Document.py 54.6 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
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
Romain Courteaud's avatar
Romain Courteaud committed
35 36
from Acquisition import aq_base
from Globals import PersistentMapping
37
from Products.CMFCore.utils import getToolByName, _checkPermission
Jean-Paul Smets's avatar
Jean-Paul Smets committed
38 39
from Products.ERP5Type import Permissions, PropertySheet, Constraint, Interface
from Products.ERP5Type.XMLObject import XMLObject
40
from Products.ERP5Type.DateUtils import convertDateToHour, number_of_hours_in_day, number_of_hours_in_year
41 42
from Products.ERP5Type.Utils import convertToUpperCase
from Products.ERP5Type.Base import WorkflowMethod
43
from Products.ERP5Type.TransactionalVariable import getTransactionalVariable
44
from Products.ERP5.Document.Url import UrlMixIn
45
from Products.ERP5.Tool.ContributionTool import MAX_REPEAT
46
from AccessControl import Unauthorized
47

Bartek Górny's avatar
Bartek Górny committed
48
_MARKER = []
49
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
50 51
# these property ids are unchangable
FIXED_PROPERTY_IDS =  ('id', 'uid', 'rid', 'sid') 
Jean-Paul Smets's avatar
Jean-Paul Smets committed
52

53 54 55 56 57
def makeSortedTuple(kw):
  items = kw.items()
  items.sort()
  return tuple(items)

58 59 60 61 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
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

99
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
100

101 102 103 104 105
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
106 107 108 109
    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).

110 111 112 113
    TODO:
    * Implement ZODB BLOB
  """
  # time of generation of various formats
Romain Courteaud's avatar
Romain Courteaud committed
114
  _cached_time = None # Defensive programming - prevent caching to RAM
115
  # generated files (cache)
Romain Courteaud's avatar
Romain Courteaud committed
116
  _cached_data = None # Defensive programming - prevent caching to RAM
117
  # mime types for cached formats XXX to be refactored
Romain Courteaud's avatar
Romain Courteaud committed
118
  _cached_mime = None # Defensive programming - prevent caching to RAM
119 120 121 122 123 124 125 126 127 128 129

  # 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
130 131 132 133 134 135 136
    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
137
    if not hasattr(aself, '_cached_time') or self._cached_time is None:
Romain Courteaud's avatar
Romain Courteaud committed
138
      self._cached_time = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
139
    if not hasattr(aself, '_cached_data') or self._cached_data is None:
Romain Courteaud's avatar
Romain Courteaud committed
140
      self._cached_data = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
141
    if not hasattr(aself, '_cached_mime') or self._cached_mime is None:
Romain Courteaud's avatar
Romain Courteaud committed
142
      self._cached_mime = PersistentMapping()
143 144 145 146 147 148

  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
149
    self.updateConversionCache()
150 151
    return self._cached_data.has_key(makeSortedTuple(format))

Bartek Górny's avatar
Bartek Górny committed
152
  security.declareProtected(Permissions.View, 'getCacheTime')
153 154
  def getCacheTime(self, **format):
    """
155
      Checks when if ever was the file produced
156
    """
Romain Courteaud's avatar
Romain Courteaud committed
157
    self.updateConversionCache()
158 159
    return self._cached_time.get(makeSortedTuple(format), 0)

Bartek Górny's avatar
Bartek Górny committed
160
  security.declareProtected(Permissions.ModifyPortalContent, 'updateConversion')
161
  def updateConversion(self, **format):
Romain Courteaud's avatar
Romain Courteaud committed
162 163
    self.updateConversionCache()
    self._cached_time[makeSortedTuple(format)] = DateTime()
164

Bartek Górny's avatar
Bartek Górny committed
165
  security.declareProtected(Permissions.ModifyPortalContent, 'setConversion')
166
  def setConversion(self, data, mime=None, **format):
Bartek Górny's avatar
Bartek Górny committed
167 168 169 170
    """
    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
171
    self.updateConversionCache()
172 173 174 175 176
    tformat = makeSortedTuple(format)
    if mime is not None:
      self._cached_mime[tformat] = mime
    if data is not None:
      self._cached_data[tformat] = data
177
      self.updateConversion(**format)
178 179
    self._p_changed = 1

Bartek Górny's avatar
Bartek Górny committed
180
  security.declareProtected(Permissions.View, 'getConversion')
181
  def getConversion(self, **format):
Bartek Górny's avatar
Bartek Górny committed
182 183 184 185 186 187 188 189 190
    """
    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
191
    self.updateConversionCache()
192 193 194
    tformat = makeSortedTuple(format)
    return self._cached_mime.get(tformat, ''), self._cached_data.get(tformat, '')

195
  security.declareProtected(Permissions.ViewManagementScreens, 'getConversionCacheInfo')
196 197 198
  def getConversionCacheInfo(self):
    """
    Get cache details as string (for debugging)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
199
    """
Romain Courteaud's avatar
Romain Courteaud committed
200
    self.updateConversionCache()
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218
    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

Bartek Górny's avatar
Bartek Górny committed
219

220
class Document(XMLObject, UrlMixIn, ConversionCacheMixin, SnapshotMixin):
Bartek Górny's avatar
Bartek Górny committed
221 222 223 224 225 226
  """
      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.

227 228 229 230 231 232 233 234 235 236
      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
237
      External Documents may be downloaded once or updated at
238 239
      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
240 241 242 243
      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.
244

Bartek Górny's avatar
Bartek Górny committed
245 246 247 248 249 250 251 252 253
      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
254 255 256 257
        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
258 259 260

      Document classes which implement conversion should use
      the ConversionCacheMixin class so that converted values are
261
      stored inside ZODB and do not need to be recalculated.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
262 263 264 265 266 267
      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
268 269 270 271 272 273 274 275 276

      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)
277 278
      (4) other possible actions to finalise the ingestion (ex. by assigning
          a reference)
Bartek Górny's avatar
Bartek Górny committed
279 280 281 282 283

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

      Metadata can be drawn from various sources:

284 285
      input      -   data supplied with http request or set on the object during (2) (e.g.
                     discovered from email text)
286
      file_name  -   data which might be encoded in file name
287 288
      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
289 290 291

      If a certain property is defined in more than one source, it is set according to
      preference order returned by a script 
292 293 294
         Document_getPreferredDocumentMetadataDiscoveryOrderList
         (or any type-based version since discovery is type dependent)

Bartek Górny's avatar
Bartek Górny committed
295
      Methods for discovering metadata are:
296

Bartek Górny's avatar
Bartek Górny committed
297 298 299 300 301
        getPropertyDictFromInput
        getPropertyDictFromFileName
        getPropertyDictFromUserLogin
        getPropertyDictFromContent

302 303 304 305 306
      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
307
        index_html (overriden in Document subclasses)
308 309 310 311 312 313 314 315 316 317 318 319 320 321 322

      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
323

324 325 326 327 328 329
      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
330

331 332
      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
333

334
      * Document_getPropertyDictFromUserLogin - finds a user (by user_login or from session)
Bartek Górny's avatar
Bartek Górny committed
335 336
        and returns properties which should be set on the document

337
      * Document_getPropertyDictFromContent - analyzes document content and returns
Bartek Górny's avatar
Bartek Górny committed
338 339
        properties which should be set on the document

340
      * Base_getImplicitSuccessorValueList - finds appropriate all documents
341
        referenced in the current content
Bartek Górny's avatar
Bartek Górny committed
342

343
      * Base_getImplicitPredecessorValueList - finds document predecessors based on
Bartek Górny's avatar
Bartek Górny committed
344 345 346 347 348 349 350 351 352
        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)

353
      * Document_getNewRevision - calculates revision number which should be set
Bartek Górny's avatar
Bartek Górny committed
354 355 356
        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
357 358 359 360
      * 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
361
      Subcontent: documents may include subcontent (files, images, etc.)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
362 363 364
      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
365 366 367 368 369

    Consistency checking:
      Default implementation uses DocumentReferenceConstraint to check if the 
      reference/language/version triplet is unique. Additional constraints
      can be added if necessary.
370 371 372 373 374 375 376 377 378 379 380 381 382

    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
383 384 385 386 387 388 389 390
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
  isPortalContent = 1
  isRADContent = 1
  isDocument = 1
391
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
392

Jean-Paul Smets's avatar
Jean-Paul Smets committed
393 394
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
395 396
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
397
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
398
  charset_parser = re.compile('charset="?([a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
399

Bartek Górny's avatar
Bartek Górny committed
400 401 402 403 404 405 406 407 408 409 410
  # 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
411 412
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
413
                    , PropertySheet.Periodicity
414
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
415 416 417 418 419
                    )

  # Declarative interfaces
  __implements__ = ()

420 421 422
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
                              'subject', 'source_reference', 'source_project_title',)
Bartek Górny's avatar
Bartek Górny committed
423

424
  ### Content processing methods
425
  security.declareProtected(Permissions.View, 'index_html')
426
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
427 428
    """
      We follow here the standard Zope API for files and images
Jean-Paul Smets's avatar
Jean-Paul Smets committed
429 430 431 432 433 434 435 436 437 438 439
      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.

440 441 442
      Should return appropriate format (calling convert
      if necessary) and set headers.

443
      format -- the format specied in the form of an extension
444
      string (ex. jpeg, html, text, txt, etc.)
445 446

      **kw -- can be various things - e.g. resolution
447

Jean-Paul Smets's avatar
Jean-Paul Smets committed
448 449 450
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
451
    """
452
    raise NotImplementedError
453

Bartek Górny's avatar
Bartek Górny committed
454 455
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
456
    """
457 458 459
      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.
460

461 462
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
463

464
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
465 466
    """
    def getPropertyListOrValue(property):
467
      """
468
        we try to get a list, else we get value and convert to list
469
      """
470 471 472 473 474 475 476
      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
477 478 479 480
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
481
          val = [str(val)]
482 483 484
        else:
          val = []
      else:
485
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
486
      return val
487

488
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
489
                                                self.searchable_property_list))
490
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
491 492 493
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
494
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
495

496 497 498 499 500 501 502
  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
503
  ### Relation getters
504
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
505
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
506
    """
507 508 509 510 511
      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.
    """
512
    text = self.getSearchableText() # XXX getSearchableText or asText ?
513
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
514
    try:
515
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
516
    except TypeError: # no regexp in preference
517 518 519
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
520
    res = rx_search.finditer(text)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
521
    res = [(r.group(), r.groupdict()) for r in res]
Bartek Górny's avatar
Bartek Górny committed
522
    return res
523
    
524
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
525 526
  def getImplicitSuccessorValueList(self):
    """
527 528 529
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
530

531
      The implementation goes in 2 steps:
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548

    - 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
549
    """
550 551 552 553 554 555 556
    tv = getTransactionalVariable(self)
    cache_key = ('getImplicitSuccessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

557
    reference_list = [r[1] for r in self.getSearchableReferenceList()]
558
    successor_list = self.Base_getImplicitSuccessorValueList(reference_list)
559
    # get unique latest (most relevant) versions
560 561 562 563 564
    temp_result = [r.getObject().getLatestVersionValue() for r in successor_list]
    result_dict = dict.fromkeys(temp_result)
    result = result_dict.keys()
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
565

566
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
567 568 569
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
570
      by reference/language etc. Implementation is passed to 
571
        Base_getImplicitPredecessorValueList
572 573 574 575 576 577 578 579 580 581 582 583 584

      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
585
    """
586 587 588 589 590 591 592
    tv = getTransactionalVariable(self)
    cache_key = ('getImplicitPredecessorValueList', self.getPhysicalPath())
    try:
      return tv[cache_key]
    except KeyError:
      pass

593 594
    method = self._getTypeBasedMethod('getImplicitPredecessorValueList', 
        fallback_script_id = 'Base_getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
595
    lst = method()
596 597 598
    # make it unique first time (before getting lastversionvalue)
    di = dict.fromkeys([r.getObject() for r in lst])
    # then get latest version and make unique again
Yusei Tahara's avatar
Yusei Tahara committed
599 600 601
    di = dict.fromkeys([o.getLatestVersionValue()
                        for o in di.keys()
                        if getattr(o, 'getLatestVersionValue', None)])
Bartek Górny's avatar
Bartek Górny committed
602
    ref = self.getReference()
603 604 605
    result = [o for o in di.keys() if o.getReference() != ref] # every object has its own reference in SearchableText
    tv[cache_key] = result
    return result
Bartek Górny's avatar
Bartek Górny committed
606

607
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
608 609 610 611 612 613 614 615 616
  def getImplicitSimilarValueList(self):
    """
      Analyses content of documents to find out by the content which documents
      are similar. Not implemented yet. 

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

617
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
618
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
619 620 621 622 623 624 625 626
    """
      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)

627 628
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
629

630
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
631
      level += 1
632 633 634 635
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
636 637 638
      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
639 640
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
641

642
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
643 644 645
    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
646 647 648 649
    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
650 651 652

    return lista_latest.keys()

653
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
654 655 656
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
657 658
      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
659

660 661
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
662

663 664 665
      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
666
    """
Bartek Górny's avatar
Bartek Górny committed
667 668
    if not self.getReference():
      return self
669
    catalog = getToolByName(self, 'portal_catalog', None)
670
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
671 672 673 674 675 676
    if language is not None: kw['language'] = language
    res = catalog(**kw)

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

677
    # if language was given return it (if there are any docs in this language)
678
    if language is not None:
679 680 681 682
      try:
        return res[0].getObject()
      except IndexError:
        return None
683 684 685 686 687 688 689
    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
690
            return in_original.getObject()
691
          else:
Bartek Górny's avatar
Bartek Górny committed
692
            return first.getObject() # this shouldn't happen in real life
693 694
        if ob.getLanguage() == user_language:
          # we found it in the user language
Bartek Górny's avatar
Bartek Górny committed
695
          return ob.getObject()
696 697 698
        if ob.getLanguage() == original_language:
          # this is in original language
          in_original = ob
699 700
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
701 702 703 704 705 706 707

  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
708
    catalog = getToolByName(self, 'portal_catalog', None)
709
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
710
                   reference=self.getReference(),
711
                   order_by=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
712
                  )
713 714 715
    if version: kw['version'] = version
    if language: kw['language'] = language
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
716

717
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
718 719
  def isVersionUnique(self):
    """
720 721 722
      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
723
    """
724 725
    if not self.getReference():
      return 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
726
    catalog = getToolByName(self, 'portal_catalog', None)
727
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
728 729 730
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
731
                                            uid=self.getUid(),
732
                                            validation_state="!=cancelled"
733 734 735 736 737
                                            )[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
738
                                            validation_state="!=cancelled"
739 740 741 742 743 744 745 746 747 748 749
                                            )[0][0]
    # If self is not indexed yet, then if count == 1, version is not unique
    return count <= self_count

  security.declareProtected(Permissions.ModifyPortalContent, 'setUniqueReference')
  def setUniqueReference(self, suffix='auto'):
    """
      Create a unique reference for the current document
      based on a suffix
    """
    # Change the document reference
750
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
751 752 753 754 755 756 757 758 759
    reference = self.getReference() + '-%s-' % suffix + '%s'
    ref_count = 0
    kw = dict(portal_type=self.getPortalDocumentTypeList())
    if self.getVersion(): kw['version'] = self.getVersion()
    if self.getLanguage(): kw['language'] = self.getLanguage()
    while catalog.unrestrictedCountResults(reference=reference % ref_count, **kw)[0][0]:
      ref_count += 1
    self._setReference(reference % ref_count)
  
760 761 762 763
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
764 765 766 767
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
      
768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784
      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:
785 786
        return str(len(filter(lambda x:x.get('action', None) in ('edit', 'upload'), history)))
    return ''
787 788 789 790 791 792 793

  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.
    """
794 795 796 797
    revision = self.getRevision()
    if revision == '':
      return []
    return [str(r) for r in range(0, int(self.getRevision()))]
798

799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
  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(),
825
                where_expression="validation_state NOT IN ('cancelled', 'deleted')")
826 827 828 829 830 831
      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
832 833 834
      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))
835 836 837 838 839 840 841 842 843 844
      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():
845
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
846 847
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
          self.setUniqueReference(suffix='unauthorized')
848
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
849 850 851
        else:
          update_kw = {}
          for k in self.propertyIds():
852
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
853 854 855 856 857 858
              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

859
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
860 861 862 863 864
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
865
    if not self.getReference(): return []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
866
    catalog = getToolByName(self, 'portal_catalog', None)
867
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
868
                           reference=self.getReference(),
869 870 871 872
                           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
873

874
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
875 876 877
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
878 879

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
880 881 882 883 884
    """
    # 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?
885 886 887 888 889 890 891
    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
892 893 894 895 896 897

  ### 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')
898
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
899 900 901 902 903
    """
      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:
904 905
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
906
        fallback_script_id='Document_getPropertyDictFromUserLogin')
907
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
908 909 910 911 912 913 914

  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
    """
915 916
    if not self.hasBaseData():
      self.convertToBaseFormat()
917
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
918
        fallback_script_id='Document_getPropertyDictFromContent')
919
    return method()
Bartek Górny's avatar
Bartek Górny committed
920 921 922 923 924 925 926 927 928 929 930 931 932 933

  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).
934 935 936 937 938

      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
939 940 941 942 943 944
    """
    if hasattr(self, '_backup_input'):
      return getattr(self, '_backup_input')
    kw = {}
    for id in self.propertyIds():
      # We should not consider file data
945 946
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
947
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
948 949 950 951 952 953
        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

954 955 956 957 958 959
  security.declareProtected(Permissions.AccessContentsInformation, 'getStandardFileName')
  def getStandardFileName(self):
    """
    Returns the document coordinates as a standard file name. This
    method is the reverse of getPropertyDictFromFileName.
    """
960 961 962
    method = self._getTypeBasedMethod('getStandardFileName', 
        fallback_script_id = 'Document_getStandardFileName')
    return method()
963

Bartek Górny's avatar
Bartek Górny committed
964 965
  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
966
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
967
    """
968 969
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
970 971 972
      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
973

974
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
975

976
      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
977
                   currently logged in, then we'll get him from session
978
    """
Bartek Górny's avatar
Bartek Górny committed
979 980 981
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList', 
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
982
    order_list = list(method())
983
    order_list.reverse()
984
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
985
    kw = {}
986
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
987
      result = None
Bartek Górny's avatar
Bartek Górny committed
988 989
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
990
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
991 992 993
      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
994 995
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
996
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
997 998
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
999 1000
      else:
        result = method()
1001 1002
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1003

1004 1005 1006
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
1007
    # Prepare the content edit parameters - portal_type should not be changed
1008 1009 1010 1011
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
1012
    self.finishIngestion()
1013
    self.reindexObject()
1014 1015
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
1016 1017
    merged_doc = self.mergeRevision()
    merged_doc.reindexObject()
1018
    return merged_doc
Bartek Górny's avatar
Bartek Górny committed
1019 1020 1021 1022

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1023 1024 1025
      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
1026
    """
1027 1028
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1029

1030 1031 1032 1033 1034 1035 1036 1037 1038 1039
  # 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

1040 1041 1042
      Default implementation returns an empty string (html, text)
      or raises an error.

1043 1044 1045 1046
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
    """
1047 1048 1049 1050 1051
    if format == 'html':
      return 'text/html', ''
    if format in ('text', 'txt'):
      return 'text/plain', ''
    raise NotImplementedError
1052 1053 1054 1055 1056 1057

  security.declareProtected(Permissions.View, 'asText')
  def asText(self):
    """
      Converts the content of the document to a textual representation.
    """
1058 1059
    mime, data = self.convert(format='txt')
    return data
1060

1061 1062
  security.declareProtected(Permissions.View, 'asEntireHTML')
  def asEntireHTML(self):
1063 1064
    """
      Returns a complete HTML representation of the document
1065 1066 1067
      (with body tags, etc.). Adds if necessary a base
      tag so that the document can be displayed in an iframe
      or standalone.
1068 1069

      Actual conversion is delegated to _asHTML
1070
    """
1071
    html = self._asHTML()
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
    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
1083

1084 1085 1086 1087 1088 1089
  security.declarePrivate('_asHTML')
  def _asHTML(self):
    """
      A private method which converts to HTML. This method
      is the one to override in subclasses.
    """
1090
    if not self.hasBaseData():
1091
      raise ConversionError('This document has not been processed yet.')
1092 1093 1094 1095 1096 1097
    if self.hasConversion(format='base-html'):
      mime, data = self.getConversion(format='base-html')
      return data
    mime, html = self.convert(format='html')
    return html

1098 1099 1100
  security.declareProtected(Permissions.View, 'asStrippedHTML')
  def asStrippedHTML(self):
    """
1101 1102
      Returns a stripped HTML representation of the document
      (without html and body tags, etc.) which can be used to inline
1103 1104
      a preview of the document.
    """
1105
    if not self.hasBaseData():
1106
      return ''
1107
    if self.hasConversion(format='stripped-html'): # XXX this is redundant since we never set it
1108 1109 1110 1111 1112
      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):
1113 1114 1115 1116
      stripped_html = body_list[0]
    else:
      stripped_html = html
    # find charset and convert to utf-8
1117 1118
    charset_list = self.charset_parser.findall(str(html)) # XXX - Not efficient is datastream 
                                                          # instance but hard to do better
1119
    if charset_list and charset_list[0] not in ('utf-8', 'UTF-8'):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1120
      stripped_html = unicode(str(stripped_html), charset_list[0]).encode('utf-8')
1121
    return stripped_html
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131

  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
1132
    html = self.asEntireHTML()
1133 1134 1135 1136 1137
    if not html: return result
    title_list = re.findall(self.title_parser, str(html))
    if title_list:
      result['title'] = title_list[0]
    return result
1138 1139

  # Base format support
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1140
  security.declareProtected(Permissions.ModifyPortalContent, 'convertToBaseFormat')
1141
  def convertToBaseFormat(self):
1142
    """
1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
      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.)
1158 1159
    """
    try:
1160 1161
      msg = self._convertToBaseFormat() # Call implemetation method
      self.clearConversionCache() # Conversion cache is now invalid
1162
      if msg is None:
1163
        msg = 'Converted to %s.' % self.getBaseContentType()
1164 1165 1166 1167 1168
      self.convertFile(comment=msg) # Invoke workflow method
    except NotImplementedError:# we don't do any workflow action if nothing has been done
      msg = '' 
    except ConversionError, e:
      msg = 'Problem: %s' % (str(e) or 'undefined.')
1169 1170
      #self.processFile(comment=msg)
      raise ConversionError, msg
1171 1172
    except Fault, e:
      msg = 'Problem: %s' % (repr(e) or 'undefined.')
1173 1174
      #self.processFile(comment=msg)
      raise Fault, msg
1175 1176
    except socket.error, e:
      msg = 'Problem: %s' % (repr(e) or 'undefined.')
1177 1178
      #self.processFile(comment=msg)
      raise socket.error, msg
1179 1180
    return msg

1181
  def _convertToBaseFormat(self):
1182
    """
1183 1184 1185 1186 1187
      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
1188

1189
      Other possible applications: conversion of HTML
1190
      text to tiddy HTML such as described here:
1191 1192
      http://www.xml.com/pub/a/2004/09/08/pyxml.html
      so that resulting text can be processed more
1193 1194 1195
      easily by XSLT parsers. Conversion of internal
      links to images of an HTML document to local
      links (in combindation with populate).
1196
    """
1197
    raise NotImplementedError
1198

1199
  def convertFile(self, **kw):
1200 1201 1202 1203 1204
    """
    Workflow transition invoked when conversion occurs.
    """
  convertFile = WorkflowMethod(convertFile)

1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
  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

1217
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1218
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
  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
1231 1232 1233 1234 1235
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1236 1237

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1238
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1239 1240 1241
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1242 1243 1244 1245 1246 1247
      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.
1248
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1249 1250
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1251

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1252
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1253
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1254 1255
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1256 1257
      Implementation is handled by ContributionTool.
    """
1258
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1259

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1260 1261
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1262
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1263 1264 1265 1266
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282
  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
1283 1284 1285 1286 1287 1288 1289 1290
  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
1291
    if len(base_url_list):
1292
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1293
        # Cut the trailing part in http://www.some.site/at/trailing.html
1294
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1295
        base_url = '/'.join(base_url_list[:-1])
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1296 1297
    return base_url

1298 1299
  security.declareProtected(Permissions.AccessContentsInformation, 'getFrequencyIndex')
  def getFrequencyIndex(self):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1300
    """
1301 1302 1303 1304
      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
1305
    """
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
    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
1340
    return method()