Document.py 53.9 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.ERP5.Document.Url import UrlMixIn
44
from Products.ERP5.Tool.ContributionTool import MAX_REPEAT
45
from AccessControl import Unauthorized
46

Bartek Górny's avatar
Bartek Górny committed
47
_MARKER = []
48
VALID_ORDER_KEY_LIST = ('user_login', 'content', 'file_name', 'input')
49 50
# these property ids are unchangable
FIXED_PROPERTY_IDS =  ('id', 'uid', 'rid', 'sid') 
Jean-Paul Smets's avatar
Jean-Paul Smets committed
51

52 53 54 55 56
def makeSortedTuple(kw):
  items = kw.items()
  items.sort()
  return tuple(items)

57 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
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

98
class ConversionError(Exception):pass
Bartek Górny's avatar
Bartek Górny committed
99

100 101 102 103 104
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
105 106 107 108
    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).

109 110 111 112
    TODO:
    * Implement ZODB BLOB
  """
  # time of generation of various formats
Romain Courteaud's avatar
Romain Courteaud committed
113
  _cached_time = None # Defensive programming - prevent caching to RAM
114
  # generated files (cache)
Romain Courteaud's avatar
Romain Courteaud committed
115
  _cached_data = None # Defensive programming - prevent caching to RAM
116
  # mime types for cached formats XXX to be refactored
Romain Courteaud's avatar
Romain Courteaud committed
117
  _cached_mime = None # Defensive programming - prevent caching to RAM
118 119 120 121 122 123 124 125 126 127 128

  # 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
129 130 131 132 133 134 135
    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
136
    if not hasattr(aself, '_cached_time') or self._cached_time is None:
Romain Courteaud's avatar
Romain Courteaud committed
137
      self._cached_time = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
138
    if not hasattr(aself, '_cached_data') or self._cached_data is None:
Romain Courteaud's avatar
Romain Courteaud committed
139
      self._cached_data = PersistentMapping()
Jean-Paul Smets's avatar
Jean-Paul Smets committed
140
    if not hasattr(aself, '_cached_mime') or self._cached_mime is None:
Romain Courteaud's avatar
Romain Courteaud committed
141
      self._cached_mime = PersistentMapping()
142 143 144 145 146 147

  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
148
    self.updateConversionCache()
149 150
    return self._cached_data.has_key(makeSortedTuple(format))

Bartek Górny's avatar
Bartek Górny committed
151
  security.declareProtected(Permissions.View, 'getCacheTime')
152 153
  def getCacheTime(self, **format):
    """
154
      Checks when if ever was the file produced
155
    """
Romain Courteaud's avatar
Romain Courteaud committed
156
    self.updateConversionCache()
157 158
    return self._cached_time.get(makeSortedTuple(format), 0)

Bartek Górny's avatar
Bartek Górny committed
159
  security.declareProtected(Permissions.ModifyPortalContent, 'updateConversion')
160
  def updateConversion(self, **format):
Romain Courteaud's avatar
Romain Courteaud committed
161 162
    self.updateConversionCache()
    self._cached_time[makeSortedTuple(format)] = DateTime()
163

Bartek Górny's avatar
Bartek Górny committed
164
  security.declareProtected(Permissions.ModifyPortalContent, 'setConversion')
165
  def setConversion(self, data, mime=None, **format):
Bartek Górny's avatar
Bartek Górny committed
166 167 168 169
    """
    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
170
    self.updateConversionCache()
171 172 173 174 175
    tformat = makeSortedTuple(format)
    if mime is not None:
      self._cached_mime[tformat] = mime
    if data is not None:
      self._cached_data[tformat] = data
176
      self.updateConversion(**format)
177 178
    self._p_changed = 1

Bartek Górny's avatar
Bartek Górny committed
179
  security.declareProtected(Permissions.View, 'getConversion')
180
  def getConversion(self, **format):
Bartek Górny's avatar
Bartek Górny committed
181 182 183 184 185 186 187 188 189
    """
    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
190
    self.updateConversionCache()
191 192 193
    tformat = makeSortedTuple(format)
    return self._cached_mime.get(tformat, ''), self._cached_data.get(tformat, '')

194
  security.declareProtected(Permissions.ViewManagementScreens, 'getConversionCacheInfo')
195 196 197
  def getConversionCacheInfo(self):
    """
    Get cache details as string (for debugging)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
198
    """
Romain Courteaud's avatar
Romain Courteaud committed
199
    self.updateConversionCache()
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
    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
218

219
class Document(XMLObject, UrlMixIn, ConversionCacheMixin, SnapshotMixin):
Bartek Górny's avatar
Bartek Górny committed
220 221 222 223 224 225
  """
      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.

226 227 228 229 230 231 232 233 234 235
      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
236
      External Documents may be downloaded once or updated at
237 238
      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
239 240 241 242
      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.
243

Bartek Górny's avatar
Bartek Górny committed
244 245 246 247 248 249 250 251 252
      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
253 254 255 256
        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
257 258 259

      Document classes which implement conversion should use
      the ConversionCacheMixin class so that converted values are
260
      stored inside ZODB and do not need to be recalculated.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
261 262 263 264 265 266
      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
267 268 269 270 271 272 273 274 275

      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)
276 277
      (4) other possible actions to finalise the ingestion (ex. by assigning
          a reference)
Bartek Górny's avatar
Bartek Górny committed
278 279 280 281 282

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

      Metadata can be drawn from various sources:

283 284
      input      -   data supplied with http request or set on the object during (2) (e.g.
                     discovered from email text)
285
      file_name  -   data which might be encoded in file name
286 287
      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
288 289 290

      If a certain property is defined in more than one source, it is set according to
      preference order returned by a script 
291 292 293
         Document_getPreferredDocumentMetadataDiscoveryOrderList
         (or any type-based version since discovery is type dependent)

Bartek Górny's avatar
Bartek Górny committed
294
      Methods for discovering metadata are:
295

Bartek Górny's avatar
Bartek Górny committed
296 297 298 299 300
        getPropertyDictFromInput
        getPropertyDictFromFileName
        getPropertyDictFromUserLogin
        getPropertyDictFromContent

301 302 303 304 305
      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
306
        index_html (overriden in Document subclasses)
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321

      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
322

323 324 325 326 327 328
      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
329

330 331
      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
332

333
      * Document_getPropertyDictFromUserLogin - finds a user (by user_login or from session)
Bartek Górny's avatar
Bartek Górny committed
334 335
        and returns properties which should be set on the document

336
      * Document_getPropertyDictFromContent - analyzes document content and returns
Bartek Górny's avatar
Bartek Górny committed
337 338
        properties which should be set on the document

339
      * Base_getImplicitSuccessorValueList - finds appropriate all documents
340
        referenced in the current content
Bartek Górny's avatar
Bartek Górny committed
341

342
      * Base_getImplicitPredecessorValueList - finds document predecessors based on
Bartek Górny's avatar
Bartek Górny committed
343 344 345 346 347 348 349 350 351
        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)

352
      * Document_getNewRevision - calculates revision number which should be set
Bartek Górny's avatar
Bartek Górny committed
353 354 355
        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
356 357 358 359
      * 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
360
      Subcontent: documents may include subcontent (files, images, etc.)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
361 362 363
      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
364 365 366 367 368

    Consistency checking:
      Default implementation uses DocumentReferenceConstraint to check if the 
      reference/language/version triplet is unique. Additional constraints
      can be added if necessary.
369 370 371 372 373 374 375 376 377 378 379 380 381

    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
382 383 384 385 386 387 388 389
  """

  meta_type = 'ERP5 Document'
  portal_type = 'Document'
  add_permission = Permissions.AddPortalContent
  isPortalContent = 1
  isRADContent = 1
  isDocument = 1
390
  __dav_collection__=0
Bartek Górny's avatar
Bartek Górny committed
391

Jean-Paul Smets's avatar
Jean-Paul Smets committed
392 393
  # Regular expressions
  href_parser = re.compile('<a[^>]*href=[\'"](.*?)[\'"]',re.IGNORECASE)
394 395
  body_parser = re.compile('<body[^>]*>(.*?)</body>', re.IGNORECASE + re.DOTALL)
  title_parser = re.compile('<title[^>]*>(.*?)</title>', re.IGNORECASE + re.DOTALL)
396
  base_parser = re.compile('<base[^>]*href=[\'"](.*?)[\'"][^>]*>', re.IGNORECASE + re.DOTALL)
397
  charset_parser = re.compile('charset="?([a-z0-9\-]+)', re.IGNORECASE)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
398

Bartek Górny's avatar
Bartek Górny committed
399 400 401 402 403 404 405 406 407 408 409
  # 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
410 411
                    , PropertySheet.ExternalDocument
                    , PropertySheet.Url
412
                    , PropertySheet.Periodicity
413
                    , PropertySheet.Snapshot
Bartek Górny's avatar
Bartek Górny committed
414 415 416 417 418
                    )

  # Declarative interfaces
  __implements__ = ()

419 420 421
  searchable_property_list = ('asText', 'title', 'description', 'id', 'reference',
                              'version', 'short_title',
                              'subject', 'source_reference', 'source_project_title',)
Bartek Górny's avatar
Bartek Górny committed
422

423
  ### Content processing methods
424
  security.declareProtected(Permissions.View, 'index_html')
425
  def index_html(self, REQUEST, RESPONSE, format=None, **kw):
426 427
    """
      We follow here the standard Zope API for files and images
Jean-Paul Smets's avatar
Jean-Paul Smets committed
428 429 430 431 432 433 434 435 436 437 438
      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.

439 440 441
      Should return appropriate format (calling convert
      if necessary) and set headers.

442
      format -- the format specied in the form of an extension
443
      string (ex. jpeg, html, text, txt, etc.)
444 445

      **kw -- can be various things - e.g. resolution
446

Jean-Paul Smets's avatar
Jean-Paul Smets committed
447 448 449
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
450
    """
451
    raise NotImplementedError
452

Bartek Górny's avatar
Bartek Górny committed
453 454
  security.declareProtected(Permissions.View, 'getSearchableText')
  def getSearchableText(self, md=None):
455
    """
456 457 458
      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.
459

460 461
      XXX-JPS - This method is nice. It should probably be moved to Base class
      searchable_property_list could become a standard class attribute.
462

463
      TODO (future): Make this property a per portal type property.
Bartek Górny's avatar
Bartek Górny committed
464 465
    """
    def getPropertyListOrValue(property):
466
      """
467
        we try to get a list, else we get value and convert to list
468
      """
469 470 471 472 473 474 475
      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
476 477 478 479
      val = self.getPropertyList(property)
      if val is None:
        val = self.getProperty(property)
        if val is not None and val != '':
480
          val = [str(val)]
481 482 483
        else:
          val = []
      else:
484
        val = [str(v) for v in list(val) if v is not None]
Bartek Górny's avatar
Bartek Górny committed
485
      return val
486

487
    searchable_text = reduce(add, map(lambda x: getPropertyListOrValue(x),
Bartek Górny's avatar
Bartek Górny committed
488
                                                self.searchable_property_list))
489
    searchable_text = ' '.join(searchable_text)
Bartek Górny's avatar
Bartek Górny committed
490 491 492
    return searchable_text

  # Compatibility with CMF Catalog
Jean-Paul Smets's avatar
Jean-Paul Smets committed
493
  SearchableText = getSearchableText
Bartek Górny's avatar
Bartek Górny committed
494

495 496 497 498 499 500 501
  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
502
  ### Relation getters
503
  security.declareProtected(Permissions.AccessContentsInformation, 'getSearchableReferenceList')
504
  def getSearchableReferenceList(self):
Bartek Górny's avatar
Bartek Górny committed
505
    """
506 507 508 509 510
      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.
    """
511
    text = self.getSearchableText() # XXX getSearchableText or asText ?
512
    regexp = self.portal_preferences.getPreferredDocumentReferenceRegularExpression()
Bartek Górny's avatar
Bartek Górny committed
513
    try:
514
      rx_search = re.compile(regexp)
Bartek Górny's avatar
Bartek Górny committed
515
    except TypeError: # no regexp in preference
516 517 518
      LOG('ERP5/Document/Document.getSearchableReferenceList', 0,
          'Document regular expression must be set in portal preferences')
      return ()
519
    res = rx_search.finditer(text)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
520
    res = [(r.group(), r.groupdict()) for r in res]
Bartek Górny's avatar
Bartek Górny committed
521
    return res
522
    
523
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSuccessorValueList')
Bartek Górny's avatar
Bartek Górny committed
524 525
  def getImplicitSuccessorValueList(self):
    """
526 527 528
      Find objects which we are referencing (if our text_content contains
      references of other documents). The whole implementation is delegated to
      Base_getImplicitSuccessorValueList script.
529

530
      The implementation goes in 2 steps:
531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547

    - 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
548 549
    """
    # XXX results should be cached as volatile attributes
550 551 552 553 554 555
    refs = [r[1] for r in self.getSearchableReferenceList()]
    res = self.Base_getImplicitSuccessorValueList(refs)
    # get unique latest (most relevant) versions
    res = [r.getObject().getLatestVersionValue() for r in res]
    res_dict = dict.fromkeys(res)
    return res_dict.keys()
Bartek Górny's avatar
Bartek Górny committed
556

557
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
558 559 560
  def getImplicitPredecessorValueList(self):
    """
      This function tries to find document which are referencing us - by reference only, or
561
      by reference/language etc. Implementation is passed to 
562
        Base_getImplicitPredecessorValueList
563 564 565 566 567 568 569 570 571 572 573 574 575

      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
576 577
    """
    # XXX results should be cached as volatile attributes
578 579
    method = self._getTypeBasedMethod('getImplicitPredecessorValueList', 
        fallback_script_id = 'Base_getImplicitPredecessorValueList')
Bartek Górny's avatar
Bartek Górny committed
580
    lst = method()
581 582 583 584
    # 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
    di = dict.fromkeys([o.getLatestVersionValue() for o in di.keys()])
Bartek Górny's avatar
Bartek Górny committed
585 586 587
    ref = self.getReference()
    return [o for o in di.keys() if o.getReference() != ref] # every object has its own reference in SearchableText

588
  security.declareProtected(Permissions.AccessContentsInformation, 'getImplicitSimilarValueList')
Bartek Górny's avatar
Bartek Górny committed
589 590 591 592 593 594 595 596 597
  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 []

598
  security.declareProtected(Permissions.AccessContentsInformation, 'getSimilarCloudValueList')
599
  def getSimilarCloudValueList(self, depth=0):
Bartek Górny's avatar
Bartek Górny committed
600 601 602 603 604 605 606 607
    """
      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)

608 609
    #gettername = 'get%sValueList' % convertToUpperCase(category)
    #relatedgettername = 'get%sRelatedValueList' % convertToUpperCase(category)
Bartek Górny's avatar
Bartek Górny committed
610

611
    def getRelatedList(ob, level=0):
Bartek Górny's avatar
Bartek Górny committed
612
      level += 1
613 614 615 616
      #getter = getattr(self, gettername)
      #relatedgetter = getattr(self, relatedgettername)
      #res = getter() + relatedgetter()
      res = ob.getSimilarValueList() + ob.getSimilarRelatedValueList()
Bartek Górny's avatar
Bartek Górny committed
617 618 619
      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
620 621
          if level != depth:
            getRelatedList(r, level)
Bartek Górny's avatar
Bartek Górny committed
622

623
    getRelatedList(self)
Bartek Górny's avatar
Bartek Górny committed
624 625 626
    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
627 628 629 630
    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
631 632 633

    return lista_latest.keys()

634
  ### Version and language getters - might be moved one day to a mixin class in base
Bartek Górny's avatar
Bartek Górny committed
635 636 637
  security.declareProtected(Permissions.View, 'getLatestVersionValue')
  def getLatestVersionValue(self, language=None):
    """
638 639
      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
640

641 642
      If language is provided, return the latest document
      in the language.
Bartek Górny's avatar
Bartek Górny committed
643

644 645 646
      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
647
    """
Bartek Górny's avatar
Bartek Górny committed
648 649
    if not self.getReference():
      return self
650
    catalog = getToolByName(self, 'portal_catalog', None)
651
    kw = dict(reference=self.getReference(), sort_on=(('version','descending'),))
652 653 654 655 656 657
    if language is not None: kw['language'] = language
    res = catalog(**kw)

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

658
    # if language was given return it (if there are any docs in this language)
659
    if language is not None:
660 661 662 663
      try:
        return res[0].getObject()
      except IndexError:
        return None
664 665 666 667 668 669 670
    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
671
            return in_original.getObject()
672
          else:
Bartek Górny's avatar
Bartek Górny committed
673
            return first.getObject() # this shouldn't happen in real life
674 675
        if ob.getLanguage() == user_language:
          # we found it in the user language
Bartek Górny's avatar
Bartek Górny committed
676
          return ob.getObject()
677 678 679
        if ob.getLanguage() == original_language:
          # this is in original language
          in_original = ob
680 681
    # this is the only doc in this version
    return self
Bartek Górny's avatar
Bartek Górny committed
682 683 684 685 686 687 688

  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
689
    catalog = getToolByName(self, 'portal_catalog', None)
690
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
691
                   reference=self.getReference(),
692
                   order_by=(('version', 'descending', 'SIGNED'),)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
693
                  )
694 695 696
    if version: kw['version'] = version
    if language: kw['language'] = language
    return catalog(**kw)
Bartek Górny's avatar
Bartek Górny committed
697

698
  security.declareProtected(Permissions.AccessContentsInformation, 'isVersionUnique')
Bartek Górny's avatar
Bartek Górny committed
699 700
  def isVersionUnique(self):
    """
701 702 703
      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
704
    """
705 706
    if not self.getReference():
      return 1
Jean-Paul Smets's avatar
Jean-Paul Smets committed
707
    catalog = getToolByName(self, 'portal_catalog', None)
708
    self_count = catalog.unrestrictedCountResults(portal_type=self.getPortalDocumentTypeList(),
709 710 711
                                            reference=self.getReference(),
                                            version=self.getVersion(),
                                            language=self.getLanguage(),
712
                                            uid=self.getUid(),
713
                                            validation_state="!=cancelled"
714 715 716 717 718
                                            )[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
719
                                            validation_state="!=cancelled"
720 721 722 723 724 725 726 727 728 729 730
                                            )[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
731
    catalog = getToolByName(self.getPortalObject(), 'portal_catalog')
732 733 734 735 736 737 738 739 740
    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)
  
741 742 743 744
  security.declareProtected(Permissions.AccessContentsInformation, 'getRevision')
  def getRevision(self):
    """
      Returns the current revision by analysing the change log
745 746 747 748
      of the current object. The return value is a string
      in order to be consistent with the property sheet
      definition.
      
749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765
      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:
766 767
        return str(len(filter(lambda x:x.get('action', None) in ('edit', 'upload'), history)))
    return ''
768 769 770 771 772 773 774

  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.
    """
775 776 777 778
    revision = self.getRevision()
    if revision == '':
      return []
    return [str(r) for r in range(0, int(self.getRevision()))]
779

780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805
  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(),
806
                where_expression="validation_state NOT IN ('cancelled', 'deleted')")
807 808 809 810 811 812
      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
813 814 815
      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))
816 817 818 819 820 821 822 823 824 825
      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():
826
          raise ValueError, "[DMS] Ingestion may not change the type of an existing document"
827 828
        elif not _checkPermission(Permissions.ModifyPortalContent, existing_document):
          self.setUniqueReference(suffix='unauthorized')
829
          raise Unauthorized, "[DMS] You are not allowed to update the existing document which has the same coordinates (id %s)" % existing_document.getId()
830 831 832
        else:
          update_kw = {}
          for k in self.propertyIds():
833
            if k not in FIXED_PROPERTY_IDS and self.hasProperty(k):
834 835 836 837 838 839
              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

840
  security.declareProtected(Permissions.AccessContentsInformation, 'getLanguageList')
Bartek Górny's avatar
Bartek Górny committed
841 842 843 844 845
  def getLanguageList(self, version=None):
    """
      Returns a list of languages which this document is available in
      for the current user.
    """
846
    if not self.getReference(): return []
Jean-Paul Smets's avatar
Jean-Paul Smets committed
847
    catalog = getToolByName(self, 'portal_catalog', None)
848
    kw = dict(portal_type=self.getPortalType(),
Jean-Paul Smets's avatar
Jean-Paul Smets committed
849
                           reference=self.getReference(),
850 851 852 853
                           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
854

855
  security.declareProtected(Permissions.AccessContentsInformation, 'getOriginalLanguage')
Bartek Górny's avatar
Bartek Górny committed
856 857 858
  def getOriginalLanguage(self):
    """
      Returns the original language of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
859 860

      XXX-JPS not implemented yet ?
Bartek Górny's avatar
Bartek Górny committed
861 862 863 864 865
    """
    # 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?
866 867 868 869 870 871 872 873
    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
874 875 876 877 878 879

  ### 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')
880
  def getPropertyDictFromUserLogin(self, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
881 882 883 884 885
    """
      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:
886 887
      user_login = str(getSecurityManager().getUser())
    method = self._getTypeBasedMethod('getPropertyDictFromUserLogin',
Bartek Górny's avatar
Bartek Górny committed
888
        fallback_script_id='Document_getPropertyDictFromUserLogin')
889
    return method(user_login)
Bartek Górny's avatar
Bartek Górny committed
890 891 892 893 894 895 896

  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
    """
897 898
    if not self.hasBaseData():
      self.convertToBaseFormat()
899
    method = self._getTypeBasedMethod('getPropertyDictFromContent',
Bartek Górny's avatar
Bartek Górny committed
900
        fallback_script_id='Document_getPropertyDictFromContent')
901
    return method()
Bartek Górny's avatar
Bartek Górny committed
902 903 904 905 906 907 908 909 910 911 912 913 914 915

  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).
916 917 918 919 920

      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
921 922 923 924 925 926
    """
    if hasattr(self, '_backup_input'):
      return getattr(self, '_backup_input')
    kw = {}
    for id in self.propertyIds():
      # We should not consider file data
927 928
      if id not in ('data', 'categories_list', 'uid', 'id',
                    'text_content', 'base_data',) \
929
            and self.hasProperty(id):
Bartek Górny's avatar
Bartek Górny committed
930 931 932 933 934 935
        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

936 937 938 939 940 941
  security.declareProtected(Permissions.AccessContentsInformation, 'getStandardFileName')
  def getStandardFileName(self):
    """
    Returns the document coordinates as a standard file name. This
    method is the reverse of getPropertyDictFromFileName.
    """
942 943 944
    method = self._getTypeBasedMethod('getStandardFileName', 
        fallback_script_id = 'Document_getStandardFileName')
    return method()
945

Bartek Górny's avatar
Bartek Górny committed
946 947
  ### Metadata disovery and ingestion methods
  security.declareProtected(Permissions.ModifyPortalContent, 'discoverMetadata')
948
  def discoverMetadata(self, file_name=None, user_login=None):
Bartek Górny's avatar
Bartek Górny committed
949
    """
950 951
      This is the main metadata discovery function - controls the process
      of discovering data from various sources. The discovery itself is
952 953 954
      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
955

956
      file_name - this parameter is a file name of the form "AA-BBB-CCC-223-en"
Bartek Górny's avatar
Bartek Górny committed
957

958
      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
959
                   currently logged in, then we'll get him from session
960
    """
Bartek Górny's avatar
Bartek Górny committed
961 962 963
    # Preference is made of a sequence of 'user_login', 'content', 'file_name', 'input'
    method = self._getTypeBasedMethod('getPreferredDocumentMetadataDiscoveryOrderList', 
        fallback_script_id = 'Document_getPreferredDocumentMetadataDiscoveryOrderList')
964
    order_list = list(method())
965
    order_list.reverse()
966
    # build a dictionary according to the order
Bartek Górny's avatar
Bartek Górny committed
967
    kw = {}
968
    for order_id in order_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
969
      result = None
Bartek Górny's avatar
Bartek Górny committed
970 971
      if order_id not in VALID_ORDER_KEY_LIST:
        # Prevent security attack or bad preferences
972
        raise AttributeError, "%s is not in valid order key list" % order_id
Bartek Górny's avatar
Bartek Górny committed
973 974 975
      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
976 977
        if file_name is not None:
          result = method(file_name)
Bartek Górny's avatar
Bartek Górny committed
978
      elif order_id == 'user_login':
Jean-Paul Smets's avatar
Jean-Paul Smets committed
979 980
        if user_login is not None:
          result = method(user_login)
Bartek Górny's avatar
Bartek Górny committed
981 982
      else:
        result = method()
983 984
      if result is not None:
        kw.update(result)
Jean-Paul Smets's avatar
Jean-Paul Smets committed
985

986 987 988
    if file_name is not None:
      # filename is often undefined....
      kw['source_reference'] = file_name
989
    # Prepare the content edit parameters - portal_type should not be changed
990 991 992 993 994
    kw.pop('portal_type', None)
    # Try not to invoke an automatic transition here
    self._edit(**kw)
    # Finish ingestion by calling method
    self.finishIngestion() 
995
    self.reindexObject()
996 997
    # Revision merge is tightly coupled
    # to metadata discovery - refer to the documentation of mergeRevision method
998 999 1000
    merged_doc = self.mergeRevision()
    merged_doc.reindexObject()
    return merged_doc 
Bartek Górny's avatar
Bartek Górny committed
1001 1002 1003 1004

  security.declareProtected(Permissions.ModifyPortalContent, 'finishIngestion')
  def finishIngestion(self):
    """
1005 1006 1007
      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
1008
    """
1009 1010
    method = self._getTypeBasedMethod('finishIngestion', fallback_script_id='Document_finishIngestion')
    return method()
Bartek Górny's avatar
Bartek Górny committed
1011

1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
  # 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

1022 1023 1024
      Default implementation returns an empty string (html, text)
      or raises an error.

1025 1026 1027 1028
      TODO:
      - implement guards API so that conversion to certain
        formats require certain permission
    """
1029 1030 1031 1032 1033
    if format == 'html':
      return 'text/html', ''
    if format in ('text', 'txt'):
      return 'text/plain', ''
    raise NotImplementedError
1034 1035 1036 1037 1038 1039

  security.declareProtected(Permissions.View, 'asText')
  def asText(self):
    """
      Converts the content of the document to a textual representation.
    """
1040 1041
    mime, data = self.convert(format='txt')
    return data
1042

1043 1044
  security.declareProtected(Permissions.View, 'asEntireHTML')
  def asEntireHTML(self):
1045 1046
    """
      Returns a complete HTML representation of the document
1047 1048 1049
      (with body tags, etc.). Adds if necessary a base
      tag so that the document can be displayed in an iframe
      or standalone.
1050 1051

      Actual conversion is delegated to _asHTML
1052
    """
1053
    html = self._asHTML()
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064
    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
1065

1066 1067 1068 1069 1070 1071
  security.declarePrivate('_asHTML')
  def _asHTML(self):
    """
      A private method which converts to HTML. This method
      is the one to override in subclasses.
    """
1072
    if not self.hasBaseData():
1073
      return ''
1074 1075 1076 1077 1078 1079
    if self.hasConversion(format='base-html'):
      mime, data = self.getConversion(format='base-html')
      return data
    mime, html = self.convert(format='html')
    return html

1080 1081 1082
  security.declareProtected(Permissions.View, 'asStrippedHTML')
  def asStrippedHTML(self):
    """
1083 1084
      Returns a stripped HTML representation of the document
      (without html and body tags, etc.) which can be used to inline
1085 1086
      a preview of the document.
    """
1087
    if not self.hasBaseData():
1088
      return ''
1089
    if self.hasConversion(format='stripped-html'): # XXX this is redundant since we never set it
1090 1091 1092 1093 1094
      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):
1095 1096 1097 1098
      stripped_html = body_list[0]
    else:
      stripped_html = html
    # find charset and convert to utf-8
1099 1100
    charset_list = self.charset_parser.findall(str(html)) # XXX - Not efficient is datastream 
                                                          # instance but hard to do better
1101
    if charset_list:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1102
      stripped_html = unicode(str(stripped_html), charset_list[0]).encode('utf-8')
1103
    return stripped_html
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113

  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
1114
    html = self.asEntireHTML()
1115 1116 1117 1118 1119
    if not html: return result
    title_list = re.findall(self.title_parser, str(html))
    if title_list:
      result['title'] = title_list[0]
    return result
1120 1121

  # Base format support
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1122
  security.declareProtected(Permissions.ModifyPortalContent, 'convertToBaseFormat')
1123
  def convertToBaseFormat(self):
1124
    """
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
      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.)
1140 1141
    """
    try:
1142 1143
      msg = self._convertToBaseFormat() # Call implemetation method
      self.clearConversionCache() # Conversion cache is now invalid
1144
      if msg is None:
1145
        msg = 'Converted to %s.' % self.getBaseContentType()
1146 1147 1148 1149 1150
      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.')
1151 1152
      #self.processFile(comment=msg)
      raise ConversionError, msg
1153 1154
    except Fault, e:
      msg = 'Problem: %s' % (repr(e) or 'undefined.')
1155 1156
      #self.processFile(comment=msg)
      raise Fault, msg
1157 1158
    except socket.error, e:
      msg = 'Problem: %s' % (repr(e) or 'undefined.')
1159 1160
      #self.processFile(comment=msg)
      raise socket.error, msg
1161 1162
    return msg

1163
  def _convertToBaseFormat(self):
1164
    """
1165 1166 1167 1168 1169
      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
1170

1171
      Other possible applications: conversion of HTML
1172
      text to tiddy HTML such as described here:
1173 1174
      http://www.xml.com/pub/a/2004/09/08/pyxml.html
      so that resulting text can be processed more
1175 1176 1177
      easily by XSLT parsers. Conversion of internal
      links to images of an HTML document to local
      links (in combindation with populate).
1178
    """
1179
    raise NotImplementedError
1180

1181 1182 1183 1184 1185 1186
  def convertFile(self, msg=None):
    """
    Workflow transition invoked when conversion occurs.
    """
  convertFile = WorkflowMethod(convertFile)

1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
  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

1199
  # Transformation API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1200
  security.declareProtected(Permissions.ModifyPortalContent, 'populateContent')
1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
  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
1213 1214 1215 1216 1217
    try:
      method = self._getTypeBasedMethod('populateContent')
    except KeyError, AttributeError:
      method = None
    if method is not None: method()
1218 1219

  # Crawling API
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1220
  security.declareProtected(Permissions.AccessContentsInformation, 'getContentURLList')
1221 1222 1223
  def getContentURLList(self):
    """
      Returns a list of URLs referenced by the content of this document.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1224 1225 1226 1227 1228 1229
      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.
1230
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1231 1232
    html_content = self.asStrippedHTML()
    return re.findall(self.href_parser, str(html_content))
1233

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1234
  security.declareProtected(Permissions.ModifyPortalContent, 'updateContentFromURL')
1235
  def updateContentFromURL(self, repeat=MAX_REPEAT, crawling_depth=0):
1236 1237
    """
      Download and update content of this document from its source URL.
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1238 1239
      Implementation is handled by ContributionTool.
    """
1240
    self.portal_contributions.updateContentFromURL(self, repeat=repeat, crawling_depth=crawling_depth)
1241

Jean-Paul Smets's avatar
Jean-Paul Smets committed
1242 1243
  security.declareProtected(Permissions.ModifyPortalContent, 'crawlContent')
  def crawlContent(self):
1244
    """
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1245 1246 1247 1248
      Initialises the crawling process on the current document.
    """
    self.portal_contributions.crawlContent(self)

1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264
  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
1265 1266 1267 1268 1269 1270 1271 1272
  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
1273
    if len(base_url_list):
1274
      if base_url_list[-1] and base_url_list[-1].find('.') > 0:
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1275
        # Cut the trailing part in http://www.some.site/at/trailing.html
1276
        # but not in http://www.some.site/at
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1277
        base_url = '/'.join(base_url_list[:-1])
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1278 1279
    return base_url

1280 1281
  security.declareProtected(Permissions.AccessContentsInformation, 'getFrequencyIndex')
  def getFrequencyIndex(self):
Jean-Paul Smets's avatar
Jean-Paul Smets committed
1282
    """
1283 1284 1285 1286
      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
1287
    """
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321
    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
1322
    return method()