blob: f1a629ff7bc17abf52b1b4d1312bc407d554cd19 [file] [log] [blame]
#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
#
"""
Module containing the Atom Pub binding-specific objects used to work with a CMIS
provider.
"""
from cmislib.cmis_services import Binding, RepositoryServiceIfc
from cmislib.domain import CmisId, CmisObject, ObjectType, Property, ACL, ACE, ChangeEntry, ResultSet, Rendition
from cmislib import messages
from cmislib.net import RESTService as Rest
from cmislib.exceptions import CmisException, \
ObjectNotFoundException, InvalidArgumentException, \
NotSupportedException
from cmislib.util import multiple_replace, parsePropValue, parseBoolValue, toCMISValue, parseDateTimeValue
from urllib import quote
from urlparse import urlparse, urlunparse
import re
import mimetypes
from xml.parsers.expat import ExpatError
import datetime
import StringIO
import logging
from xml.dom import minidom
moduleLogger = logging.getLogger('cmislib.atompub_binding')
# Namespaces
ATOM_NS = 'http://www.w3.org/2005/Atom'
APP_NS = 'http://www.w3.org/2007/app'
CMISRA_NS = 'http://docs.oasis-open.org/ns/cmis/restatom/200908/'
CMIS_NS = 'http://docs.oasis-open.org/ns/cmis/core/200908/'
# Content types
# Not all of these patterns have variability, but some do. It seemed cleaner
# just to treat them all like patterns to simplify the matching logic
ATOM_XML_TYPE = 'application/atom+xml'
ATOM_XML_ENTRY_TYPE = 'application/atom+xml;type=entry'
ATOM_XML_ENTRY_TYPE_P = re.compile('^application/atom\+xml.*type.*entry')
ATOM_XML_FEED_TYPE = 'application/atom+xml;type=feed'
ATOM_XML_FEED_TYPE_P = re.compile('^application/atom\+xml.*type.*feed')
CMIS_TREE_TYPE = 'application/cmistree+xml'
CMIS_TREE_TYPE_P = re.compile('^application/cmistree\+xml')
CMIS_QUERY_TYPE = 'application/cmisquery+xml'
CMIS_ACL_TYPE = 'application/cmisacl+xml'
# Standard rels
DOWN_REL = 'down'
FIRST_REL = 'first'
LAST_REL = 'last'
NEXT_REL = 'next'
PREV_REL = 'prev'
SELF_REL = 'self'
UP_REL = 'up'
TYPE_DESCENDANTS_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/typedescendants'
VERSION_HISTORY_REL = 'version-history'
FOLDER_TREE_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/foldertree'
RELATIONSHIPS_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/relationships'
ACL_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/acl'
CHANGE_LOG_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/changes'
POLICIES_REL = 'http://docs.oasis-open.org/ns/cmis/link/200908/policies'
RENDITION_REL = 'alternate'
# Collection types
QUERY_COLL = 'query'
TYPES_COLL = 'types'
CHECKED_OUT_COLL = 'checkedout'
UNFILED_COLL = 'unfiled'
ROOT_COLL = 'root'
class AtomPubBinding(Binding):
"""
The binding responsible for talking to the CMIS server via the AtomPub
Publishing Protocol.
"""
def __init__(self, **kwargs):
self.extArgs = kwargs
def getRepositoryService(self):
return RepositoryService()
def get(self, url, username, password, **kwargs):
"""
Does a get against the CMIS service. More than likely, you will not
need to call this method. Instead, let the other objects do it for you.
For example, if you need to get a specific object by object id, try
:class:`Repository.getObject`. If you have a path instead of an object
id, use :class:`Repository.getObjectByPath`. Or, you could start with
the root folder (:class:`Repository.getRootFolder`) and drill down from
there.
"""
# merge the cmis client extended args with the ones that got passed in
if len(self.extArgs) > 0:
kwargs.update(self.extArgs)
resp, content = Rest().get(url,
username=username,
password=password,
**kwargs)
if resp['status'] != '200':
self._processCommonErrors(resp, url)
return content
else:
try:
return minidom.parseString(content)
except ExpatError:
raise CmisException('Could not parse server response', url)
def delete(self, url, username, password, **kwargs):
"""
Does a delete against the CMIS service. More than likely, you will not
need to call this method. Instead, let the other objects do it for you.
For example, to delete a folder you'd call :class:`Folder.delete` and
to delete a document you'd call :class:`Document.delete`.
"""
# merge the cmis client extended args with the ones that got passed in
if len(self.extArgs) > 0:
kwargs.update(self.extArgs)
resp, content = Rest().delete(url,
username=username,
password=password,
**kwargs)
if resp['status'] != '200' and resp['status'] != '204':
self._processCommonErrors(resp, url)
return content
else:
pass
def post(self, url, username, password, payload, contentType, **kwargs):
"""
Does a post against the CMIS service. More than likely, you will not
need to call this method. Instead, let the other objects do it for you.
For example, to update the properties on an object, you'd call
:class:`CmisObject.updateProperties`. Or, to check in a document that's
been checked out, you'd call :class:`Document.checkin` on the PWC.
"""
# merge the cmis client extended args with the ones that got passed in
if len(self.extArgs) > 0:
kwargs.update(self.extArgs)
resp, content = Rest().post(url,
payload,
contentType,
username=username,
password=password,
**kwargs)
if resp['status'] == '200':
try:
return minidom.parseString(content)
except ExpatError:
raise CmisException('Could not parse server response', url)
elif resp['status'] == '201':
try:
return minidom.parseString(content)
except ExpatError:
raise CmisException('Could not parse server response', url)
else:
self._processCommonErrors(resp, url)
return resp
def put(self, url, username, password, payload, contentType, **kwargs):
"""
Does a put against the CMIS service. More than likely, you will not
need to call this method. Instead, let the other objects do it for you.
For example, to update the properties on an object, you'd call
:class:`CmisObject.updateProperties`. Or, to check in a document that's
been checked out, you'd call :class:`Document.checkin` on the PWC.
"""
# merge the cmis client extended args with the ones that got passed in
if len(self.extArgs) > 0:
kwargs.update(self.extArgs)
resp, content = Rest().put(url,
payload,
contentType,
username=username,
password=password,
**kwargs)
if resp['status'] != '200' and resp['status'] != '201':
self._processCommonErrors(resp, url)
return content
else:
try:
return minidom.parseString(content)
except ExpatError:
# This may happen and is normal
return None
class RepositoryService(RepositoryServiceIfc):
"""
The repository service for the AtomPub binding.
"""
def __init__(self):
self._uriTemplates = {}
self.logger = logging.getLogger('cmislib.atompub_binding.RepositoryService')
def reload(self, obj):
""" Reloads the state of the repository object."""
self.logger.debug('Reload called on object')
obj.xmlDoc = obj._cmisClient.binding.get(obj._cmisClient.repositoryUrl.encode('utf-8'),
obj._cmisClient.username,
obj._cmisClient.password)
obj._initData()
def getRepository(self, client, repositoryId):
"""
Get the repository for the specified repositoryId.
"""
doc = client.binding.get(client.repositoryUrl, client.username, client.password, **client.extArgs)
workspaceElements = doc.getElementsByTagNameNS(APP_NS, 'workspace')
for workspaceElement in workspaceElements:
idElement = workspaceElement.getElementsByTagNameNS(CMIS_NS, 'repositoryId')
if idElement[0].childNodes[0].data == repositoryId:
return AtomPubRepository(self, workspaceElement)
raise ObjectNotFoundException(url=client.repositoryUrl)
def getRepositories(self, client):
"""
Get all of the repositories provided by the server.
"""
result = client.binding.get(client.repositoryUrl, client.username, client.password, **client.extArgs)
workspaceElements = result.getElementsByTagNameNS(APP_NS, 'workspace')
# instantiate a Repository object using every workspace element
# in the service URL then ask the repository object for its ID
# and name, and return that back
repositories = []
for node in [e for e in workspaceElements if e.nodeType == e.ELEMENT_NODE]:
repository = AtomPubRepository(client, node)
repositories.append({'repositoryId': repository.getRepositoryId(),
'repositoryName': repository.getRepositoryInfo()['repositoryName']})
return repositories
def getDefaultRepository(self, client):
"""
Returns the default repository for the server via the AtomPub binding.
"""
doc = client.binding.get(client.repositoryUrl, client.username, client.password, **client.extArgs)
workspaceElements = doc.getElementsByTagNameNS(APP_NS, 'workspace')
# instantiate a Repository object with the first workspace
# element we find
repository = AtomPubRepository(client, [e for e in workspaceElements if e.nodeType == e.ELEMENT_NODE][0])
return repository
class UriTemplate(dict):
"""
Simple dictionary to represent the data stored in
a URI template entry.
"""
def __init__(self, template, templateType, mediaType):
"""
Constructor
"""
dict.__init__(self)
self['template'] = template
self['type'] = templateType
self['mediaType'] = mediaType
class AtomPubCmisObject(CmisObject):
def __init__(self, cmisClient, repository, objectId=None, xmlDoc=None, **kwargs):
""" Constructor """
self._cmisClient = cmisClient
self._repository = repository
self._objectId = objectId
self._name = None
self._properties = {}
self._allowableActions = {}
self.xmlDoc = xmlDoc
self._kwargs = kwargs
self.logger = logging.getLogger('cmislib.atompub_binding.AtomPubCmisObject')
self.logger.info('Creating an instance of CmisObject')
def __str__(self):
"""To string"""
return self.getObjectId()
def reload(self, **kwargs):
"""
Fetches the latest representation of this object from the CMIS service.
Some methods, like :class:`^Document.checkout` do this for you.
If you call reload with a properties filter, the filter will be in
effect on subsequent calls until the filter argument is changed. To
reset to the full list of properties, call reload with filter set to
'*'.
"""
self.logger.debug('Reload called on CmisObject')
if kwargs:
if self._kwargs:
self._kwargs.update(kwargs)
else:
self._kwargs = kwargs
templates = self._repository.getUriTemplates()
template = templates['objectbyid']['template']
# Doing some refactoring here. Originally, we snagged the template
# and then "filled in" the template based on the args passed in.
# However, some servers don't provide a full template which meant
# supported optional args wouldn't get passed in using the fill-the-
# template approach. What's going on now is that the template gets
# filled in where it can, but if additional, non-templated args are
# passed in, those will get tacked on to the query string as
# "additional" options.
params = {'{id}': self.getObjectId(),
'{filter}': '',
'{includeAllowableActions}': 'false',
'{includePolicyIds}': 'false',
'{includeRelationships}': '',
'{includeACL}': 'false',
'{renditionFilter}': ''}
options = {}
addOptions = {} # args specified, but not in the template
for k, v in self._kwargs.items():
pKey = "{" + k + "}"
if template.find(pKey) >= 0:
options[pKey] = toCMISValue(v)
else:
addOptions[k] = toCMISValue(v)
# merge the templated args with the default params
params.update(options)
# fill in the template
byObjectIdUrl = multiple_replace(params, template)
self.xmlDoc = self._cmisClient.binding.get(byObjectIdUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**addOptions)
self._initData()
# if a returnVersion arg was passed in, it is possible we got back
# a different object ID than the value we started with, so it needs
# to be cleared out as well
if options.has_key('returnVersion') or addOptions.has_key('returnVersion'):
self._objectId = None
def _initData(self):
"""
An internal method used to clear out any member variables that
might be out of sync if we were to fetch new XML from the
service.
"""
self._properties = {}
self._name = None
self._allowableActions = {}
def getObjectId(self):
"""
Returns the object ID for this object.
>>> doc = resultSet.getResults()[0]
>>> doc.getObjectId()
u'workspace://SpacesStore/dc26102b-e312-471b-b2af-91bfb0225339'
"""
if self._objectId is None:
if self.xmlDoc is None:
self.logger.debug('Both objectId and xmlDoc were None, reloading')
self.reload()
props = self.getProperties()
self._objectId = CmisId(props['cmis:objectId'])
return self._objectId
def getObjectParents(self, **kwargs):
"""
Gets the parents of this object as a :class:`ResultSet`.
The following optional arguments are supported:
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
- includeRelativePathSegment
"""
# get the appropriate 'up' link
parentUrl = self._getLink(UP_REL)
if parentUrl is None:
raise NotSupportedException('Root folder does not support getObjectParents')
# invoke the URL
result = self._cmisClient.binding.get(parentUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def getPaths(self):
"""
Returns the object's paths as a list of strings.
"""
# see sub-classes for implementation
pass
def getRenditions(self):
"""
Returns an array of :class:`Rendition` objects. The repository
must support the Renditions capability.
The following optional arguments are not currently supported:
- renditionFilter
- maxItems
- skipCount
"""
# if Renditions capability is None, return notsupported
if self._repository.getCapabilities()['Renditions']:
pass
else:
raise NotSupportedException
if self.xmlDoc is None:
self.reload()
linkElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'link')
renditions = []
for linkElement in linkElements:
if linkElement.attributes.has_key('rel'):
relAttr = linkElement.attributes['rel'].value
if relAttr == RENDITION_REL:
renditions.append(AtomPubRendition(linkElement))
return renditions
def getAllowableActions(self):
"""
Returns a dictionary of allowable actions, keyed off of the action name.
>>> actions = doc.getAllowableActions()
>>> for a in actions:
... print "%s:%s" % (a,actions[a])
...
canDeleteContentStream:True
canSetContentStream:True
canCreateRelationship:True
canCheckIn:False
canApplyACL:False
canDeleteObject:True
canGetAllVersions:True
canGetObjectParents:True
canGetProperties:True
"""
if self._allowableActions == {}:
self.reload(includeAllowableActions=True)
allowElements = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'allowableActions')
assert len(allowElements) == 1, "Expected response to have exactly one allowableActions element"
allowElement = allowElements[0]
for node in [e for e in allowElement.childNodes if e.nodeType == e.ELEMENT_NODE]:
actionName = node.localName
actionValue = parseBoolValue(node.childNodes[0].data)
self._allowableActions[actionName] = actionValue
return self._allowableActions
def getTitle(self):
"""
Returns the value of the object's atom:title property.
"""
if self.xmlDoc is None:
self.reload()
titleElement = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'title')[0]
if titleElement and titleElement.childNodes:
return titleElement.childNodes[0].data
def getProperties(self):
"""
Returns a dict of the object's properties. If CMIS returns an
empty element for a property, the property will be in the
dict with a value of None.
>>> props = doc.getProperties()
>>> for p in props:
... print "%s: %s" % (p, props[p])
...
cmis:contentStreamMimeType: text/html
cmis:creationDate: 2009-12-15T09:45:35.369-06:00
cmis:baseTypeId: cmis:document
cmis:isLatestMajorVersion: false
cmis:isImmutable: false
cmis:isMajorVersion: false
cmis:objectId: workspace://SpacesStore/dc26102b-e312-471b-b2af-91bfb0225339
The optional filter argument is not yet implemented.
"""
# TODO implement filter
if self._properties == {}:
if self.xmlDoc is None:
self.reload()
propertiesElement = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'properties')[0]
# cpattern = re.compile(r'^property([\w]*)')
for node in [e for e in propertiesElement.childNodes if e.nodeType == e.ELEMENT_NODE and e.namespaceURI == CMIS_NS]:
# propertyId, propertyString, propertyDateTime
# propertyType = cpattern.search(node.localName).groups()[0]
propertyName = node.attributes['propertyDefinitionId'].value
if node.childNodes and \
node.getElementsByTagNameNS(CMIS_NS, 'value')[0] and \
node.getElementsByTagNameNS(CMIS_NS, 'value')[0].childNodes:
valNodeList = node.getElementsByTagNameNS(CMIS_NS, 'value')
if len(valNodeList) == 1:
propertyValue = parsePropValue(valNodeList[0].
childNodes[0].data,
node.localName)
else:
propertyValue = []
for valNode in valNodeList:
try:
propertyValue.append(parsePropValue(valNode.
childNodes[0].data,
node.localName))
except IndexError:
pass
else:
propertyValue = None
self._properties[propertyName] = propertyValue
for node in [e for e in self.xmlDoc.childNodes if e.nodeType == e.ELEMENT_NODE and e.namespaceURI == CMISRA_NS]:
propertyName = node.nodeName
if node.childNodes:
propertyValue = node.firstChild.nodeValue
else:
propertyValue = None
self._properties[propertyName] = propertyValue
return self._properties
def getName(self):
"""
Returns the value of cmis:name from the getProperties() dictionary.
We don't need a getter for every standard CMIS property, but name
is a pretty common one so it seems to make sense.
>>> doc.getName()
u'system-overview.html'
"""
if self._name is None:
self._name = self.getProperties()['cmis:name']
return self._name
def updateProperties(self, properties):
"""
Updates the properties of an object with the properties provided.
Only provide the set of properties that need to be updated.
>>> folder = repo.getObjectByPath('/someFolder2')
>>> folder.getName()
u'someFolder2'
>>> props = {'cmis:name': 'someFolderFoo'}
>>> folder.updateProperties(props)
<cmislib.model.Folder object at 0x103ab1210>
>>> folder.getName()
u'someFolderFoo'
"""
self.logger.debug('Inside updateProperties')
# get the self link
selfUrl = self._getSelfLink()
# if we have a change token, we must pass it back, per the spec
args = {}
if self.properties.has_key('cmis:changeToken') and self.properties['cmis:changeToken'] is not None:
self.logger.debug('Change token present, adding it to args')
args = {"changeToken": self.properties['cmis:changeToken']}
# the getEntryXmlDoc function may need the object type
objectTypeId = None
if self.properties.has_key('cmis:objectTypeId') and not properties.has_key('cmis:objectTypeId'):
objectTypeId = self.properties['cmis:objectTypeId']
self.logger.debug('This object type is:%s', objectTypeId)
# build the entry based on the properties provided
xmlEntryDoc = getEntryXmlDoc(self._repository, objectTypeId, properties)
self.logger.debug('xmlEntryDoc:' + xmlEntryDoc.toxml())
# do a PUT of the entry
updatedXmlDoc = self._cmisClient.binding.put(selfUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
xmlEntryDoc.toxml(encoding='utf-8'),
ATOM_XML_TYPE,
**args)
# reset the xmlDoc for this object with what we got back from
# the PUT, then call initData we dont' want to call
# self.reload because we've already got the parsed XML--
# there's no need to fetch it again
self.xmlDoc = updatedXmlDoc
self._initData()
return self
def move(self, sourceFolder, targetFolder):
"""
Moves an object from the source folder to the target folder.
>>> sub1 = repo.getObjectByPath('/cmislib/sub1')
>>> sub2 = repo.getObjectByPath('/cmislib/sub2')
>>> doc = repo.getObjectByPath('/cmislib/sub1/testdoc1')
>>> doc.move(sub1, sub2)
"""
postUrl = targetFolder.getChildrenLink()
args = {"sourceFolderId": sourceFolder.id}
# post the Atom entry
self._cmisClient.binding.post(postUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
self.xmlDoc.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE,
**args)
def delete(self, **kwargs):
"""
Deletes this :class:`CmisObject` from the repository. Note that in the
case of a :class:`Folder` object, some repositories will refuse to
delete it if it contains children and some will delete it without
complaint. If what you really want to do is delete the folder and all
of its descendants, use :meth:`~Folder.deleteTree` instead.
>>> folder.delete()
The optional allVersions argument is supported.
"""
url = self._getSelfLink()
self._cmisClient.binding.delete(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
def applyPolicy(self, policyId):
"""
This is not yet implemented.
"""
# depends on this object's canApplyPolicy allowable action
if self.getAllowableActions()['canApplyPolicy']:
raise NotImplementedError
else:
raise CmisException('This object has canApplyPolicy set to false')
def createRelationship(self, targetObj, relTypeId):
"""
Creates a relationship between this object and a specified target
object using the relationship type specified. Returns the new
:class:`Relationship` object.
>>> rel = tstDoc1.createRelationship(tstDoc2, 'R:cmiscustom:assoc')
>>> rel.getProperties()
{u'cmis:objectId': u'workspace://SpacesStore/271c48dd-6548-4771-a8f5-0de69b7cdc25', u'cmis:creationDate': None, u'cmis:objectTypeId': u'R:cmiscustom:assoc', u'cmis:lastModificationDate': None, u'cmis:targetId': u'workspace://SpacesStore/0ca1aa08-cb49-42e2-8881-53aa8496a1c1', u'cmis:lastModifiedBy': None, u'cmis:baseTypeId': u'cmis:relationship', u'cmis:sourceId': u'workspace://SpacesStore/271c48dd-6548-4771-a8f5-0de69b7cdc25', u'cmis:changeToken': None, u'cmis:createdBy': None}
"""
if isinstance(relTypeId, str):
relTypeId = CmisId(relTypeId)
props = {}
props['cmis:sourceId'] = self.getObjectId()
props['cmis:targetId'] = targetObj.getObjectId()
props['cmis:objectTypeId'] = relTypeId
xmlDoc = getEntryXmlDoc(self._repository, properties=props)
url = self._getLink(RELATIONSHIPS_REL)
assert url is not None, 'Could not determine relationships URL'
result = self._cmisClient.binding.post(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
xmlDoc.toxml(encoding='utf-8'),
ATOM_XML_TYPE)
# instantiate CmisObject objects with the results and return the list
entryElements = result.getElementsByTagNameNS(ATOM_NS, 'entry')
assert(len(entryElements) == 1), "Expected entry element in result from relationship URL post"
return getSpecializedObject(AtomPubCmisObject(self._cmisClient, self, xmlDoc=entryElements[0]))
def getRelationships(self, **kwargs):
"""
Returns a :class:`ResultSet` of :class:`Relationship` objects for each
relationship where the source is this object.
>>> rels = tstDoc1.getRelationships()
>>> len(rels.getResults())
1
>>> rel = rels.getResults().values()[0]
>>> rel.getProperties()
{u'cmis:objectId': u'workspace://SpacesStore/271c48dd-6548-4771-a8f5-0de69b7cdc25', u'cmis:creationDate': None, u'cmis:objectTypeId': u'R:cmiscustom:assoc', u'cmis:lastModificationDate': None, u'cmis:targetId': u'workspace://SpacesStore/0ca1aa08-cb49-42e2-8881-53aa8496a1c1', u'cmis:lastModifiedBy': None, u'cmis:baseTypeId': u'cmis:relationship', u'cmis:sourceId': u'workspace://SpacesStore/271c48dd-6548-4771-a8f5-0de69b7cdc25', u'cmis:changeToken': None, u'cmis:createdBy': None}
The following optional arguments are supported:
- includeSubRelationshipTypes
- relationshipDirection
- typeId
- maxItems
- skipCount
- filter
- includeAllowableActions
"""
url = self._getLink(RELATIONSHIPS_REL)
assert url is not None, 'Could not determine relationships URL'
result = self._cmisClient.binding.get(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def removePolicy(self, policyId):
"""
This is not yet implemented.
"""
# depends on this object's canRemovePolicy allowable action
if self.getAllowableActions()['canRemovePolicy']:
raise NotImplementedError
else:
raise CmisException('This object has canRemovePolicy set to false')
def getAppliedPolicies(self):
"""
This is not yet implemented.
"""
# depends on this object's canGetAppliedPolicies allowable action
if self.getAllowableActions()['canGetAppliedPolicies']:
raise NotImplementedError
else:
raise CmisException('This object has canGetAppliedPolicies set to false')
def getACL(self):
"""
Repository.getCapabilities['ACL'] must return manage or discover.
>>> acl = folder.getACL()
>>> acl.getEntries()
{u'GROUP_EVERYONE': <cmislib.model.ACE object at 0x10071a8d0>, 'jdoe': <cmislib.model.ACE object at 0x10071a590>}
The optional onlyBasicPermissions argument is currently not supported.
"""
if self._repository.getCapabilities()['ACL']:
# if the ACL capability is discover or manage, this must be
# supported
aclUrl = self._getLink(ACL_REL)
result = self._cmisClient.binding.get(aclUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password)
return AtomPubACL(xmlDoc=result)
else:
raise NotSupportedException
def applyACL(self, acl):
"""
Updates the object with the provided :class:`ACL`.
Repository.getCapabilities['ACL'] must return manage to invoke this
call.
>>> acl = folder.getACL()
>>> acl.addEntry(ACE('jdoe', 'cmis:write', 'true'))
>>> acl.getEntries()
{u'GROUP_EVERYONE': <cmislib.model.ACE object at 0x10071a8d0>, 'jdoe': <cmislib.model.ACE object at 0x10071a590>}
"""
if self._repository.getCapabilities()['ACL'] == 'manage':
# if the ACL capability is manage, this must be
# supported
# but it also depends on the canApplyACL allowable action
# for this object
if not isinstance(acl, ACL):
raise CmisException('The ACL to apply must be an instance of the ACL class.')
aclUrl = self._getLink(ACL_REL)
assert aclUrl, "Could not determine the object's ACL URL."
result = self._cmisClient.binding.put(aclUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
acl.getXmlDoc().toxml(encoding='utf-8'),
CMIS_ACL_TYPE)
return AtomPubACL(xmlDoc=result)
else:
raise NotSupportedException
def _getSelfLink(self):
"""
Returns the URL used to retrieve this object.
"""
url = self._getLink(SELF_REL)
assert len(url) > 0, "Could not determine the self link."
return url
def _getLink(self, rel, ltype=None):
"""
Returns the HREF attribute of an Atom link element for the
specified rel.
"""
if self.xmlDoc is None:
self.reload()
linkElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'link')
for linkElement in linkElements:
if ltype:
if linkElement.attributes.has_key('rel'):
relAttr = linkElement.attributes['rel'].value
if ltype and linkElement.attributes.has_key('type'):
typeAttr = linkElement.attributes['type'].value
if relAttr == rel and ltype.match(typeAttr):
return linkElement.attributes['href'].value
else:
if linkElement.attributes.has_key('rel'):
relAttr = linkElement.attributes['rel'].value
if relAttr == rel:
return linkElement.attributes['href'].value
allowableActions = property(getAllowableActions)
name = property(getName)
id = property(getObjectId)
properties = property(getProperties)
title = property(getTitle)
ACL = property(getACL)
class AtomPubRepository(object):
"""
Represents a CMIS repository. Will lazily populate itself by
calling the repository CMIS service URL.
You must pass in an instance of a CmisClient when creating an
instance of this class.
"""
def __init__(self, cmisClient, xmlDoc=None):
""" Constructor """
self._cmisClient = cmisClient
self.xmlDoc = xmlDoc
self._repositoryId = None
self._repositoryName = None
self._repositoryInfo = {}
self._capabilities = {}
self._uriTemplates = {}
self._permDefs = {}
self._permMap = {}
self._permissions = None
self._propagation = None
self.logger = logging.getLogger('cmislib.model.Repository')
self.logger.info('Creating an instance of Repository')
def __str__(self):
"""To string"""
return self.getRepositoryId()
def reload(self):
"""
This method will re-fetch the repository's XML data from the CMIS
repository.
"""
self.logger.debug('Reload called on object')
self.xmlDoc = self._cmisClient.binding.get(self._cmisClient.repositoryUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password)
self._initData()
def _initData(self):
"""
This method clears out any local variables that would be out of sync
when data is re-fetched from the server.
"""
self._repositoryId = None
self._repositoryName = None
self._repositoryInfo = {}
self._capabilities = {}
self._uriTemplates = {}
self._permDefs = {}
self._permMap = {}
self._permissions = None
self._propagation = None
def getSupportedPermissions(self):
"""
Returns the value of the cmis:supportedPermissions element. Valid
values are:
- basic: indicates that the CMIS Basic permissions are supported
- repository: indicates that repository specific permissions are supported
- both: indicates that both CMIS basic permissions and repository specific permissions are supported
>>> repo.supportedPermissions
u'both'
"""
if not self.getCapabilities()['ACL']:
raise NotSupportedException(messages.NO_ACL_SUPPORT)
if not self._permissions:
if self.xmlDoc is None:
self.reload()
suppEls = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'supportedPermissions')
assert len(suppEls) == 1, 'Expected the repository service document to have one element named supportedPermissions'
self._permissions = suppEls[0].childNodes[0].data
return self._permissions
def getPermissionDefinitions(self):
"""
Returns a dictionary of permission definitions for this repository. The
key is the permission string or technical name of the permission
and the value is the permission description.
>>> for permDef in repo.permissionDefinitions:
... print permDef
...
cmis:all
{http://www.alfresco.org/model/system/1.0}base.LinkChildren
{http://www.alfresco.org/model/content/1.0}folder.Consumer
{http://www.alfresco.org/model/security/1.0}All.All
{http://www.alfresco.org/model/system/1.0}base.CreateAssociations
{http://www.alfresco.org/model/system/1.0}base.FullControl
{http://www.alfresco.org/model/system/1.0}base.AddChildren
{http://www.alfresco.org/model/system/1.0}base.ReadAssociations
{http://www.alfresco.org/model/content/1.0}folder.Editor
{http://www.alfresco.org/model/content/1.0}cmobject.Editor
{http://www.alfresco.org/model/system/1.0}base.DeleteAssociations
cmis:read
cmis:write
"""
if not self.getCapabilities()['ACL']:
raise NotSupportedException(messages.NO_ACL_SUPPORT)
if self._permDefs == {}:
if self.xmlDoc is None:
self.reload()
aclEls = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'aclCapability')
assert len(aclEls) == 1, 'Expected the repository service document to have one element named aclCapability'
aclEl = aclEls[0]
perms = {}
for e in aclEl.childNodes:
if e.localName == 'permissions':
permEls = e.getElementsByTagNameNS(CMIS_NS, 'permission')
assert len(permEls) == 1, 'Expected permissions element to have a child named permission'
descEls = e.getElementsByTagNameNS(CMIS_NS, 'description')
assert len(descEls) == 1, 'Expected permissions element to have a child named description'
perm = permEls[0].childNodes[0].data
desc = descEls[0].childNodes[0].data
perms[perm] = desc
self._permDefs = perms
return self._permDefs
def getPermissionMap(self):
"""
Returns a dictionary representing the permission mapping table where
each key is a permission key string and each value is a list of one or
more permissions the principal must have to perform the operation.
>>> for (k,v) in repo.permissionMap.items():
... print 'To do this: %s, you must have these perms:' % k
... for perm in v:
... print perm
...
To do this: canCreateFolder.Folder, you must have these perms:
cmis:all
{http://www.alfresco.org/model/system/1.0}base.CreateChildren
To do this: canAddToFolder.Folder, you must have these perms:
cmis:all
{http://www.alfresco.org/model/system/1.0}base.CreateChildren
To do this: canDelete.Object, you must have these perms:
cmis:all
{http://www.alfresco.org/model/system/1.0}base.DeleteNode
To do this: canCheckin.Document, you must have these perms:
cmis:all
{http://www.alfresco.org/model/content/1.0}lockable.CheckIn
"""
if not self.getCapabilities()['ACL']:
raise NotSupportedException(messages.NO_ACL_SUPPORT)
if self._permMap == {}:
if self.xmlDoc is None:
self.reload()
aclEls = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'aclCapability')
assert len(aclEls) == 1, 'Expected the repository service document to have one element named aclCapability'
aclEl = aclEls[0]
permMap = {}
for e in aclEl.childNodes:
permList = []
if e.localName == 'mapping':
keyEls = e.getElementsByTagNameNS(CMIS_NS, 'key')
assert len(keyEls) == 1, 'Expected mapping element to have a child named key'
permEls = e.getElementsByTagNameNS(CMIS_NS, 'permission')
assert len(permEls) >= 1, 'Expected mapping element to have at least one permission element'
key = keyEls[0].childNodes[0].data
for permEl in permEls:
permList.append(permEl.childNodes[0].data)
permMap[key] = permList
self._permMap = permMap
return self._permMap
def getPropagation(self):
"""
Returns the value of the cmis:propagation element. Valid values are:
- objectonly: indicates that the repository is able to apply ACEs
without changing the ACLs of other objects
- propagate: indicates that the repository is able to apply ACEs to a
given object and propagate this change to all inheriting objects
>>> repo.propagation
u'propagate'
"""
if not self.getCapabilities()['ACL']:
raise NotSupportedException(messages.NO_ACL_SUPPORT)
if not self._propagation:
if self.xmlDoc is None:
self.reload()
propEls = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'propagation')
assert len(propEls) == 1, 'Expected the repository service document to have one element named propagation'
self._propagation = propEls[0].childNodes[0].data
return self._propagation
def getRepositoryId(self):
"""
Returns this repository's unique identifier
>>> repo = client.getDefaultRepository()
>>> repo.getRepositoryId()
u'83beb297-a6fa-4ac5-844b-98c871c0eea9'
"""
if self._repositoryId is None:
if self.xmlDoc is None:
self.reload()
self._repositoryId = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'repositoryId')[0].firstChild.data
return self._repositoryId
def getRepositoryName(self):
"""
Returns this repository's name
>>> repo = client.getDefaultRepository()
>>> repo.getRepositoryName()
u'Main Repository'
"""
if self._repositoryName is None:
if self.xmlDoc is None:
self.reload()
if self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'repositoryName')[0].firstChild:
self._repositoryName = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'repositoryName')[0].firstChild.data
else:
self._repositoryName = u''
return self._repositoryName
def getRepositoryInfo(self):
"""
Returns a dict of repository information.
>>> repo = client.getDefaultRepository()>>> repo.getRepositoryName()
u'Main Repository'
>>> info = repo.getRepositoryInfo()
>>> for k,v in info.items():
... print "%s:%s" % (k,v)
...
cmisSpecificationTitle:Version 1.0 Committee Draft 04
cmisVersionSupported:1.0
repositoryDescription:None
productVersion:3.2.0 (r2 2440)
rootFolderId:workspace://SpacesStore/aa1ecedf-9551-49c5-831a-0502bb43f348
repositoryId:83beb297-a6fa-4ac5-844b-98c871c0eea9
repositoryName:Main Repository
vendorName:Alfresco
productName:Alfresco Repository (Community)
"""
if not self._repositoryInfo:
if self.xmlDoc is None:
self.reload()
repoInfoElement = self.xmlDoc.getElementsByTagNameNS(CMISRA_NS, 'repositoryInfo')[0]
for node in repoInfoElement.childNodes:
if node.nodeType == node.ELEMENT_NODE and \
node.localName != 'capabilities' and \
node.localName != 'aclCapability':
try:
data = node.childNodes[0].data
except IndexError:
data = None
except AttributeError:
data = None
self._repositoryInfo[node.localName] = data
return self._repositoryInfo
def getCapabilities(self):
"""
Returns a dict of repository capabilities.
>>> caps = repo.getCapabilities()
>>> for k,v in caps.items():
... print "%s:%s" % (k,v)
...
PWCUpdatable:True
VersionSpecificFiling:False
Join:None
ContentStreamUpdatability:anytime
AllVersionsSearchable:False
Renditions:None
Multifiling:True
GetFolderTree:True
GetDescendants:True
ACL:None
PWCSearchable:True
Query:bothcombined
Unfiling:False
Changes:None
"""
if not self._capabilities:
if self.xmlDoc is None:
self.reload()
capabilitiesElement = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'capabilities')[0]
for node in [e for e in capabilitiesElement.childNodes if e.nodeType == e.ELEMENT_NODE]:
key = node.localName.replace('capability', '')
value = parseBoolValue(node.childNodes[0].data)
self._capabilities[key] = value
return self._capabilities
def getRootFolder(self):
"""
Returns the root folder of the repository
>>> root = repo.getRootFolder()
>>> root.getObjectId()
u'workspace://SpacesStore/aa1ecedf-9551-49c5-831a-0502bb43f348'
"""
# get the root folder id
rootFolderId = self.getRepositoryInfo()['rootFolderId']
# instantiate a Folder object using the ID
folder = AtomPubFolder(self._cmisClient, self, rootFolderId)
# return it
return folder
def getFolder(self, folderId):
"""
Returns a :class:`Folder` object for a specified folderId
>>> someFolder = repo.getFolder('workspace://SpacesStore/aa1ecedf-9551-49c5-831a-0502bb43f348')
>>> someFolder.getObjectId()
u'workspace://SpacesStore/aa1ecedf-9551-49c5-831a-0502bb43f348'
"""
retObject = self.getObject(folderId)
return AtomPubFolder(self._cmisClient, self, xmlDoc=retObject.xmlDoc)
def getTypeChildren(self,
typeId=None):
"""
Returns a list of :class:`ObjectType` objects corresponding to the
child types of the type specified by the typeId.
If no typeId is provided, the result will be the same as calling
`self.getTypeDefinitions`
These optional arguments are current unsupported:
- includePropertyDefinitions
- maxItems
- skipCount
>>> baseTypes = repo.getTypeChildren()
>>> for baseType in baseTypes:
... print baseType.getTypeId()
...
cmis:folder
cmis:relationship
cmis:document
cmis:policy
"""
# Unfortunately, the spec does not appear to present a way to
# know how to get the children of a specific type without first
# retrieving the type, then asking it for one of its navigational
# links.
# if a typeId is specified, get it, then get its "down" link
if typeId:
targetType = self.getTypeDefinition(typeId)
childrenUrl = targetType.getLink(DOWN_REL, ATOM_XML_FEED_TYPE_P)
typesXmlDoc = self._cmisClient.binding.get(childrenUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password)
entryElements = typesXmlDoc.getElementsByTagNameNS(ATOM_NS, 'entry')
types = []
for entryElement in entryElements:
objectType = AtomPubObjectType(self._cmisClient,
self,
xmlDoc=entryElement)
types.append(objectType)
# otherwise, if a typeId is not specified, return
# the list of base types
else:
types = self.getTypeDefinitions()
return types
def getTypeDescendants(self, typeId=None, **kwargs):
"""
Returns a list of :class:`ObjectType` objects corresponding to the
descendant types of the type specified by the typeId.
If no typeId is provided, the repository's "typesdescendants" URL
will be called to determine the list of descendant types.
>>> allTypes = repo.getTypeDescendants()
>>> for aType in allTypes:
... print aType.getTypeId()
...
cmis:folder
F:cm:systemfolder
F:act:savedactionfolder
F:app:configurations
F:fm:forums
F:wcm:avmfolder
F:wcm:avmplainfolder
F:wca:webfolder
F:wcm:avmlayeredfolder
F:st:site
F:app:glossary
F:fm:topic
These optional arguments are supported:
- depth
- includePropertyDefinitions
>>> types = repo.getTypeDescendants('cmis:folder')
>>> len(types)
17
>>> types = repo.getTypeDescendants('cmis:folder', depth=1)
>>> len(types)
12
>>> types = repo.getTypeDescendants('cmis:folder', depth=2)
>>> len(types)
17
"""
# Unfortunately, the spec does not appear to present a way to
# know how to get the children of a specific type without first
# retrieving the type, then asking it for one of its navigational
# links.
if typeId:
targetType = self.getTypeDefinition(typeId)
descendUrl = targetType.getLink(DOWN_REL, CMIS_TREE_TYPE_P)
else:
descendUrl = self.getLink(TYPE_DESCENDANTS_REL)
if not descendUrl:
raise NotSupportedException("Could not determine the type descendants URL")
typesXmlDoc = self._cmisClient.binding.get(descendUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
entryElements = typesXmlDoc.getElementsByTagNameNS(ATOM_NS, 'entry')
types = []
for entryElement in entryElements:
objectType = AtomPubObjectType(self._cmisClient,
self,
xmlDoc=entryElement)
types.append(objectType)
return types
def getTypeDefinitions(self, **kwargs):
"""
Returns a list of :class:`ObjectType` objects representing
the base types in the repository.
>>> baseTypes = repo.getTypeDefinitions()
>>> for baseType in baseTypes:
... print baseType.getTypeId()
...
cmis:folder
cmis:relationship
cmis:document
cmis:policy
"""
typesUrl = self.getCollectionLink(TYPES_COLL)
typesXmlDoc = self._cmisClient.binding.get(typesUrl,
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
entryElements = typesXmlDoc.getElementsByTagNameNS(ATOM_NS, 'entry')
types = []
for entryElement in entryElements:
objectType = AtomPubObjectType(self._cmisClient,
self,
xmlDoc=entryElement)
types.append(objectType)
# return the result
return types
def getTypeDefinition(self, typeId):
"""
Returns an :class:`ObjectType` object for the specified object type id.
>>> folderType = repo.getTypeDefinition('cmis:folder')
"""
objectType = AtomPubObjectType(self._cmisClient, self, typeId)
objectType.reload()
return objectType
def getLink(self, rel):
"""
Returns the HREF attribute of an Atom link element for the
specified rel.
"""
if self.xmlDoc is None:
self.reload()
linkElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'link')
for linkElement in linkElements:
if linkElement.attributes.has_key('rel'):
relAttr = linkElement.attributes['rel'].value
if relAttr == rel:
return linkElement.attributes['href'].value
def getCheckedOutDocs(self, **kwargs):
"""
Returns a ResultSet of :class:`CmisObject` objects that
are currently checked out.
>>> rs = repo.getCheckedOutDocs()
>>> len(rs.getResults())
2
>>> for doc in repo.getCheckedOutDocs().getResults():
... doc.getTitle()
...
u'sample-a (Working Copy).pdf'
u'sample-b (Working Copy).pdf'
These optional arguments are supported:
- folderId
- maxItems
- skipCount
- orderBy
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
"""
return self.getCollection(CHECKED_OUT_COLL, **kwargs)
def getUnfiledDocs(self, **kwargs):
"""
Returns a ResultSet of :class:`CmisObject` objects that
are currently unfiled.
>>> rs = repo.getUnfiledDocs()
>>> len(rs.getResults())
2
>>> for doc in repo.getUnfiledDocs().getResults():
... doc.getTitle()
...
u'sample-a.pdf'
u'sample-b.pdf'
These optional arguments are supported:
- folderId
- maxItems
- skipCount
- orderBy
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
"""
return self.getCollection(UNFILED_COLL, **kwargs)
def getObject(self,
objectId,
**kwargs):
"""
Returns an object given the specified object ID.
>>> doc = repo.getObject('workspace://SpacesStore/f0c8b90f-bec0-4405-8b9c-2ab570589808')
>>> doc.getTitle()
u'sample-b.pdf'
The following optional arguments are supported:
- returnVersion
- filter
- includeRelationships
- includePolicyIds
- renditionFilter
- includeACL
- includeAllowableActions
"""
return getSpecializedObject(AtomPubCmisObject(self._cmisClient, self, CmisId(objectId), **kwargs), **kwargs)
def getObjectByPath(self, path, **kwargs):
"""
Returns an object given the path to the object.
>>> doc = repo.getObjectByPath('/jeff test/sample-b.pdf')
>>> doc.getTitle()
u'sample-b.pdf'
The following optional arguments are not currently supported:
- filter
- includeAllowableActions
"""
# get the uritemplate
template = self.getUriTemplates()['objectbypath']['template']
# fill in the template with the path provided
params = {'{path}': quote(path, '/'),
'{filter}': '',
'{includeAllowableActions}': 'false',
'{includePolicyIds}': 'false',
'{includeRelationships}': '',
'{includeACL}': 'false',
'{renditionFilter}': ''}
options = {}
addOptions = {} # args specified, but not in the template
for k, v in kwargs.items():
pKey = "{" + k + "}"
if template.find(pKey) >= 0:
options[pKey] = toCMISValue(v)
else:
addOptions[k] = toCMISValue(v)
# merge the templated args with the default params
params.update(options)
byObjectPathUrl = multiple_replace(params, template)
# do a GET against the URL
result = self._cmisClient.binding.get(byObjectPathUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**addOptions)
# instantiate CmisObject objects with the results and return the list
entryElements = result.getElementsByTagNameNS(ATOM_NS, 'entry')
assert(len(entryElements) == 1), "Expected entry element in result from calling %s" % byObjectPathUrl
return getSpecializedObject(AtomPubCmisObject(self._cmisClient, self, xmlDoc=entryElements[0], **kwargs), **kwargs)
def query(self, statement, **kwargs):
"""
Returns a list of :class:`CmisObject` objects based on the CMIS
Query Language passed in as the statement. The actual objects
returned will be instances of the appropriate child class based
on the object's base type ID.
In order for the results to be properly instantiated as objects,
make sure you include 'cmis:objectId' as one of the fields in
your select statement, or just use "SELECT \*".
If you want the search results to automatically be instantiated with
the appropriate sub-class of :class:`CmisObject` you must either
include cmis:baseTypeId as one of the fields in your select statement
or just use "SELECT \*".
>>> q = "select * from cmis:document where cmis:name like '%test%'"
>>> resultSet = repo.query(q)
>>> len(resultSet.getResults())
1
>>> resultSet.hasNext()
False
The following optional arguments are supported:
- searchAllVersions
- includeRelationships
- renditionFilter
- includeAllowableActions
- maxItems
- skipCount
>>> q = 'select * from cmis:document'
>>> rs = repo.query(q)
>>> len(rs.getResults())
148
>>> rs = repo.query(q, maxItems='5')
>>> len(rs.getResults())
5
>>> rs.hasNext()
True
"""
if self.xmlDoc is None:
self.reload()
# get the URL this repository uses to accept query POSTs
queryUrl = self.getCollectionLink(QUERY_COLL)
# build the CMIS query XML that we're going to POST
xmlDoc = self._getQueryXmlDoc(statement, **kwargs)
# do the POST
# print 'posting:%s' % xmlDoc.toxml(encoding='utf-8')
result = self._cmisClient.binding.post(queryUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
xmlDoc.toxml(encoding='utf-8'),
CMIS_QUERY_TYPE)
# return the result set
return AtomPubResultSet(self._cmisClient, self, result)
def getContentChanges(self, **kwargs):
"""
Returns a :class:`ResultSet` containing :class:`ChangeEntry` objects.
>>> for changeEntry in rs:
... changeEntry.objectId
... changeEntry.id
... changeEntry.changeType
... changeEntry.changeTime
...
'workspace://SpacesStore/0e2dc775-16b7-4634-9e54-2417a196829b'
u'urn:uuid:0e2dc775-16b7-4634-9e54-2417a196829b'
u'created'
datetime.datetime(2010, 2, 11, 12, 55, 14)
'workspace://SpacesStore/bd768f9f-99a7-4033-828d-5b13f96c6923'
u'urn:uuid:bd768f9f-99a7-4033-828d-5b13f96c6923'
u'updated'
datetime.datetime(2010, 2, 11, 12, 55, 13)
'workspace://SpacesStore/572c2cac-6b26-4cd8-91ad-b2931fe5b3fb'
u'urn:uuid:572c2cac-6b26-4cd8-91ad-b2931fe5b3fb'
u'updated'
The following optional arguments are supported:
- changeLogToken
- includeProperties
- includePolicyIDs
- includeACL
- maxItems
You can get the latest change log token by inspecting the repository
info via :meth:`Repository.getRepositoryInfo`.
>>> repo.info['latestChangeLogToken']
u'2692'
>>> rs = repo.getContentChanges(changeLogToken='2692')
>>> len(rs)
1
>>> rs[0].id
u'urn:uuid:8e88f694-93ef-44c5-9f70-f12fff824be9'
>>> rs[0].changeType
u'updated'
>>> rs[0].changeTime
datetime.datetime(2010, 2, 16, 20, 6, 37)
"""
if self.getCapabilities()['Changes'] is None:
raise NotSupportedException(messages.NO_CHANGE_LOG_SUPPORT)
changesUrl = self.getLink(CHANGE_LOG_REL)
result = self._cmisClient.binding.get(changesUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubChangeEntryResultSet(self._cmisClient, self, result)
def createDocumentFromString(self,
name,
properties={},
parentFolder=None,
contentString=None,
contentType=None,
contentEncoding=None):
"""
Creates a new document setting the content to the string provided. If
the repository supports unfiled objects, you do not have to pass in
a parent :class:`Folder` otherwise it is required.
This method is essentially a convenience method that wraps your string
with a StringIO and then calls createDocument.
>>> repo.createDocumentFromString('testdoc5', parentFolder=testFolder, contentString='Hello, World!', contentType='text/plain')
<cmislib.model.Document object at 0x101352ed0>
"""
# if you didn't pass in a parent folder
if parentFolder is None:
# if the repository doesn't require fileable objects to be filed
if self.getCapabilities()['Unfiling']:
# has not been implemented
# postUrl = self.getCollectionLink(UNFILED_COLL)
raise NotImplementedError
else:
# this repo requires fileable objects to be filed
raise InvalidArgumentException
return parentFolder.createDocument(name, properties, StringIO.StringIO(contentString),
contentType, contentEncoding)
def createDocument(self,
name,
properties={},
parentFolder=None,
contentFile=None,
contentType=None,
contentEncoding=None):
"""
Creates a new :class:`Document` object. If the repository
supports unfiled objects, you do not have to pass in
a parent :class:`Folder` otherwise it is required.
To create a document with an associated contentFile, pass in a
File object. The method will attempt to guess the appropriate content
type and encoding based on the file. To specify it yourself, pass them
in via the contentType and contentEncoding arguments.
>>> f = open('sample-a.pdf', 'rb')
>>> doc = folder.createDocument('sample-a.pdf', contentFile=f)
<cmislib.model.Document object at 0x105be5e10>
>>> f.close()
>>> doc.getTitle()
u'sample-a.pdf'
The following optional arguments are not currently supported:
- versioningState
- policies
- addACEs
- removeACEs
"""
postUrl = ''
# if you didn't pass in a parent folder
if parentFolder is None:
# if the repository doesn't require fileable objects to be filed
if self.getCapabilities()['Unfiling']:
# has not been implemented
# postUrl = self.getCollectionLink(UNFILED_COLL)
raise NotImplementedError
else:
# this repo requires fileable objects to be filed
raise InvalidArgumentException
else:
postUrl = parentFolder.getChildrenLink()
# make sure a name is set
properties['cmis:name'] = name
# hardcoding to cmis:document if it wasn't
# passed in via props
if not properties.has_key('cmis:objectTypeId'):
properties['cmis:objectTypeId'] = CmisId('cmis:document')
# and if it was passed in, making sure it is a CmisId
elif not isinstance(properties['cmis:objectTypeId'], CmisId):
properties['cmis:objectTypeId'] = CmisId(properties['cmis:objectTypeId'])
# build the Atom entry
xmlDoc = getEntryXmlDoc(self, None, properties, contentFile,
contentType, contentEncoding)
# post the Atom entry
result = self._cmisClient.binding.post(postUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
xmlDoc.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE)
# what comes back is the XML for the new document,
# so use it to instantiate a new document
# then return it
return AtomPubDocument(self._cmisClient, self, xmlDoc=result)
def createDocumentFromSource(self,
sourceId,
properties={},
parentFolder=None):
"""
This is not yet implemented.
The following optional arguments are not yet supported:
- versioningState
- policies
- addACEs
- removeACEs
"""
# TODO: To be implemented
raise NotImplementedError
def createFolder(self,
parentFolder,
name,
properties={}):
"""
Creates a new :class:`Folder` object in the specified parentFolder.
>>> root = repo.getRootFolder()
>>> folder = repo.createFolder(root, 'someFolder2')
>>> folder.getTitle()
u'someFolder2'
>>> folder.getObjectId()
u'workspace://SpacesStore/2224a63c-350b-438c-be72-8f425e79ce1f'
The following optional arguments are not yet supported:
- policies
- addACEs
- removeACEs
"""
return parentFolder.createFolder(name, properties)
def createRelationship(self, sourceObj, targetObj, relType):
"""
Creates a relationship of the specific type between a source object
and a target object and returns the new :class:`Relationship` object.
The following optional arguments are not currently supported:
- policies
- addACEs
- removeACEs
"""
return sourceObj.createRelationship(targetObj, relType)
def createPolicy(self, properties):
"""
This has not yet been implemented.
The following optional arguments are not currently supported:
- folderId
- policies
- addACEs
- removeACEs
"""
# TODO: To be implemented
raise NotImplementedError
def getUriTemplates(self):
"""
Returns a list of the URI templates the repository service knows about.
>>> templates = repo.getUriTemplates()
>>> templates['typebyid']['mediaType']
u'application/atom+xml;type=entry'
>>> templates['typebyid']['template']
u'http://localhost:8080/alfresco/s/cmis/type/{id}'
"""
if self._uriTemplates == {}:
if self.xmlDoc is None:
self.reload()
uriTemplateElements = self.xmlDoc.getElementsByTagNameNS(CMISRA_NS, 'uritemplate')
for uriTemplateElement in uriTemplateElements:
template = None
templType = None
mediatype = None
for node in [e for e in uriTemplateElement.childNodes if e.nodeType == e.ELEMENT_NODE]:
if node.localName == 'template':
template = node.childNodes[0].data
elif node.localName == 'type':
templType = node.childNodes[0].data
elif node.localName == 'mediatype':
mediatype = node.childNodes[0].data
self._uriTemplates[templType] = UriTemplate(template,
templType,
mediatype)
return self._uriTemplates
def getCollection(self, collectionType, **kwargs):
"""
Returns a list of objects returned for the specified collection.
If the query collection is requested, an exception will be raised.
That collection isn't meant to be retrieved.
If the types collection is specified, the method returns the result of
`getTypeDefinitions` and ignores any optional params passed in.
>>> from cmislib.atompub.atompub_binding import TYPES_COLL
>>> types = repo.getCollection(TYPES_COLL)
>>> len(types)
4
>>> types[0].getTypeId()
u'cmis:folder'
Otherwise, the collection URL is invoked, and a :class:`ResultSet` is
returned.
>>> from cmislib.atompub.atompub_binding import CHECKED_OUT_COLL
>>> resultSet = repo.getCollection(CHECKED_OUT_COLL)
>>> len(resultSet.getResults())
1
"""
if collectionType == QUERY_COLL:
raise NotSupportedException
elif collectionType == TYPES_COLL:
return self.getTypeDefinitions()
result = self._cmisClient.binding.get(self.getCollectionLink(collectionType).encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self, result)
def getCollectionLink(self, collectionType):
"""
Returns the link HREF from the specified collectionType
('checkedout', for example).
>>> from cmislib.atompub.atompub_binding import CHECKED_OUT_COLL
>>> repo.getCollectionLink(CHECKED_OUT_COLL)
u'http://localhost:8080/alfresco/s/cmis/checkedout'
"""
collectionElements = self.xmlDoc.getElementsByTagNameNS(APP_NS, 'collection')
for collectionElement in collectionElements:
link = collectionElement.attributes['href'].value
for node in [e for e in collectionElement.childNodes if e.nodeType == e.ELEMENT_NODE]:
if node.localName == 'collectionType':
if node.childNodes[0].data == collectionType:
return link
def _getQueryXmlDoc(self, query, **kwargs):
"""
Utility method that knows how to build CMIS query xml around the
specified query statement.
"""
cmisXmlDoc = minidom.Document()
queryElement = cmisXmlDoc.createElementNS(CMIS_NS, "query")
queryElement.setAttribute('xmlns', CMIS_NS)
cmisXmlDoc.appendChild(queryElement)
statementElement = cmisXmlDoc.createElementNS(CMIS_NS, "statement")
# CMIS-703
# cdataSection = cmisXmlDoc.createCDATASection(query)
# statementElement.appendChild(cdataSection)
textNode = cmisXmlDoc.createTextNode(query)
statementElement.appendChild(textNode)
queryElement.appendChild(statementElement)
for (k, v) in kwargs.items():
optionElement = cmisXmlDoc.createElementNS(CMIS_NS, k)
optionText = cmisXmlDoc.createTextNode(v)
optionElement.appendChild(optionText)
queryElement.appendChild(optionElement)
return cmisXmlDoc
capabilities = property(getCapabilities)
id = property(getRepositoryId)
info = property(getRepositoryInfo)
name = property(getRepositoryName)
rootFolder = property(getRootFolder)
permissionDefinitions = property(getPermissionDefinitions)
permissionMap = property(getPermissionMap)
propagation = property(getPropagation)
supportedPermissions = property(getSupportedPermissions)
class AtomPubResultSet(ResultSet):
"""
Represents a paged result set. In CMIS, this is most often an Atom feed.
"""
def __init__(self, cmisClient, repository, xmlDoc):
""" Constructor """
self._cmisClient = cmisClient
self._repository = repository
self._xmlDoc = xmlDoc
self._results = []
self.logger = logging.getLogger('cmislib.model.ResultSet')
self.logger.info('Creating an instance of ResultSet')
def __iter__(self):
""" Iterator for the result set """
return iter(self.getResults())
def __getitem__(self, index):
""" Getter for the result set """
return self.getResults()[index]
def __len__(self):
""" Len method for the result set """
return len(self.getResults())
def _getLink(self, rel):
"""
Returns the link found in the feed's XML for the specified rel.
"""
linkElements = self._xmlDoc.getElementsByTagNameNS(ATOM_NS, 'link')
for linkElement in linkElements:
if linkElement.attributes.has_key('rel'):
relAttr = linkElement.attributes['rel'].value
if relAttr == rel:
return linkElement.attributes['href'].value
def _getPageResults(self, rel):
"""
Given a specified rel, does a get using that link (if one exists)
and then converts the resulting XML into a dictionary of
:class:`CmisObject` objects or its appropriate sub-type.
The results are kept around to facilitate repeated calls without moving
the cursor.
"""
link = self._getLink(rel)
if link:
result = self._cmisClient.binding.get(link.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password)
# return the result
self._xmlDoc = result
self._results = []
return self.getResults()
def reload(self):
"""
Re-invokes the self link for the current set of results.
>>> resultSet = repo.getCollection(CHECKED_OUT_COLL)
>>> resultSet.reload()
"""
self.logger.debug('Reload called on result set')
self._getPageResults(SELF_REL)
def getResults(self):
"""
Returns the results that were fetched and cached by the get*Page call.
>>> resultSet = repo.getCheckedOutDocs()
>>> resultSet.hasNext()
False
>>> for result in resultSet.getResults():
... result
...
<cmislib.model.Document object at 0x104851810>
"""
if self._results:
return self._results
if self._xmlDoc:
entryElements = self._xmlDoc.getElementsByTagNameNS(ATOM_NS, 'entry')
entries = []
for entryElement in entryElements:
cmisObject = getSpecializedObject(AtomPubCmisObject(self._cmisClient,
self._repository,
xmlDoc=entryElement))
entries.append(cmisObject)
self._results = entries
return self._results
def hasObject(self, objectId):
"""
Returns True if the specified objectId is found in the list of results,
otherwise returns False.
"""
for obj in self.getResults():
if obj.id == objectId:
return True
return False
def getFirst(self):
"""
Returns the first page of results as a dictionary of
:class:`CmisObject` objects or its appropriate sub-type. This only
works when the server returns a "first" link. Not all of them do.
>>> resultSet.hasFirst()
True
>>> results = resultSet.getFirst()
>>> for result in results:
... result
...
<cmislib.model.Document object at 0x10480bc90>
"""
return self._getPageResults(FIRST_REL)
def getPrev(self):
"""
Returns the prev page of results as a dictionary of
:class:`CmisObject` objects or its appropriate sub-type. This only
works when the server returns a "prev" link. Not all of them do.
>>> resultSet.hasPrev()
True
>>> results = resultSet.getPrev()
>>> for result in results:
... result
...
<cmislib.model.Document object at 0x10480bc90>
"""
return self._getPageResults(PREV_REL)
def getNext(self):
"""
Returns the next page of results as a dictionary of
:class:`CmisObject` objects or its appropriate sub-type.
>>> resultSet.hasNext()
True
>>> results = resultSet.getNext()
>>> for result in results:
... result
...
<cmislib.model.Document object at 0x10480bc90>
"""
return self._getPageResults(NEXT_REL)
def getLast(self):
"""
Returns the last page of results as a dictionary of
:class:`CmisObject` objects or its appropriate sub-type. This only
works when the server is returning a "last" link. Not all of them do.
>>> resultSet.hasLast()
True
>>> results = resultSet.getLast()
>>> for result in results:
... result
...
<cmislib.model.Document object at 0x10480bc90>
"""
return self._getPageResults(LAST_REL)
def hasNext(self):
"""
Returns True if this page contains a next link.
>>> resultSet.hasNext()
True
"""
if self._getLink(NEXT_REL):
return True
else:
return False
def hasPrev(self):
"""
Returns True if this page contains a prev link. Not all CMIS providers
implement prev links consistently.
>>> resultSet.hasPrev()
True
"""
if self._getLink(PREV_REL):
return True
else:
return False
def hasFirst(self):
"""
Returns True if this page contains a first link. Not all CMIS providers
implement first links consistently.
>>> resultSet.hasFirst()
True
"""
if self._getLink(FIRST_REL):
return True
else:
return False
def hasLast(self):
"""
Returns True if this page contains a last link. Not all CMIS providers
implement last links consistently.
>>> resultSet.hasLast()
True
"""
if self._getLink(LAST_REL):
return True
else:
return False
class AtomPubDocument(AtomPubCmisObject):
"""
An object typically associated with file content.
"""
def checkout(self):
"""
Performs a checkout on the :class:`Document` and returns the
Private Working Copy (PWC), which is also an instance of
:class:`Document`
>>> doc.getObjectId()
u'workspace://SpacesStore/f0c8b90f-bec0-4405-8b9c-2ab570589808;1.0'
>>> doc.isCheckedOut()
False
>>> pwc = doc.checkout()
>>> doc.isCheckedOut()
True
"""
# get the checkedout collection URL
checkoutUrl = self._repository.getCollectionLink(CHECKED_OUT_COLL)
assert len(checkoutUrl) > 0, "Could not determine the checkedout collection url."
# get this document's object ID
# build entry XML with it
properties = {'cmis:objectId': self.getObjectId()}
entryXmlDoc = getEntryXmlDoc(self._repository, properties=properties)
# post it to to the checkedout collection URL
result = self._cmisClient.binding.post(checkoutUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
entryXmlDoc.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE)
# now that the doc is checked out, we need to refresh the XML
# to pick up the prop updates related to a checkout
self.reload()
return AtomPubDocument(self._cmisClient, self._repository, xmlDoc=result)
def cancelCheckout(self):
"""
Cancels the checkout of this object by retrieving the Private Working
Copy (PWC) and then deleting it. After the PWC is deleted, this object
will be reloaded to update properties related to a checkout.
>>> doc.isCheckedOut()
True
>>> doc.cancelCheckout()
>>> doc.isCheckedOut()
False
"""
pwcDoc = self.getPrivateWorkingCopy()
if pwcDoc:
pwcDoc.delete()
self.reload()
def getPrivateWorkingCopy(self):
"""
Retrieves the object using the object ID in the property:
cmis:versionSeriesCheckedOutId then uses getObject to instantiate
the object.
>>> doc.isCheckedOut()
False
>>> doc.checkout()
<cmislib.model.Document object at 0x103a25ad0>
>>> pwc = doc.getPrivateWorkingCopy()
>>> pwc.getTitle()
u'sample-b (Working Copy).pdf'
"""
# reloading the document just to make sure we've got the latest
# and greatest PWC ID
self.reload()
pwcDocId = self.getProperties()['cmis:versionSeriesCheckedOutId']
if pwcDocId:
return self._repository.getObject(pwcDocId)
def isCheckedOut(self):
"""
Returns true if the document is checked out.
>>> doc.isCheckedOut()
True
>>> doc.cancelCheckout()
>>> doc.isCheckedOut()
False
"""
# reloading the document just to make sure we've got the latest
# and greatest checked out prop
self.reload()
return parseBoolValue(self.getProperties()['cmis:isVersionSeriesCheckedOut'])
def getCheckedOutBy(self):
"""
Returns the ID who currently has the document checked out.
>>> pwc = doc.checkout()
>>> pwc.getCheckedOutBy()
u'admin'
"""
# reloading the document just to make sure we've got the latest
# and greatest checked out prop
self.reload()
return self.getProperties()['cmis:versionSeriesCheckedOutBy']
def checkin(self, checkinComment=None, **kwargs):
"""
Checks in this :class:`Document` which must be a private
working copy (PWC).
>>> doc.isCheckedOut()
False
>>> pwc = doc.checkout()
>>> doc.isCheckedOut()
True
>>> pwc.checkin()
<cmislib.model.Document object at 0x103a8ae90>
>>> doc.isCheckedOut()
False
The following optional arguments are supported:
- major
- properties
- contentStream
- policies
- addACEs
- removeACEs
"""
# major = true is supposed to be the default but inmemory 0.9 is throwing an error 500 without it
if not kwargs.has_key('major'):
kwargs['major'] = 'true'
# Add checkin to kwargs and checkinComment, if it exists
kwargs['checkin'] = 'true'
kwargs['checkinComment'] = checkinComment
# Build an empty ATOM entry
entryXmlDoc = getEmptyXmlDoc()
# Get the self link
# Do a PUT of the empty ATOM to the self link
url = self._getSelfLink()
result = self._cmisClient.binding.put(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
entryXmlDoc.toxml(encoding='utf-8'),
ATOM_XML_TYPE,
**kwargs)
return AtomPubDocument(self._cmisClient, self._repository, xmlDoc=result)
def getLatestVersion(self, **kwargs):
"""
Returns a :class:`Document` object representing the latest version in
the version series.
The following optional arguments are supported:
- major
- filter
- includeRelationships
- includePolicyIds
- renditionFilter
- includeACL
- includeAllowableActions
>>> latestDoc = doc.getLatestVersion()
>>> latestDoc.getProperties()['cmis:versionLabel']
u'2.1'
>>> latestDoc = doc.getLatestVersion(major='false')
>>> latestDoc.getProperties()['cmis:versionLabel']
u'2.1'
>>> latestDoc = doc.getLatestVersion(major='true')
>>> latestDoc.getProperties()['cmis:versionLabel']
u'2.0'
"""
doc = None
if kwargs.has_key('major') and kwargs['major'] == 'true':
doc = self._repository.getObject(self.getObjectId(), returnVersion='latestmajor')
else:
doc = self._repository.getObject(self.getObjectId(), returnVersion='latest')
return doc
def getPropertiesOfLatestVersion(self, **kwargs):
"""
Like :class:`^CmisObject.getProperties`, returns a dict of properties
from the latest version of this object in the version series.
The optional major and filter arguments are supported.
"""
latestDoc = self.getLatestVersion(**kwargs)
return latestDoc.getProperties()
def getAllVersions(self, **kwargs):
"""
Returns a :class:`ResultSet` of document objects for the entire
version history of this object, including any PWC's.
The optional filter and includeAllowableActions are
supported.
"""
# get the version history link
versionsUrl = self._getLink(VERSION_HISTORY_REL)
# invoke the URL
result = self._cmisClient.binding.get(versionsUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def getContentStream(self):
"""
Returns the CMIS service response from invoking the 'enclosure' link.
>>> doc.getName()
u'sample-b.pdf'
>>> o = open('tmp.pdf', 'wb')
>>> result = doc.getContentStream()
>>> o.write(result.read())
>>> result.close()
>>> o.close()
>>> import os.path
>>> os.path.getsize('tmp.pdf')
117248
The optional streamId argument is not yet supported.
"""
# TODO: Need to implement the streamId
contentElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'content')
# CMIS-701
if len(contentElements) != 1:
self.reload()
contentElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'content')
assert(len(contentElements) == 1), 'Expected to find exactly one atom:content element.'
# if the src element exists, follow that
if contentElements[0].attributes.has_key('src'):
srcUrl = contentElements[0].attributes['src'].value
# the cmis client class parses non-error responses
result, content = Rest().get(srcUrl.encode('utf-8'),
username=self._cmisClient.username,
password=self._cmisClient.password,
**self._cmisClient.extArgs)
if result['status'] != '200':
raise CmisException(result['status'])
return StringIO.StringIO(content)
else:
# otherwise, try to return the value of the content element
if contentElements[0].childNodes:
return contentElements[0].childNodes[0].data
def setContentStream(self, contentFile, contentType=None):
"""
Sets the content stream on this object.
The following optional arguments are not yet supported:
- overwriteFlag=None
"""
# get this object's content stream link
contentElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'content')
assert(len(contentElements) == 1), 'Expected to find exactly one atom:content element.'
# if the src element exists, follow that
if contentElements[0].attributes.has_key('src'):
srcUrl = contentElements[0].attributes['src'].value
# there may be times when this URL is absent, but I'm not sure how to
# set the content stream when that is the case
assert srcUrl, 'Unable to determine content stream URL.'
# need to determine the mime type
mimetype = contentType
if not mimetype and hasattr(contentFile, 'name'):
mimetype, encoding = mimetypes.guess_type(contentFile.name)
if not mimetype:
mimetype = 'application/binary'
# if we have a change token, we must pass it back, per the spec
args = {}
if self.properties.has_key('cmis:changeToken') and self.properties['cmis:changeToken'] is not None:
self.logger.debug('Change token present, adding it to args')
args = {"changeToken": self.properties['cmis:changeToken']}
# put the content file
result = self._cmisClient.binding.put(srcUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
contentFile.read(),
mimetype,
**args)
# what comes back is the XML for the updated document,
# which is not required by the spec to be the same document
# we just updated, so use it to instantiate a new document
# then return it
return AtomPubDocument(self._cmisClient, self._repository, xmlDoc=result)
def deleteContentStream(self):
"""
Delete's the content stream associated with this object.
"""
# get this object's content stream link
contentElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'content')
assert(len(contentElements) == 1), 'Expected to find exactly one atom:content element.'
# if the src element exists, follow that
if contentElements[0].attributes.has_key('src'):
srcUrl = contentElements[0].attributes['src'].value
# there may be times when this URL is absent, but I'm not sure how to
# delete the content stream when that is the case
assert srcUrl, 'Unable to determine content stream URL.'
# if we have a change token, we must pass it back, per the spec
args = {}
if self.properties.has_key('cmis:changeToken') and self.properties['cmis:changeToken'] is not None:
self.logger.debug('Change token present, adding it to args')
args = {"changeToken": self.properties['cmis:changeToken']}
# delete the content stream
self._cmisClient.binding.delete(srcUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**args)
checkedOut = property(isCheckedOut)
def getPaths(self):
"""
Returns the Document's paths by asking for the parents with the
includeRelativePathSegment flag set to true, then concats the value
of cmis:path with the relativePathSegment.
"""
# get the appropriate 'up' link
parentUrl = self._getLink(UP_REL)
if parentUrl is None:
raise NotSupportedException('Root folder does not support getObjectParents')
# invoke the URL
result = self._cmisClient.binding.get(parentUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
filter='cmis:path',
includeRelativePathSegment=True)
paths = []
rs = AtomPubResultSet(self._cmisClient, self._repository, result)
for res in rs:
path = res.properties['cmis:path']
relativePathSegment = res.properties['cmisra:relativePathSegment']
# concat with a slash
# add it to the list
paths.append(path + '/' + relativePathSegment)
return paths
class AtomPubFolder(AtomPubCmisObject):
"""
A container object that can hold other :class:`CmisObject` objects
"""
def createFolder(self, name, properties={}):
"""
Creates a new :class:`Folder` using the properties provided.
Right now I expect a property called 'cmis:name' but I don't
complain if it isn't there (although the CMIS provider will). If a
cmis:name property isn't provided, the value passed in to the name
argument will be used.
To specify a custom folder type, pass in a property called
cmis:objectTypeId set to the :class:`CmisId` representing the type ID
of the instance you want to create. If you do not pass in an object
type ID, an instance of 'cmis:folder' will be created.
>>> subFolder = folder.createFolder('someSubfolder')
>>> subFolder.getName()
u'someSubfolder'
The following optional arguments are not supported:
- policies
- addACEs
- removeACEs
"""
# get the folder represented by folderId.
# we'll use his 'children' link post the new child
postUrl = self.getChildrenLink()
# make sure the name property gets set
properties['cmis:name'] = name
# hardcoding to cmis:folder if it wasn't passed in via props
if not properties.has_key('cmis:objectTypeId'):
properties['cmis:objectTypeId'] = CmisId('cmis:folder')
# and checking to make sure the object type ID is an instance of CmisId
elif not isinstance(properties['cmis:objectTypeId'], CmisId):
properties['cmis:objectTypeId'] = CmisId(properties['cmis:objectTypeId'])
# build the Atom entry
entryXml = getEntryXmlDoc(self._repository, properties=properties)
# post the Atom entry
result = self._cmisClient.binding.post(postUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
entryXml.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE)
# what comes back is the XML for the new folder,
# so use it to instantiate a new folder then return it
return AtomPubFolder(self._cmisClient, self._repository, xmlDoc=result)
def createDocumentFromString(self,
name,
properties={},
contentString=None,
contentType=None,
contentEncoding=None):
"""
Creates a new document setting the content to the string provided. If
the repository supports unfiled objects, you do not have to pass in
a parent :class:`Folder` otherwise it is required.
This method is essentially a convenience method that wraps your string
with a StringIO and then calls createDocument.
>>> testFolder.createDocumentFromString('testdoc3', contentString='hello, world', contentType='text/plain')
"""
return self._repository.createDocumentFromString(name, properties,
self, contentString, contentType, contentEncoding)
def createDocument(self, name, properties={}, contentFile=None,
contentType=None, contentEncoding=None):
"""
Creates a new Document object in the repository using
the properties provided.
Right now this is basically the same as createFolder,
but this deals with contentStreams. The common logic should
probably be moved to CmisObject.createObject.
The method will attempt to guess the appropriate content
type and encoding based on the file. To specify it yourself, pass them
in via the contentType and contentEncoding arguments.
>>> f = open('250px-Cmis_logo.png', 'rb')
>>> subFolder.createDocument('logo.png', contentFile=f)
<cmislib.model.Document object at 0x10410fa10>
>>> f.close()
If you wanted to set one or more properties when creating the doc, pass
in a dict, like this:
>>> props = {'cmis:someProp':'someVal'}
>>> f = open('250px-Cmis_logo.png', 'rb')
>>> subFolder.createDocument('logo.png', props, contentFile=f)
<cmislib.model.Document object at 0x10410fa10>
>>> f.close()
To specify a custom object type, pass in a property called
cmis:objectTypeId set to the :class:`CmisId` representing the type ID
of the instance you want to create. If you do not pass in an object
type ID, an instance of 'cmis:document' will be created.
The following optional arguments are not yet supported:
- versioningState
- policies
- addACEs
- removeACEs
"""
return self._repository.createDocument(name,
properties,
self,
contentFile,
contentType,
contentEncoding)
def getChildren(self, **kwargs):
"""
Returns a paged :class:`ResultSet`. The result set contains a list of
:class:`CmisObject` objects for each child of the Folder. The actual
type of the object returned depends on the object's CMIS base type id.
For example, the method might return a list that contains both
:class:`Document` objects and :class:`Folder` objects.
>>> childrenRS = subFolder.getChildren()
>>> children = childrenRS.getResults()
The following optional arguments are supported:
- maxItems
- skipCount
- orderBy
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
- includePathSegment
"""
# get the appropriate 'down' link
childrenUrl = self.getChildrenLink()
# invoke the URL
result = self._cmisClient.binding.get(childrenUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def getChildrenLink(self):
"""
Gets the Atom link that knows how to return this object's children.
"""
url = self._getLink(DOWN_REL, ATOM_XML_FEED_TYPE_P)
assert len(url) > 0, "Could not find the children url"
return url
def getDescendantsLink(self):
"""
Returns the 'down' link of type `CMIS_TREE_TYPE`
>>> folder.getDescendantsLink()
u'http://localhost:8080/alfresco/s/cmis/s/workspace:SpacesStore/i/86f6bf54-f0e8-4a72-8cb1-213599ba086c/descendants'
"""
url = self._getLink(DOWN_REL, CMIS_TREE_TYPE_P)
assert len(url) > 0, "Could not find the descendants url"
# some servers return a depth arg as part of this URL
# so strip it off but keep other args
if url.find("?") >= 0:
u = list(urlparse(url))
u[4] = '&'.join([p for p in u[4].split('&') if not p.startswith('depth=')])
url = urlunparse(u)
return url
def getDescendants(self, **kwargs):
"""
Gets the descendants of this folder. The descendants are returned as
a paged :class:`ResultSet` object. The result set contains a list of
:class:`CmisObject` objects where the actual type of each object
returned will vary depending on the object's base type id. For example,
the method might return a list that contains both :class:`Document`
objects and :class:`Folder` objects.
The following optional argument is supported:
- depth. Use depth=-1 for all descendants, which is the default if no
depth is specified.
>>> resultSet = folder.getDescendants()
>>> len(resultSet.getResults())
105
>>> resultSet = folder.getDescendants(depth=1)
>>> len(resultSet.getResults())
103
The following optional arguments *may* also work but haven't been
tested:
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
- includePathSegment
"""
if not self._repository.getCapabilities()['GetDescendants']:
raise NotSupportedException('This repository does not support getDescendants')
# default the depth to -1, which is all descendants
if "depth" not in kwargs:
kwargs['depth'] = -1
# get the appropriate 'down' link
descendantsUrl = self.getDescendantsLink()
# invoke the URL
result = self._cmisClient.binding.get(descendantsUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def getTree(self, **kwargs):
"""
Unlike :class:`Folder.getChildren` or :class:`Folder.getDescendants`,
this method returns only the descendant objects that are folders. The
results do not include the current folder.
The following optional arguments are supported:
- depth
- filter
- includeRelationships
- renditionFilter
- includeAllowableActions
- includePathSegment
>>> rs = folder.getTree(depth='2')
>>> len(rs.getResults())
3
>>> for folder in rs.getResults().values():
... folder.getTitle()
...
u'subfolder2'
u'parent test folder'
u'subfolder'
"""
# Get the descendants link and do a GET against it
url = self._getLink(FOLDER_TREE_REL)
assert url is not None, 'Unable to determine folder tree link'
result = self._cmisClient.binding.get(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# return the result set
return AtomPubResultSet(self._cmisClient, self._repository, result)
def getParent(self):
"""
This is not yet implemented.
The optional filter argument is not yet supported.
"""
# get the appropriate 'up' link
parentUrl = self._getLink(UP_REL)
# invoke the URL
result = self._cmisClient.binding.get(parentUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password)
# return the result set
return AtomPubFolder(self._cmisClient, self._repository, xmlDoc=result)
def deleteTree(self, **kwargs):
"""
Deletes the folder and all of its descendant objects.
>>> resultSet = subFolder.getDescendants()
>>> len(resultSet.getResults())
2
>>> subFolder.deleteTree()
The following optional arguments are supported:
- allVersions
- unfileObjects
- continueOnFailure
"""
# Per the spec, the repo must have the GetDescendants capability
# to support deleteTree
if not self._repository.getCapabilities()['GetDescendants']:
raise NotSupportedException('This repository does not support deleteTree')
# Get the descendants link and do a DELETE against it
url = self._getLink(DOWN_REL, CMIS_TREE_TYPE_P)
result = self._cmisClient.binding.delete(url.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
def addObject(self, cmisObject, **kwargs):
"""
Adds the specified object as a child of this object. No new object is
created. The repository must support multifiling for this to work.
>>> sub1 = repo.getObjectByPath("/cmislib/sub1")
>>> sub2 = repo.getObjectByPath("/cmislib/sub2")
>>> doc = sub1.createDocument("testdoc1")
>>> len(sub1.getChildren())
1
>>> len(sub2.getChildren())
0
>>> sub2.addObject(doc)
>>> len(sub2.getChildren())
1
>>> sub2.getChildren()[0].name
u'testdoc1'
The following optional arguments are supported:
- allVersions
"""
if not self._repository.getCapabilities()['Multifiling']:
raise NotSupportedException('This repository does not support multifiling')
postUrl = self.getChildrenLink()
# post the Atom entry
self._cmisClient.binding.post(postUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
cmisObject.xmlDoc.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE,
**kwargs)
def removeObject(self, cmisObject):
"""
Removes the specified object from this folder. The repository must
support unfiling for this to work.
"""
if not self._repository.getCapabilities()['Unfiling']:
raise NotSupportedException('This repository does not support unfiling')
postUrl = self._repository.getCollectionLink(UNFILED_COLL)
args = {"removeFrom": self.getObjectId()}
# post the Atom entry to the unfiled collection
self._cmisClient.binding.post(postUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
cmisObject.xmlDoc.toxml(encoding='utf-8'),
ATOM_XML_ENTRY_TYPE,
**args)
def getPaths(self):
"""
Returns the paths as a list of strings. The spec says folders cannot
be multi-filed, so this should always be one value. We return a list
to be symmetric with the same method in :class:`Document`.
"""
return [self.properties['cmis:path']]
class AtomPubRelationship(AtomPubCmisObject):
"""
Defines a relationship object between two :class:`CmisObjects` objects
"""
def getSourceId(self):
"""
Returns the :class:`CmisId` on the source side of the relationship.
"""
if self.xmlDoc is None:
self.reload()
props = self.getProperties()
return AtomPubCmisId(props['cmis:sourceId'])
def getTargetId(self):
"""
Returns the :class:`CmisId` on the target side of the relationship.
"""
if self.xmlDoc is None:
self.reload()
props = self.getProperties()
return AtomPubCmisId(props['cmis:targetId'])
def getSource(self):
"""
Returns an instance of the appropriate child-type of :class:`CmisObject`
for the source side of the relationship.
"""
sourceId = self.getSourceId()
return getSpecializedObject(self._repository.getObject(sourceId))
def getTarget(self):
"""
Returns an instance of the appropriate child-type of :class:`CmisObject`
for the target side of the relationship.
"""
targetId = self.getTargetId()
return getSpecializedObject(self._repository.getObject(targetId))
sourceId = property(getSourceId)
targetId = property(getTargetId)
source = property(getSource)
target = property(getTarget)
class AtomPubPolicy(AtomPubCmisObject):
"""
An arbirary object that can 'applied' to objects that the
repository identifies as being 'controllable'.
"""
pass
class AtomPubObjectType(ObjectType):
"""
Represents the CMIS object type such as 'cmis:document' or 'cmis:folder'.
Contains metadata about the type.
"""
def __init__(self, cmisClient, repository, typeId=None, xmlDoc=None):
""" Constructor """
self._cmisClient = cmisClient
self._repository = repository
self._kwargs = None
self._typeId = typeId
self.xmlDoc = xmlDoc
self.logger = logging.getLogger('cmislib.model.ObjectType')
self.logger.info('Creating an instance of ObjectType')
def __str__(self):
"""To string"""
return self.getTypeId()
def getTypeId(self):
"""
Returns the type ID for this object.
>>> docType = repo.getTypeDefinition('cmis:document')
>>> docType.getTypeId()
'cmis:document'
"""
if self._typeId is None:
if self.xmlDoc is None:
self.reload()
self._typeId = CmisId(self._getElementValue(CMIS_NS, 'id'))
return self._typeId
def _getElementValue(self, namespace, elementName):
"""
Helper method to retrieve child element values from type XML.
"""
if self.xmlDoc is None:
self.reload()
# typeEls = self.xmlDoc.getElementsByTagNameNS(CMISRA_NS, 'type')
# assert len(typeEls) == 1, "Expected to find exactly one type element but instead found %d" % len(typeEls)
# typeEl = typeEls[0]
typeEl = None
for e in self.xmlDoc.childNodes:
if e.nodeType == e.ELEMENT_NODE and e.localName == "type":
typeEl = e
break
assert typeEl, "Expected to find one child element named type"
els = typeEl.getElementsByTagNameNS(namespace, elementName)
if len(els) >= 1:
el = els[0]
if el and len(el.childNodes) >= 1:
return el.childNodes[0].data
def getLocalName(self):
"""Getter for cmis:localName"""
return self._getElementValue(CMIS_NS, 'localName')
def getLocalNamespace(self):
"""Getter for cmis:localNamespace"""
return self._getElementValue(CMIS_NS, 'localNamespace')
def getDisplayName(self):
"""Getter for cmis:displayName"""
return self._getElementValue(CMIS_NS, 'displayName')
def getQueryName(self):
"""Getter for cmis:queryName"""
return self._getElementValue(CMIS_NS, 'queryName')
def getDescription(self):
"""Getter for cmis:description"""
return self._getElementValue(CMIS_NS, 'description')
def getBaseId(self):
"""Getter for cmis:baseId"""
return AtomPubCmisId(self._getElementValue(CMIS_NS, 'baseId'))
def isCreatable(self):
"""Getter for cmis:creatable"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'creatable'))
def isFileable(self):
"""Getter for cmis:fileable"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'fileable'))
def isQueryable(self):
"""Getter for cmis:queryable"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'queryable'))
def isFulltextIndexed(self):
"""Getter for cmis:fulltextIndexed"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'fulltextIndexed'))
def isIncludedInSupertypeQuery(self):
"""Getter for cmis:includedInSupertypeQuery"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'includedInSupertypeQuery'))
def isControllablePolicy(self):
"""Getter for cmis:controllablePolicy"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'controllablePolicy'))
def isControllableACL(self):
"""Getter for cmis:controllableACL"""
return parseBoolValue(self._getElementValue(CMIS_NS, 'controllableACL'))
def getLink(self, rel, linkType):
"""
Gets the HREF for the link element with the specified rel and linkType.
>>> from cmislib.atompub.atompub_binding import ATOM_XML_FEED_TYPE
>>> docType.getLink('down', ATOM_XML_FEED_TYPE)
u'http://localhost:8080/alfresco/s/cmis/type/cmis:document/children'
"""
linkElements = self.xmlDoc.getElementsByTagNameNS(ATOM_NS, 'link')
for linkElement in linkElements:
if linkElement.attributes.has_key('rel') and linkElement.attributes.has_key('type'):
relAttr = linkElement.attributes['rel'].value
typeAttr = linkElement.attributes['type'].value
if relAttr == rel and linkType.match(typeAttr):
return linkElement.attributes['href'].value
def getProperties(self):
"""
Returns a list of :class:`Property` objects representing each property
defined for this type.
>>> objType = repo.getTypeDefinition('cmis:relationship')
>>> for prop in objType.properties:
... print 'Id:%s' % prop.id
... print 'Cardinality:%s' % prop.cardinality
... print 'Description:%s' % prop.description
... print 'Display name:%s' % prop.displayName
... print 'Local name:%s' % prop.localName
... print 'Local namespace:%s' % prop.localNamespace
... print 'Property type:%s' % prop.propertyType
... print 'Query name:%s' % prop.queryName
... print 'Updatability:%s' % prop.updatability
... print 'Inherited:%s' % prop.inherited
... print 'Orderable:%s' % prop.orderable
... print 'Queryable:%s' % prop.queryable
... print 'Required:%s' % prop.required
... print 'Open choice:%s' % prop.openChoice
"""
if self.xmlDoc is None:
self.reload(includePropertyDefinitions='true')
# Currently, property defs don't have an enclosing element. And, the
# element name varies depending on type. Until that changes, I'm going
# to find all elements unique to a prop, then grab its parent node.
propTypeElements = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'propertyType')
if len(propTypeElements) <= 0:
self.reload(includePropertyDefinitions='true')
propTypeElements = self.xmlDoc.getElementsByTagNameNS(CMIS_NS, 'propertyType')
assert len(propTypeElements) > 0, 'Could not retrieve object type property definitions'
props = {}
for typeEl in propTypeElements:
prop = AtomPubProperty(typeEl.parentNode)
props[prop.id] = prop
return props
def reload(self, **kwargs):
"""
This method will reload the object's data from the CMIS service.
"""
if kwargs:
if self._kwargs:
self._kwargs.update(kwargs)
else:
self._kwargs = kwargs
templates = self._repository.getUriTemplates()
template = templates['typebyid']['template']
params = {'{id}': self._typeId}
byTypeIdUrl = multiple_replace(params, template)
result = self._cmisClient.binding.get(byTypeIdUrl.encode('utf-8'),
self._cmisClient.username,
self._cmisClient.password,
**kwargs)
# instantiate CmisObject objects with the results and return the list
entryElements = result.getElementsByTagNameNS(ATOM_NS, 'entry')
assert(len(entryElements) == 1), "Expected entry element in result from calling %s" % byTypeIdUrl
self.xmlDoc = entryElements[0]
id = property(getTypeId)
localName = property(getLocalName)
localNamespace = property(getLocalNamespace)
displayName = property(getDisplayName)
queryName = property(getQueryName)
description = property(getDescription)
baseId = property(getBaseId)
creatable = property(isCreatable)
fileable = property(isFileable)
queryable = property(isQueryable)
fulltextIndexed = property(isFulltextIndexed)
includedInSupertypeQuery = property(isIncludedInSupertypeQuery)
controllablePolicy = property(isControllablePolicy)
controllableACL = property(isControllableACL)
properties = property(getProperties)
class AtomPubProperty(Property):
"""
This class represents an attribute or property definition of an object
type.
"""