.DateUtils import convertDateToHour, number_of_hours_in_day, number_of_hours_in_year
43 44
from Products.ERP5Type.Utils import convertToUpperCase
from Products.ERP5Type.Base import WorkflowMethod
45
from Products.ERP5Type.TransactionalVariable import getTransactionalVariable
46 47
from Products.ERP5Type.ExtensibleTraversable import ExtensibleTraversableMixIn
from Products.ERP5Type.Cache import getReadOnlyTransactionCache
48
from Products.ERP5.Document.Url import UrlMixIn
49
from Products.ERP5.Tool.ContributionTool import MAX_REPEAT
50
from AccessControl import Unauthorized
51

Bartek Górny's avatar
Bartek Górny committed
52
_MARKER = []
53
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
54 55
# these property ids are unchangable
FIXED_PROPERTY_IDS =  ('id', 'uid', 'rid', 'sid') 
Jean-Paul Smets's avatar
Jean-Paul Smets committed
56

57 58 59 60 61
def makeSortedTuple(kw):
  items = kw.items()
  items.sort()
  return tuple(items)

62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
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

103
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
104

105 106
class NotConvertedError(Exception):pass

107 108 109 110 111
class ConversionCacheMixin:
  """
    This class provides a generic API to store in the ZODB
    various converted versions of a file or of a string.

Bartek Górny's avatar
Bartek Górny committed
112 113 114 115
    Versions are stored in dictionaries; the class stores also
    generation time of every format and its mime-type string.
    Format can be a string or a tuple (e.g. format, resolution).

116 117 118 119
    TODO:
    * Implement ZODB BLOB
  """
  # time of generation of various formats
Romain Courteaud's avatar
Romain Courteaud committed
120
  _cached_time = None # Defensive programming - prevent caching to RAM
121
  # generated files (cache)
Romain Courteaud's avatar
Romain Courteaud committed
122
  _cached_data = None # Defensive programming - prevent caching to RAM
123
  # mime types for cached formats XXX to be refactored
Romain Courteaud's avatar
Romain Courteaud committed
124
  _cached_mime = None # Defensive programming - prevent caching to RAM
125 126 127 128 129 130 131 132 133 134 135

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

  security.declareProtected(Permissions.ModifyPortalContent, 'clearConversionCache')
  def clearConversionCache(self):
    """
    Clear cache (invoked by interaction workflow upon file upload
    needed here to overwrite class attribute with instance attrs
    """
Romain Courteaud's avatar
Romain Courteaud committed
136 137 138 139 140 141 142
    self._cached_time = PersistentMapping()
    self._cached_data = PersistentMapping()
    self._cached_mime = PersistentMapping()

  security.declareProtected(Permissions.View, 'updateConversionCache')
  def updateConversionCache(self):
    aself = aq_base(self)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
143
    if not hasattr(aself, '_cached_time') or self._cached_time is None:
Romain Courteaud's avatar
Romain Courteaud committed
144
      self._cached_time = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
145
    if not hasattr(aself, '_cached_data') or self._cached_data is None:
Romain Courteaud's avatar
Romain Courteaud committed
146
      self._cached_data = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
147
    if not hasattr(aself, '_cached_mime') or self._cached_mime is None:
Romain Courteaud's avatar
Romain Courteaud committed
148
      self._cached_mime = PersistentMapping()
149 150 151 152 153 154

  security.declareProtected(Permissions.View, 'hasConversion')
  def hasConversion(self, **format):
    """
      Checks whether we have a version in this format
    """
Romain Courteaud's avatar
Romain Courteaud committed
155
    self.updateConversionCache()
156 157
    return self._cached_data.has_key(makeSortedTuple(format))

Bartek Górny's avatar
Bartek Górny committed
158
  security.declareProtected(Permissions.View, 'getCacheTime')
159 160
  def getCacheTime(self, **format):
    """
161
      Checks when if ever was the file produced
162
    """
Romain Courteaud's avatar
Romain Courteaud committed
163
    self.updateConversionCache()
164 165
    return self._cached_time.get(makeSortedTuple(format), 0)

Bartek Górny's avatar
Bartek Górny committed
166
  security.declareProtected(Permissions.ModifyPortalContent, 'updateConversion')
167
  def updateConversion(self, **format):
Romain Courteaud's avatar
Romain Courteaud committed
168 169
    self.updateConversionCache()
    self._cached_time[makeSortedTuple(format)] = DateTime()
170

Bartek Górny's avatar
Bartek Górny committed
171
  security.declareProtected(Permissions.ModifyPortalContent, 'setConversion')
172
  def setConversion(self, data, mime=None, **format):
Bartek Górny's avatar
Bartek Górny committed
173 174 175 176
    """
    Saves a version of the document in a given format; records mime type
    and conversion time (which is right now).
    """
Romain Courteaud's avatar
Romain Courteaud committed
177
    self.updateConversionCache()
178 179 180 181
    tformat = makeSortedTuple(format)
    if mime is not None:
      self._cached_mime[tformat] = mime
    if data is not None:
182 183 184
      self._cached_data[tformat] = aq_base(data) # Use of aq_base 
        # is useful to remove the wrapper from a temp object
        # which may have been used to generate data
185
      self.updateConversion(**format)
186 187
    self._p_changed = 1

Bartek Górny's avatar
Bartek Górny committed
188
  security.declareProtected(Permissions.View, 'getConversion')
189
  def getConversion(self, **format):
Bartek Górny's avatar
Bartek Górny committed
190 191 192 193 194 195 196 197 198
    """
    Returns version of the document in a given format, if it has it; otherwise
    returns empty string (the caller should check hasConversion before calling
    this function.

    (we could be much cooler here - pass testing and updating methods to this function
    so that it does it all by itself; this'd eliminate the need for setConversion public method)
    XXX-BG: I'm not sure now what I meant by this...
    """
Romain Courteaud's avatar
Romain Courteaud committed
199
    self.updateConversionCache()
200 201 202
    tformat = makeSortedTuple(format)
    return self._cached_mime.get(tformat, ''), self._cached_data.get(tformat, '')

203
  security.declareProtected(Permissions.ViewManagementScreens, 'getConversionCacheInfo')
204 205 206
  def getConversionCacheInfo(self):
    """
    Get cache details as string (for debugging)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
207
    """
Romain Courteaud's avatar
Romain Courteaud committed
208
    self.updateConversionCache()
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
    s = 'CACHE INFO:<br/><table><tr><td>format</td><td>size</td><td>time</td><td>is changed</td></tr>'
    for f in self._cached_time.keys():
      t = self._cached_time[f]
      data = self._cached_data.get(f)
      if data:
        if isinstance(data, str):
          ln = len(data)
        else:
          ln = 0
          while data is not None:
            ln += len(data.data)
            data = data.next
      else:
        ln = 'no data!!!'
      s += '<tr><td>%s</td><td>%s</td><td>%s</td><td>%s</td></tr>' % (f, str(ln), str(t), '-')
    s += '</table>'
    return s

227 228 229 230 231 232 233
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
234

235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
  ### 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':
        user_folder = getToolByName(self, 'acl_users', None)
        if user_folder is not None:
          try:
            if request.get('PUBLISHED', _MARKER) is _MARKER:
              # request['PUBLISHED'] is required by validate
              request.other['PUBLISHED'] = self
              has_published = False
            else:
              has_published = True
            user = user_folder.validate(request)
            if not has_published:
              del request.other['PUBLISHED']
          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)


class Document(PermanentURLMixIn, XMLObject, UrlMixIn, ConversionCacheMixin, SnapshotMixin):
Bartek Górny's avatar
Bartek Górny committed
294 295 296 297 298 299
  """
      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.

300 301 302 303 304 305 306 307 308 309
      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).

Jean-Paul Smets's avatar
Jean-Paul Smets committed
310
      External Documents may be downloaded once or updated at
311 312
      regular interval. The later can be useful to update the content
      of an external source. Previous versions may be stored
Jean-Paul Smets's avatar
Jean-Paul Smets committed
313 314 315 316
      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.
317

Bartek Górny's avatar
Bartek Górny committed
318 319 320 321 322 323 324 325 326
      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
327 328 329 330
        methods. TextDocument itself has a subclass
        (XSLTDocument) which provides XSLT based analysis
        and transformation of XML content based on XSLT
        templates. 
Bartek Górny's avatar
Bartek Górny committed
331 332 333

      Document classes which implement conversion should use
      the ConversionCacheMixin class so that converted values are
334
      stored inside ZODB and do not need to be recalculated.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
335 336 337 338 339 340
      More generally, 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.
Bartek Górny's avatar
Bartek Górny committed
341 342 343 344 345 346 347 348 349

      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)
350 351
      (4) other possible actions to finalise the ingestion (ex. by assigning
          a reference)
Bartek Górny's avatar
Bartek Górny committed
352 353 354 355 356

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

      Metadata can be drawn from various sources:

357 358
      input      -   data supplied with http request or set on the object during (2) (e.g.
                     discovered from email text)
359
      file_name  -   data which might be encoded in file name
360 361
      user_login -   information about user who is contributing the file
      content    -   data which might be derived from document content
Bartek Górny's avatar
Bartek Górny committed
362 363 364

      If a certain property is defined in more than one source, it is set according to
      preference order returned by a script 
365 366 367
         Document_getPreferredDocumentMetadataDiscoveryOrderList
         (or any type-based version since discovery is type dependent)

Bartek Górny's avatar
Bartek Górny committed
368
      Methods for discovering metadata are:
369

Bartek Górny's avatar
Bartek Górny committed
370 371 372 373 374
        getPropertyDictFromInput
        getPropertyDictFromFileName
        getPropertyDictFromUserLogin
        getPropertyDictFromContent

375 376 377 378 379
      Methods for processing content are implemented either in 
      Document class or in Base class:

        getSearchableReferenceList (Base)
        getSearchableText (Base)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
380
        index_html (overriden in Document subclasses)
381 382 383 384 385 386 387 388 389 390 391 392 393 394 395

      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).
Bartek Górny's avatar
Bartek Górny committed
396

397 398 399 400 401 402
      document.getSearchableReferenceList() returns
        [
         {'reference':' INV-12367'},
         {'reference': 'INV-1112', 'version':'012}', 
         {'reference': 'AB-CC-DRK', 'version':'011', 'language': 'en'}
        ]
Bartek Górny's avatar
Bartek Górny committed
403

404 405
      The Document class behaviour can be extended / customized through scripts
      (which are type-based so can be adjusted per portal type).
Bartek Górny's avatar
Bartek Górny committed
406

407
      * Document_getPropertyDictFromUserLogin - finds a user (by user_login or from session)
Bartek Górny's avatar
Bartek Górny committed
408 409
        and returns properties which should be set on the document

410
      * Document_getPropertyDictFromContent - analyzes document content and returns
Bartek Górny's avatar
Bartek Górny committed
411 412
        properties which should be set on the document

413
      * Base_getImplicitSuccessorValueList - finds appropriate all documents
414
        referenced in the current content
Bartek Górny's avatar
Bartek Górny committed
415

416
      * Base_getImplicitPredecessorValueList - finds document predecessors based on
Bartek Górny's avatar
Bartek Górny committed
417 418 419 420 421 422 423 424 425
        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)

426
      * Document_getNewRevision - calculates revision number which should be set
Bartek Górny's avatar
Bartek Górny committed
427 428 429
        on this document. Implementation depends on revision numbering policy which
        can be very different. Interaction workflow should call setNewRevision method.

Jean-Paul Smets's avatar
Jean-Paul Smets committed
430 431 432 433
      * 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.

Bartek Górny's avatar
Bartek Górny committed
434
      Subcontent: documents may include subcontent (files, images, etc.)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
435 436 437
      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.
Bartek Górny's avatar
Bartek Górny committed
438 439 440 441 442

    Consistency checking:
      Default implementation uses DocumentReferenceConstraint to check if the 
      reference/language/version triplet is unique. Additional constraints
      can be added if necessary.
443 444 445 446 447 448 449 450 451 452 453 454 455

    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.
Bartek Górny's avatar
Bartek Górny committed
456 457 458 459 460 461 462 463
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
  isPortalContent = 1
  isRADContent = 1
  isDocument = 1
464
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
465

Jean-Paul Smets's avatar
Jean-Paul Smets committed
466 467
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
468 469
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
470
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
471
  charset_parser = re.compile('charset="?([a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
472

Bartek Górny's avatar
Bartek Górny committed
473 474 475 476 477 478 479 480 481 482 483
  # 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
484 485
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
486
                    , PropertySheet.Periodicity
487
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
488 489 490 491 492
                    )

  # Declarative interfaces
  __implements__ = ()

493 494 495
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
                              'subject', 'source_reference', 'source_project_title',)
Bartek Górny's avatar
Bartek Górny committed
496

497
  ### Content processing methods
498
  security.declareProtected(Permissions.View, 'index_html')
499
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
500 501
    """
      We follow here the standard Zope API for files and images
Jean-Paul Smets's avatar
Jean-Paul Smets committed
502 503 504 505 506 507 508 509 510 511 512
      and extend it to support format conversion. The idea
      is that an image which ID is "something.jpg" should
      ne directly accessible through the URL
      /a/b/something.jpg. The same is true for a file and
      for any document type which primary purpose is to
      be used by a helper application rather than displayed
      as HTML in a web browser. Exceptions to this approach
      include Web Pages which are intended to be primarily rendered
      withing the layout of a Web Site or withing a standard ERP5 page.
      Please refer to the index_html of TextDocument.

513 514 515
      Should return appropriate format (calling convert
      if necessary) and set headers.

516
      format -- the format specied in the form of an extension
517
      string (ex. jpeg, html, text, txt, etc.)
518 519

      **kw -- can be various things - e.g. resolution
520

Jean-Paul Smets's avatar
Jean-Paul Smets committed
521 522 523
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
524
    """
525
    raise NotImplementedError
526

Bartek Górny's avatar
Bartek Górny committed
527 528
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
529
    """
530 531 532
      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.
533

534 535
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
536

537
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
538 539
    """
    def getPropertyListOrValue(property):
540
      """
541
        we try to get a list, else we get value and convert to list
542
      """
543 544 545 546 547 548 549
      method = getattr(self, property, None)
      if method is not None:
        if callable(method):
          val = method()
          if isinstance(val, list) or isinstance(val, tuple):
            return list(val)
          return [str(val)]
Bartek Górny's avatar
Bartek Górny committed
550 551 552 553
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
554
          val = [str(val)]
555 556 557
        else:
          val = []
      else:
558
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
559
      return val
560

561
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
562
                                                self.searchable_property_list))
563
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
564 565 566
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
567
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
568

569 570 571 572 573 574 575
  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
576
  ### Relation getters
577
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
578
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
579
    """
580 581 582 583 584
      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.
    """
585
    text = self.getSearchableText() # XXX getSearchableText or asText ?
586
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
587
    try:
588
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
589
    except TypeError: # no regexp in preference
590 591 592
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
593 594 595 596 597 598 599 600 601 602 603 604
    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

605
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
606 607
  def getImplicitSuccessorValueList(self):
    """
608 609 610
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
611

612
      The implementation goes in 2 steps:
613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629

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

    - Step 2: read the list of dictionaries
      and build a list of values by calling portal_catalog
      with appropriate parameters (and if possible build 
      a complex query whenever this becomes available in
      portal catalog)
      
      The script is reponsible for calling getSearchableReferenceList
      so that it can use another approach if needed.
      
      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
630
    """
631
    tv = getTransactionalVariable(self) # XXX Performance improvement required
632 633 634 635 636 637
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

638
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
639
    result = self.Base_getImplicitSuccessorValueList(reference_list)
640 641
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
642

643
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
644 645 646
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
647
      by reference/language etc. Implementation is passed to 
648
        Base_getImplicitPredecessorValueList
649 650 651 652 653 654 655 656 657 658 659 660 661

      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)
      ( if possible build  a complex query whenever 
      this becomes available in portal catalog )

      NOTE: passing a group_by parameter may be useful at a
      later stage of the implementation.
Bartek Górny's avatar
Bartek Górny committed
662
    """
663
    tv = getTransactionalVariable(self) # XXX Performance improvement required
664 665 666 667 668 669
    cache_key = ('getImplicitPredecessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

670 671
    method = self._getTypeBasedMethod('getImplicitPredecessorValueList', 
        fallback_script_id = 'Base_getImplicitPredecessorValueList')
672
    result = method()
673 674
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
675

676
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
677 678 679
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
680
      are similar. Not implemented yet.
Bartek Górny's avatar
Bartek Górny committed
681 682 683 684 685

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

686
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
687
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
688 689 690 691 692 693 694 695
    """
      Returns all documents which are similar to us, directly or indirectly, and
      in both directions. In other words, it is a transitive closure of similar 
      relation. Every document is returned in the latest version available.
    """
    lista = {}
    depth = int(depth)

696 697
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
698

699
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
700
      level += 1
701 702 703 704
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
705 706 707
      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
708 709
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
710

711
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
712 713 714
    lista_latest = {}
    for o in lista.keys():
      lista_latest[o.getLatestVersionValue()] = True # get latest versions avoiding duplicates again
Bartek Górny's avatar
Bartek Górny committed
715 716 717 718
    if lista_latest.has_key(self): 
      lista_latest.pop(self) # remove this document
    if lista_latest.has_key(self.getLatestVersionValue()): 
      lista_latest.pop(self()) # remove this document
Bartek Górny's avatar
Bartek Górny committed
719 720 721

    return lista_latest.keys()

722
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
723 724 725
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
726 727
      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
728

729 730
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
731

732 733 734
      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
735
    """
Bartek Górny's avatar
Bartek Górny committed
736 737
    if not self.getReference():
      return self
738
    catalog = getToolByName(self, 'portal_catalog', None)
739
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
740 741 742 743 744 745
    if language is not None: kw['language'] = language
    res = catalog(**kw)

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

746
    # if language was given return it (if there are any docs in this language)
747
    if language is not None:
748 749 750 751
      try:
        return res[0].getObject()
      except IndexError:
        return None
752 753 754 755 756 757 758
    else:
      first = res[0]
      in_original = None
      for ob in res:
        if ob.getVersion() != first.getVersion():
          # we are out of the latest version - return in_original or first
          if in_original is not None:
Bartek Górny's avatar
Bartek Górny committed
759
            return in_original.getObject()
760
          else:
Bartek Górny's avatar
Bartek Górny committed
761
            return first.getObject() # this shouldn't happen in real life
762 763
        if ob.getLanguage() == user_language:
          # we found it in the user language
Bartek Górny's avatar
Bartek Górny committed
764
          return ob.getObject()
765 766 767
        if ob.getLanguage() == original_language:
          # this is in original language
          in_original = ob
768 769
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
770 771 772 773 774 775 776

  security.declareProtected(Permissions.View, 'getVersionValueList')
  def getVersionValueList(self, version=None, language=None):
    """
      Returns a list of documents with same reference, same portal_type
      but different version and given language or any language if not given.
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
777
    catalog = getToolByName(self, 'portal_catalog', None)
778
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
779
                   reference=self.getReference(),
780
                   order_by=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
781
                  )
782 783 784
    if version: kw['version'] = version
    if language: kw['language'] = language
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
785

786
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
787 788
  def isVersionUnique(self):
    """
789 790 791
      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
792
    """
793 794
    if not self.getReference():
      return 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
795
    catalog = getToolByName(self, 'portal_catalog', None)
796
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
797 798 799
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
800
                                            uid=self.getUid(),
801
                                            validation_state="!=cancelled"
802 803 804 805 806
                                            )[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
807
                                            validation_state="!=cancelled"
808 809 810 811
                                            )[0][0]
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

812 813 814 815
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
816 817 818 819
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
      
820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836
      NOTE: for now, workflow choice is hardcoded. This is
      an optimisation hack. If a document does neither use
      edit_workflow or processing_status_workflow, the
      first workflow in the chain has prioriot. Better
      implementation would require to be able to define
      which workflow in a chain is the default one for
      revision tracking (and for modification date).
    """
    portal_workflow = getToolByName(self, 'portal_workflow')
    wf_list = list(portal_workflow.getWorkflowsFor(self))
    wf = portal_workflow.getWorkflowById('edit_workflow')
    if wf is not None: wf_list = [wf] + wf_list
    wf = portal_workflow.getWorkflowById('processing_status_workflow')
    if wf is not None: wf_list = [wf] + wf_list
    for wf in wf_list:
      history = wf.getInfoFor(self, 'history', None)
      if history:
837 838
        return str(len(filter(lambda x:x.get('action', None) in ('edit', 'upload'), history)))
    return ''
839 840 841 842 843 844 845

  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.
    """
846 847 848 849
    revision = self.getRevision()
    if revision == '':
      return []
    return [str(r) for r in range(0, int(self.getRevision()))]
850

851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876
  security.declareProtected(Permissions.ModifyPortalContent, 'mergeRevision')
  def mergeRevision(self):
    """
      Merge the current document with any previous revision
      or change its version to make sure it is still unique.

      NOTE: revision support is implemented in the Document
      class rather than within the ContributionTool
      because the ingestion process requires to analyse the content
      of the document first. Hence, it is not possible to
      do any kind of update operation until the whole ingestion
      process is completed, since update requires to know
      reference, version, language, etc. In addition,
      we have chosen to try to merge revisions after each
      metadata discovery as a way to make sure that any
      content added in the system through the ContributionTool
      (ex. through webdav) will be merged if necessary.
      It may be posssible though to split disoverMetadata and
      finishIngestion.
    """
    document = self
    catalog = getToolByName(self, 'portal_catalog', None)
    if self.getReference():
      # Find all document with same (reference, version, language)
      kw = dict(portal_type=self.getPortalDocumentTypeList(),
                reference=self.getReference(),
877
                where_expression="validation_state NOT IN ('cancelled', 'deleted')")
878 879 880 881 882 883
      if self.getVersion(): kw['version'] = self.getVersion()
      if self.getLanguage(): kw['language'] = self.getLanguage()
      document_list = catalog.unrestrictedSearchResults(**kw)
      existing_document = None
      # Select the first one which is not self and which
      # shares the same coordinates
884 885 886
      document_list = list(document_list)
      document_list.sort(lambda x,y: cmp(x.getId(), y.getId() ))
      LOG('[DMS] Existing documents for %s' %self.getSourceReference(), INFO, len(document_list))
887 888 889 890 891 892 893 894 895 896
      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():
897
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
898 899
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
          self.setUniqueReference(suffix='unauthorized')
900
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
901 902 903
        else:
          update_kw = {}
          for k in self.propertyIds():
904
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
905 906 907 908 909 910
              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

911
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
912 913 914 915 916
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
917
    if not self.getReference(): return []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
918
    catalog = getToolByName(self, 'portal_catalog', None)
919
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
920
                           reference=self.getReference(),
921 922 923 924
                           group_by=('language',))
    if version is not None:
      kw['version'] = version
    return map(lambda o:o.getLanguage(), catalog(**kw))
Bartek Górny's avatar
Bartek Górny committed
925

926
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
927 928 929
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
930 931

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
932 933 934 935 936
    """
    # 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?
937 938 939 940 941 942 943
    reference = self.getReference()
    if not reference:
      return 
    catalog = getToolByName(self, 'portal_catalog', None)
    res = catalog(reference=self.getReference(), sort_on=(('creation_date','ascending'),))
    # XXX this should be security-unaware - delegate to script with proxy roles
    return res[0].getLanguage() # XXX what happens if it is empty?
Bartek Górny's avatar
Bartek Górny committed
944 945 946 947 948 949

  ### 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')
950
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
951 952 953 954 955
    """
      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:
956 957
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
958
        fallback_script_id='Document_getPropertyDictFromUserLogin')
959
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
960 961 962 963 964 965 966

  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
    """
967
    if not self.hasBaseData():
968
      raise NotConvertedError
969
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
970
        fallback_script_id='Document_getPropertyDictFromContent')
971
    return method()
Bartek Górny's avatar
Bartek Górny committed
972 973 974 975 976 977 978 979 980 981 982 983 984 985

  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).
986 987 988 989 990

      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
991 992 993 994 995 996
    """
    if hasattr(self, '_backup_input'):
      return getattr(self, '_backup_input')
    kw = {}
    for id in self.propertyIds():
      # We should not consider file data
997 998
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
999
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
1000 1001 1002 1003 1004 1005
        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

1006 1007 1008 1009 1010 1011
  security.declareProtected(Permissions.AccessContentsInformation, 'getStandardFileName')
  def getStandardFileName(self):
    """
    Returns the document coordinates as a standard file name. This
    method is the reverse of getPropertyDictFromFileName.
    """
1012 1013 1014
    method = self._getTypeBasedMethod('getStandardFileName', 
        fallback_script_id = 'Document_getStandardFileName')
    return method()
1015

Bartek Górny's avatar
Bartek Górny committed
1016 1017
  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
1018
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
1019
    """
1020 1021
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
1022 1023 1024
      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
1025

1026
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
1027

1028
      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
1029
                   currently logged in, then we'll get him from session
1030
    """
Bartek Górny's avatar
Bartek Górny committed
1031 1032 1033
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList', 
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
1034
    order_list = list(method())
1035
    order_list.reverse()
1036
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
1037
    kw = {}
1038
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1039
      result = None
Bartek Górny's avatar
Bartek Górny committed
1040 1041
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
1042
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
1043 1044 1045
      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
1046 1047
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
1048
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1049 1050
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
1051 1052
      else:
        result = method()
1053 1054
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1055

1056 1057 1058
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
1059
    # Prepare the content edit parameters - portal_type should not be changed
1060 1061 1062 1063
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
1064
    self.finishIngestion()
1065
    self.reindexObject()
1066 1067
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
1068 1069
    merged_doc = self.mergeRevision()
    merged_doc.reindexObject()
1070
    return merged_doc
Bartek Górny's avatar
Bartek Górny committed
1071 1072 1073 1074

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1075 1076 1077
      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
1078
    """
1079 1080
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1081

1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
  # Conversion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'convert')
  def convert(self, format, **kw):
    """
      Main content conversion function, returns result which should
      be returned and stored in cache.
      format - the format specied in the form of an extension
      string (ex. jpeg, html, text, txt, etc.)
      **kw can be various things - e.g. resolution

1092 1093 1094
      Default implementation returns an empty string (html, text)
      or raises an error.

1095 1096 1097 1098
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
    """
1099 1100 1101 1102 1103
    if format == 'html':
      return 'text/html', ''
    if format in ('text', 'txt'):
      return 'text/plain', ''
    raise NotImplementedError
1104 1105 1106 1107 1108 1109

  security.declareProtected(Permissions.View, 'asText')
  def asText(self):
    """
      Converts the content of the document to a textual representation.
    """
1110
    mime, data = self.convert(format='txt')
1111
    return str(data)
1112

1113 1114
  security.declareProtected(Permissions.View, 'asEntireHTML')
  def asEntireHTML(self):
1115 1116
    """
      Returns a complete HTML representation of the document
1117 1118 1119
      (with body tags, etc.). Adds if necessary a base
      tag so that the document can be displayed in an iframe
      or standalone.
1120 1121

      Actual conversion is delegated to _asHTML
1122
    """
1123
    html = self._asHTML()
1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134
    if self.getUrlString():
      # If a URL is defined, add the base tag
      # if base is defined yet.
      html = str(html)
      if not html.find('<base') >= 0:
        base = '<base href="%s">' % self.getContentBaseURL()
        html = html.replace('<head>', '<head>%s' % base)
      # We do not implement cache yet since it increases ZODB
      # for probably no reason. More research needed
      # self.setConversion(html, mime='text/html', format='base-html')
    return html
1135

1136 1137 1138 1139 1140 1141
  security.declarePrivate('_asHTML')
  def _asHTML(self):
    """
      A private method which converts to HTML. This method
      is the one to override in subclasses.
    """
1142
    if not self.hasBaseData():
1143
      raise ConversionError('This document has not been processed yet.')
1144 1145 1146 1147 1148 1149
    if self.hasConversion(format='base-html'):
      mime, data = self.getConversion(format='base-html')
      return data
    mime, html = self.convert(format='html')
    return html

1150 1151 1152
  security.declareProtected(Permissions.View, 'asStrippedHTML')
  def asStrippedHTML(self):
    """
1153 1154
      Returns a stripped HTML representation of the document
      (without html and body tags, etc.) which can be used to inline
1155 1156
      a preview of the document.
    """
1157
    if not self.hasBaseData():
1158
      return ''
1159
    if self.hasConversion(format='stripped-html'): # XXX this is redundant since we never set it
1160 1161 1162 1163 1164
      mime, data = self.getConversion(format='stripped-html')
      return data
    mime, html = self.convert(format='html')
    body_list = re.findall(self.body_parser, str(html))
    if len(body_list):
1165 1166 1167 1168
      stripped_html = body_list[0]
    else:
      stripped_html = html
    # find charset and convert to utf-8
1169 1170
    charset_list = self.charset_parser.findall(str(html)) # XXX - Not efficient is datastream 
                                                          # instance but hard to do better
1171
    if charset_list and charset_list[0] not in ('utf-8', 'UTF-8'):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1172
      stripped_html = unicode(str(stripped_html), charset_list[0]).encode('utf-8')
1173
    return stripped_html
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183

  security.declareProtected(Permissions.AccessContentsInformation, 'getContentInformation')
  def getContentInformation(self):
    """
    Returns the content information from the HTML conversion.
    The default implementation tries to build a dictionnary
    from the HTML conversion of the document and extract
    the document title.
    """
    result = {}
Yusei Tahara's avatar
Yusei Tahara committed
1184
    html = self.asEntireHTML()
1185 1186 1187 1188 1189
    if not html: return result
    title_list = re.findall(self.title_parser, str(html))
    if title_list:
      result['title'] = title_list[0]
    return result
1190 1191

  # Base format support
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1192
  security.declareProtected(Permissions.ModifyPortalContent, 'convertToBaseFormat')
1193
  def convertToBaseFormat(self):
1194
    """
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
      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.)
1210 1211
    """
    try:
1212
      message = self._convertToBaseFormat() # Call implemetation method
1213
      self.clearConversionCache() # Conversion cache is now invalid
1214 1215 1216 1217 1218 1219 1220
      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
1221

1222
  def _convertToBaseFormat(self):
1223
    """
1224 1225 1226 1227 1228
      Placeholder method. Must be subclassed by classes
      which need a base format. Refer to OOoDocument
      for an example of ODF base format which is used
      as a way to convert about any file format into
      about any file format.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1229

1230
      Other possible applications: conversion of HTML
1231
      text to tiddy HTML such as described here:
1232 1233
      http://www.xml.com/pub/a/2004/09/08/pyxml.html
      so that resulting text can be processed more
1234 1235 1236
      easily by XSLT parsers. Conversion of internal
      links to images of an HTML document to local
      links (in combindation with populate).
1237
    """
1238
    raise NotImplementedError
1239

1240 1241 1242 1243 1244 1245 1246 1247
  security.declareProtected(Permissions.ModifyPortalContent, 'isSupportBaseDataConversion')
  def isSupportBaseDataConversion(self):
    """
    This is a public interface to check a document that is support conversion
    to base format and can be overridden in subclasses.
    """
    return False

1248
  def convertFile(self, **kw):
1249 1250 1251 1252 1253
    """
    Workflow transition invoked when conversion occurs.
    """
  convertFile = WorkflowMethod(convertFile)

1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
  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 {}

1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
  security.declareProtected(Permissions.ModifyPortalContent, 'updateBaseMetadata')
  def updateBaseMetadata(self, **kw):
    """
    Update the base format data with the latest properties entered
    by the user. For example, if title is changed in ERP5 interface,
    the base format file should be updated accordingly.

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

1282
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1283
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295
  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
1296 1297 1298 1299 1300
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1301 1302

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1303
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1304 1305 1306
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1307 1308 1309 1310 1311 1312
      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.
1313
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1314 1315
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1316

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1317
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1318
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1319 1320
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1321 1322
      Implementation is handled by ContributionTool.
    """
1323
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1324

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1325 1326
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1327
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1328 1329 1330 1331
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
  security.declareProtected(Permissions.View, 'isIndexContent')
  def isIndexContent(self, container=None):
    """
      Ask container if we are and index, or a content.
      In the vast majority of cases we are content.
      This method is required in a crawling process to make
      a difference between URLs which return an index (ex. the 
      list of files in remote server which is accessed through HTTP)
      and the files themselves.
    """
    if container is None:
      container = self.getParentValue()
    if hasattr(aq_base(container), 'isIndexContent'):
      return container.isIndexContent(self)
    return False

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1348 1349 1350 1351 1352 1353 1354 1355
  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
1356
    if len(base_url_list):
1357
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1358
        # Cut the trailing part in http://www.some.site/at/trailing.html
1359
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1360
        base_url = '/'.join(base_url_list[:-1])
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1361 1362
    return base_url

1363 1364
  security.declareProtected(Permissions.AccessContentsInformation, 'getFrequencyIndex')
  def getFrequencyIndex(self):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1365
    """
1366 1367 1368 1369
      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).
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1370
    """
1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
    try:
      return self.getUpdateFrequencyValue().getIntIndex()
    except AttributeError:
      # Catch Attribute error or Key error - XXX not beautiful
      return 0

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

    return creation_date_index

  security.declareProtected(Permissions.AccessContentsInformation, 'isUpdatable')
  def isUpdatable(self):
    """
      This method is used to decide which document can be updated
      in the crawling process. This can depend for example on
      workflow states (publication state,
      validation state) or on roles on the document.
    """
    method = self._getTypeBasedMethod('isUpdatable', 
        fallback_script_id = 'Document_isUpdatable')
Yusei Tahara's avatar
Yusei Tahara committed
1405
    return method()