zope.wfmc 3.5.0

Workflow-Management Coalition Workflow Engine

This package provides an implementation of a Workflow-Management Coalition (WFMC) workflow engine. The engine is provided as a collection of workflow process components. Workflow processes can be defined in Python or via the XML Process-Definition Language, XPDL.

Detailed Documentation

Workflow-Management Coalition Workflow Engine

This package provides an implementation of a Workflow-Management Coalition (WFMC) workflow engine. The engine is provided as a collection of workflow process components. Workflow processes can be defined in Python or via the XML Process-Definition Language, XPDL.

In this document, we’ll look at Python-defined process definitions:

>>> from zope.wfmc import process
>>> pd = process.ProcessDefinition('sample')

The argument to the process is a process id.

A process has a number of parts. Let’s look at a sample review process:

                            -----------
                         -->| Publish |
----------   ---------- /   -----------
| Author |-->| Review |-    ----------
----------   ---------- \-->| Reject |
                            ----------

Here we have a single start activity and 2 end activities. We could have modeled this with a single end activity, but that is not required. A single start activity is required. A process definition has a set of activities, with transitions between them. Let’s define the activities for our process definition:

>>> pd.defineActivities(
...     author  = process.ActivityDefinition(),
...     review  = process.ActivityDefinition(),
...     publish = process.ActivityDefinition(),
...     reject  = process.ActivityDefinition(),
...     )

We supply activities as keyword arguments. The argument names provide activity ids that we’ll use when defining transitions:

>>> pd.defineTransitions(
...     process.TransitionDefinition('author', 'review'),
...     process.TransitionDefinition('review', 'publish'),
...     process.TransitionDefinition('review', 'reject'),
...     )

Each transition is constructed with an identifier for a starting activity, and an identifier for an ending activity.

Before we can use a workflow definition, we have to register it as a utility. This is necessary so that process instances can find their definitions. In addition, the utility name must match the process id:

>>> import zope.component
>>> zope.component.provideUtility(pd, name=pd.id)

Now, with this definition, we can execute our workflow. We haven’t defined any work yet, but we can see the workflow execute. We’ll see the workflow executing by registering a subscriber that logs workflow events:

>>> def log_workflow(event):
...     print event

>>> import zope.event
>>> zope.event.subscribers.append(log_workflow)

To use the workflow definition, we need to create an instance:

>>> proc = pd()

Now, if we start the workflow:

>>> proc.start()
ProcessStarted(Process('sample'))
Transition(None, Activity('sample.author'))
ActivityStarted(Activity('sample.author'))
ActivityFinished(Activity('sample.author'))
Transition(Activity('sample.author'), Activity('sample.review'))
ActivityStarted(Activity('sample.review'))
ActivityFinished(Activity('sample.review'))
Transition(Activity('sample.review'), Activity('sample.publish'))
ActivityStarted(Activity('sample.publish'))
ActivityFinished(Activity('sample.publish'))
ProcessFinished(Process('sample'))

we see that we transition immediately into the author activity, then into review and publish. Normally, we’d need to do some work in each activity, and transitions would continue only after work had been done, however, in this case, we didn’t define any work, so each activity completed immediately.

Note that we didn’t transition into the rejected activity. By default, when an activity is completed, the first transition for which its condition evaluates to True is used. By default, transitions have boolean conditions [1] that evaluate to True, so the transition to publish is used because it was defined before the transition to reject. What we want is to transition to publish if a reviewer approves the content for publication, but to reject if the reviewer rejects the content for publication. We can use a condition for this:

>>> pd = process.ProcessDefinition('sample')
>>> zope.component.provideUtility(pd, name=pd.id)
Unregistered event:
UtilityRegistration(<BaseGlobalComponents base>, IProcessDefinition,
                    'sample', ProcessDefinition('sample'), None, u'')

>>> pd.defineActivities(
...     author = process.ActivityDefinition(),
...     review = process.ActivityDefinition(),
...     publish = process.ActivityDefinition(),
...     reject = process.ActivityDefinition(),
...     )
>>> pd.defineTransitions(
...     process.TransitionDefinition('author', 'review'),
...     process.TransitionDefinition(
...         'review', 'publish', condition=lambda data: data.publish),
...     process.TransitionDefinition('review', 'reject'),
...     )

We redefined the workflow process, specifying a condition for the transition to publish. Boolean conditions are just callable objects that take a data object and return a boolean value. The data object is called “workflow-relevant data”. A process instance has a data object containing this data. In the example, above, the condition simply returned the value of the publish attribute. How does this attribute get set? It needs to be set by the review activity. To do that, we need to arrange for the activity to set the data. This brings us to applications.

Process definitions are meant to be used with different applications. For this reason, process definitions don’t include application logic. What they do include is a specifications of the applications to be invoked and the flow of work-flow-relevant data to and from the application. Now, we can define our applications:

>>> pd.defineApplications(
...     author = process.Application(),
...     review = process.Application(
...         process.OutputParameter('publish')),
...     publish = process.Application(),
...     reject = process.Application(),
...     )

We used the same names for the applications that we used for our activities. This isn’t required, but is a common practice. Note that the review application includes a specification of an output parameter. Now that we’ve defined our applications, we need to modify our activities to use them:

>>> pd.activities['author'].addApplication('author')
>>> pd.activities['review'].addApplication('review', ['publish'])
>>> pd.activities['publish'].addApplication('publish')
>>> pd.activities['reject'].addApplication('reject')

An activity can use many applications, so we call addApplication. In the application definition for the ‘review’ application, we provided the name of a workflow-relevent data variable corresponding to the output parameter defined for the application. When using an application in an activity, a workflow-relevent data variable name must be provided for each of the parameters in the identified applications’s signature. When an application is used in an activity, workflow-relevent data are passed for each of the input parameters and are set by each of the output parameters. In this example, the output parameter, will be used to add a publish attribute to the workflow relevant data.

Participants

We’ve declared some applications, and we’ve wired them up to activities, but we still haven’t specified any application code. Before we can specify application code, we need to consider who will be performing the application. Workflow applications are normally executed by people, or other external actors. As with applications, process definitions allow participants in the workflow to be declared and identified with activities. We declare participants much as we declare applications, except without parameters:

>>> pd.defineParticipants(
...     author   = process.Participant(),
...     reviewer = process.Participant(),
...     )

In this case, we happened to reuse an activity name for one, but not both of the participants. Having defined these participants, we can associate them with activities:

>>> pd.activities['author'].definePerformer('author')
>>> pd.activities['review'].definePerformer('reviewer')

Application Integration

To use a process definition to control application logic, we need to associate it with an “integration” object.

When a process needs to get a participant, it calls createParticipant on its integration attribute, passing the process id and the performer id. If an activity doesn’t have a performer, then the procedure above is used with an empty performer id.

Similarly, when a process needs a work item, it calls createWorkItem on its integration attribute, passing the process id and the application id.

Work items provide a start method, which is used to start the work and pass input arguments. It is the responsibility of the work item, at some later time, to call the workItemFinished method on the activity, to notify the activity that the work item was completed. Output parameters are passed to the workItemFinished method.

A simple way to create integration objects is with zope.wfmc.attributeintegration.AttributeIntegration.

>>> from zope.wfmc.attributeintegration import AttributeIntegration
>>> integration = AttributeIntegration()
>>> pd.integration = integration

We’ll start by defining a simple Participant class:

>>> import zope.interface
>>> from zope.wfmc import interfaces

>>> class Participant(object):
...     zope.component.adapts(interfaces.IActivity)
...     zope.interface.implements(interfaces.IParticipant)
...
...     def __init__(self, activity):
...         self.activity = activity

We set attributes on the integration for each participant:

>>> integration.authorParticipant   = Participant
>>> integration.reviewerParticipant = Participant

We also define an attribute for participants for activities that don’t have performers:

>>> integration.Participant = Participant

Now we’ll define our work-items. First we’ll define some classes:

>>> work_list = []

>>> class ApplicationBase:
...     zope.component.adapts(interfaces.IParticipant)
...     zope.interface.implements(interfaces.IWorkItem)
...
...     def __init__(self, participant):
...         self.participant = participant
...         work_list.append(self)
...
...     def start(self):
...         pass
...
...     def finish(self):
...         self.participant.activity.workItemFinished(self)

>>> class Review(ApplicationBase):
...     def finish(self, publish):
...         self.participant.activity.workItemFinished(self, publish)

>>> class Publish(ApplicationBase):
...     def start(self):
...         print "Published"
...         self.finish()

>>> class Reject(ApplicationBase):
...     def start(self):
...         print "Rejected"
...         self.finish()

and then we’ll hook them up with the integration object:

>>> integration.authorWorkItem  = ApplicationBase
>>> integration.reviewWorkItem  = Review
>>> integration.publishWorkItem = Publish
>>> integration.rejectWorkItem  = Reject

Using workflow processes

To use a process definition, instantiate it and call its start method to start execution:

>>> proc = pd()
>>> proc.start()
... # doctest: +NORMALIZE_WHITESPACE
ProcessStarted(Process('sample'))
Transition(None, Activity('sample.author'))
ActivityStarted(Activity('sample.author'))

We transition into the author activity and wait for work to get done. To move forward, we need to get at the authoring work item, so we can finish it. Our work items add themselves to a work list, so we can get the item from the list.

>>> item = work_list.pop()

Now we can finish the work item, by calling its finish method:

>>> item.finish()
WorkItemFinished('author')
ActivityFinished(Activity('sample.author'))
Transition(Activity('sample.author'), Activity('sample.review'))
ActivityStarted(Activity('sample.review'))

We see that we transitioned to the review activity. Note that the finish method isn’t a part of the workflow APIs. It was defined by our sample classes. Other applications could use different mechanisms.

Now, we’ll finish the review process by calling the review work item’s finish. We’ll pass False, indicating that the content should not be published:

>>> work_list.pop().finish(False)
WorkItemFinished('review')
ActivityFinished(Activity('sample.review'))
Transition(Activity('sample.review'), Activity('sample.reject'))
ActivityStarted(Activity('sample.reject'))
Rejected
WorkItemFinished('reject')
ActivityFinished(Activity('sample.reject'))
ProcessFinished(Process('sample'))

Ordering output transitions

Normally, outgoing transitions are ordered in the order of transition definition and all transitions from a given activity are used.

If transitions are defined in an inconvenient order, then the workflow might not work as expected. For example, let’s modify the above process by switching the order of definition of some of the transitions. We’ll reuse our integration object from the previous example by passing it to the definition constructor:

>>> pd = process.ProcessDefinition('sample', integration)
>>> zope.component.provideUtility(pd, name=pd.id)
Unregistered event:
UtilityRegistration(<BaseGlobalComponents base>, IProcessDefinition,
                    'sample', ProcessDefinition('sample'), None, u'')

>>> pd.defineActivities(
...     author = process.ActivityDefinition(),
...     review = process.ActivityDefinition(),
...     publish = process.ActivityDefinition(),
...     reject = process.ActivityDefinition(),
...     )
>>> pd.defineTransitions(
...     process.TransitionDefinition('author', 'review'),
...     process.TransitionDefinition('review', 'reject'),
...     process.TransitionDefinition(
...         'review', 'publish', condition=lambda data: data.publish),
...     )

>>> pd.defineApplications(
...     author = process.Application(),
...     review = process.Application(
...         process.OutputParameter('publish')),
...     publish = process.Application(),
...     reject = process.Application(),
...     )

>>> pd.activities['author'].addApplication('author')
>>> pd.activities['review'].addApplication('review', ['publish'])
>>> pd.activities['publish'].addApplication('publish')
>>> pd.activities['reject'].addApplication('reject')

>>> pd.defineParticipants(
...     author   = process.Participant(),
...     reviewer = process.Participant(),
...     )

>>> pd.activities['author'].definePerformer('author')
>>> pd.activities['review'].definePerformer('reviewer')

and run our process:

>>> proc = pd()
>>> proc.start()
... # doctest: +NORMALIZE_WHITESPACE
ProcessStarted(Process('sample'))
Transition(None, Activity('sample.author'))
ActivityStarted(Activity('sample.author'))

>>> work_list.pop().finish()
WorkItemFinished('author')
ActivityFinished(Activity('sample.author'))
Transition(Activity('sample.author'), Activity('sample.review'))
ActivityStarted(Activity('sample.review'))

This time, we’ll say that we should publish:

>>> work_list.pop().finish(True)
WorkItemFinished('review')
ActivityFinished(Activity('sample.review'))
Transition(Activity('sample.review'), Activity('sample.reject'))
ActivityStarted(Activity('sample.reject'))
Rejected
WorkItemFinished('reject')
ActivityFinished(Activity('sample.reject'))
ProcessFinished(Process('sample'))

But we went to the reject activity anyway. Why? Because transitions are tested in order. Because the transition to the reject activity was tested first and had no condition, we followed it without checking the condition for the transition to the publish activity. We can fix this by specifying outgoing transitions on the reviewer activity directly. To do this, we’ll also need to specify ids in our transitions. Let’s redefine the process:

>>> pd = process.ProcessDefinition('sample', integration)
>>> zope.component.provideUtility(pd, name=pd.id)
Unregistered event:
UtilityRegistration(<BaseGlobalComponents base>, IProcessDefinition,
                    'sample', ProcessDefinition('sample'), None, u'')

>>> pd.defineActivities(
...     author = process.ActivityDefinition(),
...     review = process.ActivityDefinition(),
...     publish = process.ActivityDefinition(),
...     reject = process.ActivityDefinition(),
...     )
>>> pd.defineTransitions(
...     process.TransitionDefinition('author', 'review'),
...     process.TransitionDefinition('review', 'reject', id='reject'),
...     process.TransitionDefinition(
...         'review', 'publish', id='publish',
...         condition=lambda data: data.publish),
...     )

>>> pd.defineApplications(
...     author = process.Application(),
...     review = process.Application(
...         process.OutputParameter('publish')),
...     publish = process.Application(),
...     reject = process.Application(),
...     )

>>> pd.activities['author'].addApplication('author')
>>> pd.activities['review'].addApplication('review', ['publish'])
>>> pd.activities['publish'].addApplication('publish')
>>> pd.activities['reject'].addApplication('reject')

>>> pd.defineParticipants(
...     author   = process.Participant(),
...     reviewer = process.Participant(),
...     )

>>> pd.activities['author'].definePerformer('author')
>>> pd.activities['review'].definePerformer('reviewer')

>>> pd.activities['review'].addOutgoing('publish')
>>> pd.activities['review'].addOutgoing('reject')

Now, when we run the process, we’ll go to the publish activity as expected:

>>> proc = pd()
>>> proc.start()
... # doctest: +NORMALIZE_WHITESPACE
ProcessStarted(Process('sample'))
Transition(None, Activity('sample.author'))
ActivityStarted(Activity('sample.author'))

>>> work_list.pop().finish()
WorkItemFinished('author')
ActivityFinished(Activity('sample.author'))
Transition(Activity('sample.author'), Activity('sample.review'))
ActivityStarted(Activity('sample.review'))

>>> work_list.pop().finish(True)
WorkItemFinished('review')
ActivityFinished(Activity('sample.review'))
Transition(Activity('sample.review'), Activity('sample.publish'))
ActivityStarted(Activity('sample.publish'))
Published
WorkItemFinished('publish')
ActivityFinished(Activity('sample.publish'))
ProcessFinished(Process('sample'))

Let’s see the other way also, where we should transition to reject:

>>> proc = pd()
>>> proc.start()
... # doctest: +NORMALIZE_WHITESPACE
ProcessStarted(Process('sample'))
Transition(None, Activity('sample.author'))
ActivityStarted(Activity('sample.author'))

>>> work_list.pop().finish()
WorkItemFinished('author')
ActivityFinished(Activity('sample.author'))
Transition(Activity('sample.author'), Activity('sample.review'))
ActivityStarted(Activity('sample.review'))

>>> work_list.pop().finish(False)
WorkItemFinished('review')
ActivityFinished(Activity('sample.review'))
Transition(Activity('sample.review'), Activity('sample.reject'))
ActivityStarted(Activity('sample.reject'))
Rejected
WorkItemFinished('reject')
ActivityFinished(Activity('sample.reject'))
ProcessFinished(Process('sample'))

Complex Flows

Lets look at a more complex example. In this example, we’ll extend the process to work with multiple reviewers. We’ll also make the work-list handling a bit more sophisticated. We’ll also introduce some new concepts:

• splits and joins

• process arguments

Consider the publication process shown below:

Author:      Tech          Tech          Editorial
             Reviewer 1:   Reviewer 2:   Reviewer:
===========  ===========   ===========   ==============
                                                         ---------
     ----------------------------------------------------| Start |
    /                                                    ---------
    |
    V
-----------
| Prepare |<------------------------------\
-----------                                \
    |        ------------                   \
    |        | Tech     |--------------- \   \
    |------->| Review 1 |                 V   |
    |        ------------  ----------    -------------
     \                     | Tech   |    | Editorial |   ----------
       ------------------->| Review |--->| Review    |-->| Reject |
                           | 2      |    -------------   ----------
                           ----------      |      |
-----------                               /        \
| Prepare |                              /          \--------\
| Final   |<----------------------------/                    |
-----------                                                  |
   ^   |                                 ----------          V
   |    \------------------------------->| Review |      -----------
    \                                    | Final  |----->| Publish |
     ------------------------------------|        |      -----------
                                         ----------

Here we’ve arranged the process diagram into columns, with the activities for each participant. We have four participants, the author, two technical reviewers, and an editorial reviewer. The author prepares a draft. The author sends the draft to both technical reviewers for review. When the technical reviews have completed, the editorial review does an initial editorial review. Based on the technical reviews, the editor may choose to:

• Reject the document

• Publish the document as is

• Request technical changes (based on the technical reviewers’ comments), or

• Request editorial changes.

If technical changes are required, the work flows back to the “Prepare” activity. If editorial changes are necessary, then work flows to the “Prepare Final” activity. When the author has made the editorial changes, work flows to “Review Final”. The editor may request additional changes, in which case, work flows back to “Prepare Final”, otherwise, the work flows to “Publish”.

This example illustrates different kinds of “joins” and “splits”. The term “join” refers to the way incoming transitions to an activity are handled. There are two kinds of joins: “and” and “xor”. With an “and” join, the activity waits for each of the incoming transitions. In this example, the inputs to the “Editorial Review” activity form an “and” join. Editorial review waits until each of the technical reviews are completed. The rest of the joins in this example are “xor” joins. The activity starts on any transition into the activity.

The term “split” refers to way outgoing transitions from an activity are handled. Normally, exactly one transition out of an activity is used. This is called an “xor” split. With an “and” split, all transitions with boolean conditions that evaluate to True are used. In this example, the “Prepare” activity has an “and” split. Work flows simultaneously to the two technical review activities. The rest of the splits in this example are “xor” splits.

Lets create our new workflow process. We’ll reuse our existing integration object:

>>> Publication = process.ProcessDefinition('Publication')
>>> Publication.integration = integration
>>> zope.component.provideUtility(Publication, name=Publication.id)

>>> Publication.defineActivities(
...     start   = process.ActivityDefinition("Start"),
...     prepare = process.ActivityDefinition("Prepare"),
...     tech1   = process.ActivityDefinition("Technical Review 1"),
...     tech2   = process.ActivityDefinition("Technical Review 2"),
...     review  = process.ActivityDefinition("Editorial Review"),
...     final   = process.ActivityDefinition("Final Preparation"),
...     rfinal  = process.ActivityDefinition("Review Final"),
...     publish = process.ActivityDefinition("Publish"),
...     reject  = process.ActivityDefinition("Reject"),
...     )

Here, we’ve passed strings to the activity definitions providing names. Names must be either unicode or ASCII strings.

We define our transitions:

>>> Publication.defineTransitions(
...     process.TransitionDefinition('start', 'prepare'),
...     process.TransitionDefinition('prepare', 'tech1'),
...     process.TransitionDefinition('prepare', 'tech2'),
...     process.TransitionDefinition('tech1', 'review'),
...     process.TransitionDefinition('tech2', 'review'),
...
...     process.TransitionDefinition(
...         'review', 'reject',
...         condition=lambda data: not data.publish
...         ),
...     process.TransitionDefinition(
...         'review', 'prepare',
...         condition=lambda data: data.tech_changes
...         ),
...     process.TransitionDefinition(
...         'review', 'final',
...         condition=lambda data: data.ed_changes
...         ),
...     process.TransitionDefinition('review', 'publish'),
...
...     process.TransitionDefinition('final', 'rfinal'),
...     process.TransitionDefinition(
...         'rfinal', 'final',
...         condition=lambda data: data.ed_changes
...         ),
...     process.TransitionDefinition('rfinal', 'publish'),
...     )

We specify our “and” split and join:

>>> Publication.activities['prepare'].andSplit(True)
>>> Publication.activities['review'].andJoin(True)

We define our participants and applications:

>>> Publication.defineParticipants(
...     author   = process.Participant("Author"),
...     tech1    = process.Participant("Technical Reviewer 1"),
...     tech2    = process.Participant("Technical Reviewer 2"),
...     reviewer = process.Participant("Editorial Reviewer"),
...     )

>>> Publication.defineApplications(
...     prepare = process.Application(),
...     tech_review = process.Application(
...         process.OutputParameter('publish'),
...         process.OutputParameter('tech_changes'),
...         ),
...     ed_review = process.Application(
...         process.InputParameter('publish1'),
...         process.InputParameter('tech_changes1'),
...         process.InputParameter('publish2'),
...         process.InputParameter('tech_changes2'),
...         process.OutputParameter('publish'),
...         process.OutputParameter('tech_changes'),
...         process.OutputParameter('ed_changes'),
...         ),
...     publish = process.Application(),
...     reject = process.Application(),
...     final = process.Application(),
...     rfinal = process.Application(
...         process.OutputParameter('ed_changes'),
...         ),
...     )

>>> Publication.activities['prepare'].definePerformer('author')
>>> Publication.activities['prepare'].addApplication('prepare')

>>> Publication.activities['tech1'].definePerformer('tech1')
>>> Publication.activities['tech1'].addApplication(
...     'tech_review', ['publish1', 'tech_changes1'])

>>> Publication.activities['tech2'].definePerformer('tech2')
>>> Publication.activities['tech2'].addApplication(
...     'tech_review', ['publish2', 'tech_changes2'])

>>> Publication.activities['review'].definePerformer('reviewer')
>>> Publication.activities['review'].addApplication(
...     'ed_review',
...     ['publish1', 'tech_changes1', 'publish2', 'tech_changes2',
...      'publish', 'tech_changes', 'ed_changes'],
...     )

>>> Publication.activities['final'].definePerformer('author')
>>> Publication.activities['final'].addApplication('final')

>>> Publication.activities['rfinal'].definePerformer('reviewer')
>>> Publication.activities['rfinal'].addApplication(
...     'rfinal', ['ed_changes'],
...     )

>>> Publication.activities['publish'].addApplication('publish')
>>> Publication.activities['reject'].addApplication('reject')

We want to be able to specify an author when we start the process. We’d also like to be told the final disposition of the process. To accomplish this, we’ll define parameters for our process:

>>> Publication.defineParameters(
...     process.InputParameter('author'),
...     process.OutputParameter('publish'),
...     )

Now that we’ve defined the process, we need to provide participant and application components. Let’s start with our participants. Rather than sharing a single work list, we’ll give each user their own work list. We’ll also create preexisting participants and return them. Finally, we’ll create multiple authors and use the selected one:

>>> class User:
...     def __init__(self):
...         self.work_list = []

>>> authors = {'bob': User(), 'ted': User(), 'sally': User()}

>>> reviewer = User()
>>> tech1 = User()
>>> tech2 = User()

>>> class Author(Participant):
...     def __init__(self, activity):
...         Participant.__init__(self, activity)
...         author_name = activity.process.workflowRelevantData.author
...         print "Author `%s` selected" % author_name
...         self.user = authors[author_name]

In this example, we need to define a separate attribute for each participant:

>>> integration.authorParticipant = Author

When the process is created, the author name will be passed in and assigned to the workflow-relevant data. Our author class uses this information to select the named user.

>>> class Reviewer(Participant):
...     user = reviewer
>>> integration.reviewerParticipant = Reviewer

>>> class Tech1(Participant):
...     user = tech1
>>> integration.tech1Participant = Tech1

>>> class Tech2(Participant):
...     user = tech2
>>> integration.tech2Participant = Tech2

We’ll use our orginal participation class for activities without performers:

>>> integration.Participant = Participant

Now we’ll create our applications. Let’s start with our author:

>>> class ApplicationBase(object):
...     zope.component.adapts(interfaces.IParticipant)
...     zope.interface.implements(interfaces.IWorkItem)
...
...     def __init__(self, participant):
...         self.participant = participant
...         self.activity = participant.activity
...         participant.user.work_list.append(self)
...
...     def start(self):
...         pass
...
...     def finish(self):
...         self.participant.activity.workItemFinished(self)

>>> class Prepare(ApplicationBase):
...
...     def summary(self):
...         process = self.activity.process
...         doc = getattr(process.applicationRelevantData, 'doc', '')
...         if doc:
...             print 'Previous draft:'
...             print doc
...             print 'Changes we need to make:'
...             for change in process.workflowRelevantData.tech_changes:
...                 print change
...         else:
...             print 'Please write the initial draft'
...
...     def finish(self, doc):
...         self.activity.process.applicationRelevantData.doc = doc
...         super(Prepare, self).finish()

>>> integration.prepareWorkItem = Prepare

Since we used the prepare application for revisions as well as initial preparation, we provide a summary method to show us what we have to do.

Here we get the document created by the author passed in as an argument to the finish method. In a more realistic implementation, the author task would create the document at the start of the task and provide a user interface for the user to edit it. We store the document as application-relevant data, since we’ll want reviewers to be able to access it, but we don’t need it directly for workflow control.

>>> class TechReview(ApplicationBase):
...
...     def getDoc(self):
...         return self.activity.process.applicationRelevantData.doc
...
...     def finish(self, decision, changes):
...         self.activity.workItemFinished(self, decision, changes)

>>> integration.tech_reviewWorkItem = TechReview

Here, we provided a method to access the original document.

>>> class Review(TechReview):
...
...     def start(self, publish1, changes1, publish2, changes2):
...         if not (publish1 and publish2):
...             # Reject if either tech reviewer rejects
...             self.activity.workItemFinished(
...                 self, False, changes1 + changes2, ())
...
...         if changes1 or changes2:
...             # we won't do anything if there are tech changes
...             self.activity.workItemFinished(
...                 self, True, changes1 + changes2, ())
...
...     def finish(self, ed_changes):
...         self.activity.workItemFinished(self, True, (), ed_changes)

>>> integration.ed_reviewWorkItem = Review

In this implementation, we decided to reject outright if either technical editor recommended rejection and to send work back to preparation if there are any technical changes. We also subclassed TechReview to get the getDoc method.

We’ll reuse the publish and reject application from the previous example.

>>> class Final(ApplicationBase):
...
...     def summary(self):
...         process = self.activity.process
...         doc = getattr(process.applicationRelevantData, 'doc', '')
...         print 'Previous draft:'
...         print self.activity.process.applicationRelevantData.doc
...         print 'Changes we need to make:'
...         for change in process.workflowRelevantData.ed_changes:
...            print change
...
...     def finish(self, doc):
...         self.activity.process.applicationRelevantData.doc = doc
...         super(Final, self).finish()

>>> integration.finalWorkItem = Final

In our this application, we simply update the document to reflect changes.

>>> class ReviewFinal(TechReview):
...
...     def finish(self, ed_changes):
...         self.activity.workItemFinished(self, ed_changes)

>>> integration.rfinalWorkItem = ReviewFinal

Our process now returns data. When we create a process, we need to supply an object that it can call back to:

>>> class PublicationContext:
...     zope.interface.implements(interfaces.IProcessContext)
...
...     def processFinished(self, process, decision):
...         self.decision = decision

Now, let’s try out our process:

>>> context = PublicationContext()
>>> proc = Publication(context)
>>> proc.start('bob')
ProcessStarted(Process('Publication'))
Transition(None, Activity('Publication.start'))
ActivityStarted(Activity('Publication.start'))
ActivityFinished(Activity('Publication.start'))
Author `bob` selected
Transition(Activity('Publication.start'), Activity('Publication.prepare'))
ActivityStarted(Activity('Publication.prepare'))

We should have added an item to bob’s work list. Let’s get it and finish it, submitting a document:

>>> item = authors['bob'].work_list.pop()
>>> item.finish("I give my pledge, as an American\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my Country.")
WorkItemFinished('prepare')
ActivityFinished(Activity('Publication.prepare'))
Transition(Activity('Publication.prepare'), Activity('Publication.tech1'))
ActivityStarted(Activity('Publication.tech1'))
Transition(Activity('Publication.prepare'), Activity('Publication.tech2'))
ActivityStarted(Activity('Publication.tech2'))

Notice that we transitioned to two activities, tech1 and tech2. This is because the prepare activity has an “and” split. Now we’ll do a tech review. Let’s see what tech1 has:

>>> item = tech1.work_list.pop()
>>> print item.getDoc()
I give my pledge, as an American
to save, and faithfully to defend from waste
the natural resources of my Country.

Let’s tell the author to change “American” to “Earthling”:

>>> item.finish(True, ['Change "American" to "Earthling"'])
WorkItemFinished('tech_review')
ActivityFinished(Activity('Publication.tech1'))
Transition(Activity('Publication.tech1'), Activity('Publication.review'))

Here we transitioned to the editorial review activity, but we didn’t start it. This is because the editorial review activity has an “and” join, meaning that it won’t start until both transitions have occurred.

Now we’ll do the other technical review:

>>> item = tech2.work_list.pop()
>>> item.finish(True, ['Change "Country" to "planet"'])
WorkItemFinished('tech_review')
ActivityFinished(Activity('Publication.tech2'))
Transition(Activity('Publication.tech2'), Activity('Publication.review'))
ActivityStarted(Activity('Publication.review'))
WorkItemFinished('ed_review')
ActivityFinished(Activity('Publication.review'))
Author `bob` selected
Transition(Activity('Publication.review'), Activity('Publication.prepare'))
ActivityStarted(Activity('Publication.prepare'))

Now when we transitioned to the editorial review activity, we started it, because each of the input transitions had happened. Our editorial review application automatically sent the work back to preparation, because there were technical comments. Of course the author is still bob. Let’s address the comments:

>>> item = authors['bob'].work_list.pop()
>>> item.summary()
Previous draft:
I give my pledge, as an American
to save, and faithfully to defend from waste
the natural resources of my Country.
Changes we need to make:
Change "American" to "Earthling"
Change "Country" to "planet"

>>> item.finish("I give my pledge, as an Earthling\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my planet.")
WorkItemFinished('prepare')
ActivityFinished(Activity('Publication.prepare'))
Transition(Activity('Publication.prepare'), Activity('Publication.tech1'))
ActivityStarted(Activity('Publication.tech1'))
Transition(Activity('Publication.prepare'), Activity('Publication.tech2'))
ActivityStarted(Activity('Publication.tech2'))

As before, after completing the initial edits, we start the technical review activities again. We’ll review it again. This time, we have no comments, because the author applied our requested changes:

>>> item = tech1.work_list.pop()
>>> item.finish(True, [])
WorkItemFinished('tech_review')
ActivityFinished(Activity('Publication.tech1'))
Transition(Activity('Publication.tech1'), Activity('Publication.review'))

>>> item = tech2.work_list.pop()
>>> item.finish(True, [])
WorkItemFinished('tech_review')
ActivityFinished(Activity('Publication.tech2'))
Transition(Activity('Publication.tech2'), Activity('Publication.review'))
ActivityStarted(Activity('Publication.review'))

This time, we are left in the technical review activity because there weren’t any technical changes. We’re ready to do our editorial review. We’ll request an editorial change:

>>> item = reviewer.work_list.pop()
>>> print item.getDoc()
I give my pledge, as an Earthling
to save, and faithfully to defend from waste
the natural resources of my planet.

>>> item.finish(['change "an" to "a"'])
WorkItemFinished('ed_review')
ActivityFinished(Activity('Publication.review'))
Author `bob` selected
Transition(Activity('Publication.review'), Activity('Publication.final'))
ActivityStarted(Activity('Publication.final'))

Because we requested editorial changes, we transitioned to the final editing activity, so that the author (still bob) can make the changes:

>>> item = authors['bob'].work_list.pop()
>>> item.summary()
Previous draft:
I give my pledge, as an Earthling
to save, and faithfully to defend from waste
the natural resources of my planet.
Changes we need to make:
change "an" to "a"

>>> item.finish("I give my pledge, as a Earthling\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my planet.")
WorkItemFinished('final')
ActivityFinished(Activity('Publication.final'))
Transition(Activity('Publication.final'), Activity('Publication.rfinal'))
ActivityStarted(Activity('Publication.rfinal'))

We transition to the activity for reviewing the final edits. We review the document and approve it for publication:

>>> item = reviewer.work_list.pop()
>>> print item.getDoc()
I give my pledge, as a Earthling
to save, and faithfully to defend from waste
the natural resources of my planet.

>>> item.finish([])
WorkItemFinished('rfinal')
ActivityFinished(Activity('Publication.rfinal'))
Transition(Activity('Publication.rfinal'), Activity('Publication.publish'))
ActivityStarted(Activity('Publication.publish'))
Published
WorkItemFinished('publish')
ActivityFinished(Activity('Publication.publish'))
ProcessFinished(Process('Publication'))

At this point, the rest of the process finished automatically. In addition, the decision was recorded in the process context object:

>>> context.decision
True

Coming Soon

• XPDL support

• Timeouts/exceptions

• “otherwise” conditions

[1]

There are other kinds of conditions, namely “otherwise” and “exception” conditions.

See also

http://www.wfmc.org http://www.wfmc.org/standards/standards.htm

XPDL Import

We can import process definitions from files in the XML Process Definition Language (XPDL) format. An XPDL file contains multiple process definitions arranged in a package. When we load the file, we get a package containing some number of process definitions.

Let’s look at an example. The file publication.xpdl contains a definition for the publication example developed in the “README.txt” file. We can read it using the xpdl module:

>>> from zope.wfmc import xpdl
>>> import os
>>> package = xpdl.read(open(os.path.join(this_directory,
...                                       'publication.xpdl')))

This package contains a single definition:

>>> package
{u'Publication': ProcessDefinition(u'Publication')}

>>> pd = package[u'Publication']
>>> from zope.wfmc.attributeintegration import AttributeIntegration
>>> integration = AttributeIntegration()
>>> pd.integration = integration

Now, having read the process definition, we can use it as we did before (in “README.txt”). As before, we’ll create an event subscriber so that we can see what’s going on:

>>> def log_workflow(event):
...     print event

>>> import zope.event
>>> zope.event.subscribers.append(log_workflow)

and we’ll register the process definition as a utility:

>>> import zope.component
>>> zope.component.provideUtility(pd, name=pd.id)

and we’ll define and register participant and application adapters:

>>> import zope.interface
>>> from zope.wfmc import interfaces

>>> class Participant(object):
...     zope.component.adapts(interfaces.IActivity)
...     zope.interface.implements(interfaces.IParticipant)
...
...     def __init__(self, activity):
...         self.activity = activity

>>> class User:
...     def __init__(self):
...         self.work_list = []

>>> authors = {'bob': User(), 'ted': User(), 'sally': User()}

>>> reviewer = User()
>>> tech1 = User()
>>> tech2 = User()

>>> class Author(Participant):
...     def __init__(self, activity):
...         Participant.__init__(self, activity)
...         author_name = activity.process.workflowRelevantData.author
...         print "Author `%s` selected" % author_name
...         self.user = authors[author_name]

>>> integration.authorParticipant = Author

>>> class Reviewer(Participant):
...     user = reviewer
>>> integration.reviewerParticipant = Reviewer

>>> class Tech1(Participant):
...     user = tech1
>>> integration.tech1Participant = Tech1

>>> class Tech2(Participant):
...     user = tech2
>>> integration.tech2Participant = Tech2

>>> integration.SystemParticipant = Participant

>>> class ApplicationBase(object):
...     zope.component.adapts(interfaces.IParticipant)
...     zope.interface.implements(interfaces.IWorkItem)
...
...     def __init__(self, participant):
...         self.participant = participant
...         self.activity = participant.activity
...         participant.user.work_list.append(self)
...
...     def start(self):
...         pass
...
...     def finish(self):
...         self.participant.activity.workItemFinished(self)

>>> class Prepare(ApplicationBase):
...
...     def summary(self):
...         process = self.activity.process
...         doc = getattr(process.applicationRelevantData, 'doc', '')
...         if doc:
...             print 'Previous draft:'
...             print doc
...             print 'Changes we need to make:'
...             for change in process.workflowRelevantData.tech_changes:
...                 print change
...         else:
...             print 'Please write the initial draft'
...
...     def finish(self, doc):
...         self.activity.process.applicationRelevantData.doc = doc
...         super(Prepare, self).finish()

>>> integration.prepareWorkItem = Prepare

>>> class TechReview(ApplicationBase):
...
...     def getDoc(self):
...         return self.activity.process.applicationRelevantData.doc
...
...     def finish(self, decision, changes):
...         self.activity.workItemFinished(self, decision, changes)

>>> integration.tech_reviewWorkItem = TechReview

>>> class Review(TechReview):
...
...     def start(self, publish1, changes1, publish2, changes2):
...         if not (publish1 and publish2):
...             # Reject if either tech reviewer rejects
...             self.activity.workItemFinished(
...                 self, False, changes1 + changes2, ())
...
...         if changes1 or changes2:
...             # we won't do anyting if there are tech changes
...             self.activity.workItemFinished(
...                 self, True, changes1 + changes2, ())
...
...     def finish(self, ed_changes):
...         self.activity.workItemFinished(self, True, (), ed_changes)

>>> integration.ed_reviewWorkItem = Review

>>> class Final(ApplicationBase):
...
...     def summary(self):
...         process = self.activity.process
...         doc = getattr(process.applicationRelevantData, 'doc', '')
...         print 'Previous draft:'
...         print self.activity.process.applicationRelevantData.doc
...         print 'Changes we need to make:'
...         for change in process.workflowRelevantData.ed_changes:
...            print change
...
...     def finish(self, doc):
...         self.activity.process.applicationRelevantData.doc = doc
...         super(Final, self).finish()

>>> integration.finalWorkItem = Final

>>> class ReviewFinal(TechReview):
...
...     def finish(self, ed_changes):
...         self.activity.workItemFinished(self, ed_changes)

>>> integration.rfinalWorkItem = ReviewFinal

>>> class Publish:
...     zope.component.adapts(interfaces.IParticipant)
...     zope.interface.implements(interfaces.IWorkItem)
...
...     def __init__(self, participant):
...         self.participant = participant
...
...     def start(self):
...         print "Published"
...         self.finish()
...
...     def finish(self):
...         self.participant.activity.workItemFinished(self)

>>> integration.publishWorkItem = Publish

>>> class Reject(Publish):
...     def start(self):
...         print "Rejected"
...         self.finish()

>>> integration.rejectWorkItem = Reject

and a process context, so we can pass parameters:

>>> class PublicationContext:
...     zope.interface.implements(interfaces.IProcessContext)
...
...     def processFinished(self, process, decision):
...         self.decision = decision

Now, let’s try out our process. We’ll follow the same steps we did in “README.txt”, getting the same results:

>>> context = PublicationContext()
>>> proc = pd(context)
>>> proc.start('bob')
ProcessStarted(Process(u'Publication'))
Transition(None, Activity(u'Publication.start'))
ActivityStarted(Activity(u'Publication.start'))
ActivityFinished(Activity(u'Publication.start'))
Author `bob` selected
Transition(Activity(u'Publication.start'),
           Activity(u'Publication.prepare'))
ActivityStarted(Activity(u'Publication.prepare'))

>>> item = authors['bob'].work_list.pop()
>>> item.finish("I give my pledge, as an American\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my Country.")
WorkItemFinished(u'prepare')
ActivityFinished(Activity(u'Publication.prepare'))
Transition(Activity(u'Publication.prepare'),
           Activity(u'Publication.tech1'))
ActivityStarted(Activity(u'Publication.tech1'))
Transition(Activity(u'Publication.prepare'),
           Activity(u'Publication.tech2'))
ActivityStarted(Activity(u'Publication.tech2'))

>>> item = tech1.work_list.pop()
>>> print item.getDoc()
I give my pledge, as an American
to save, and faithfully to defend from waste
the natural resources of my Country.

>>> item.finish(True, ['Change "American" to "human"'])
WorkItemFinished(u'tech_review')
ActivityFinished(Activity(u'Publication.tech1'))
Transition(Activity(u'Publication.tech1'),
           Activity(u'Publication.review'))

>>> item = tech2.work_list.pop()
>>> item.finish(True, ['Change "Country" to "planet"'])
WorkItemFinished(u'tech_review')
ActivityFinished(Activity(u'Publication.tech2'))
Transition(Activity(u'Publication.tech2'),
           Activity(u'Publication.review'))
ActivityStarted(Activity(u'Publication.review'))
WorkItemFinished(u'ed_review')
ActivityFinished(Activity(u'Publication.review'))
Author `bob` selected
Transition(Activity(u'Publication.review'),
           Activity(u'Publication.prepare'))
ActivityStarted(Activity(u'Publication.prepare'))

>>> item = authors['bob'].work_list.pop()
>>> item.summary()
Previous draft:
I give my pledge, as an American
to save, and faithfully to defend from waste
the natural resources of my Country.
Changes we need to make:
Change "American" to "human"
Change "Country" to "planet"

>>> item.finish("I give my pledge, as an human\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my planet.")
WorkItemFinished(u'prepare')
ActivityFinished(Activity(u'Publication.prepare'))
Transition(Activity(u'Publication.prepare'),
           Activity(u'Publication.tech1'))
ActivityStarted(Activity(u'Publication.tech1'))
Transition(Activity(u'Publication.prepare'),
           Activity(u'Publication.tech2'))
ActivityStarted(Activity(u'Publication.tech2'))

>>> item = tech1.work_list.pop()
>>> item.finish(True, [])
WorkItemFinished(u'tech_review')
ActivityFinished(Activity(u'Publication.tech1'))
Transition(Activity(u'Publication.tech1'),
           Activity(u'Publication.review'))

>>> item = tech2.work_list.pop()
>>> item.finish(True, [])
WorkItemFinished(u'tech_review')
ActivityFinished(Activity(u'Publication.tech2'))
Transition(Activity(u'Publication.tech2'),
           Activity(u'Publication.review'))
ActivityStarted(Activity(u'Publication.review'))

>>> item = reviewer.work_list.pop()
>>> print item.getDoc()
I give my pledge, as an human
to save, and faithfully to defend from waste
the natural resources of my planet.

>>> item.finish(['change "an" to "a"'])
WorkItemFinished(u'ed_review')
ActivityFinished(Activity(u'Publication.review'))
Author `bob` selected
Transition(Activity(u'Publication.review'),
           Activity(u'Publication.final'))
ActivityStarted(Activity(u'Publication.final'))

>>> item = authors['bob'].work_list.pop()
>>> item.summary()
Previous draft:
I give my pledge, as an human
to save, and faithfully to defend from waste
the natural resources of my planet.
Changes we need to make:
change "an" to "a"

>>> item.finish("I give my pledge, as a human\n"
...             "to save, and faithfully to defend from waste\n"
...             "the natural resources of my planet.")
WorkItemFinished(u'final')
ActivityFinished(Activity(u'Publication.final'))
Transition(Activity(u'Publication.final'),
           Activity(u'Publication.rfinal'))
ActivityStarted(Activity(u'Publication.rfinal'))

>>> item = reviewer.work_list.pop()
>>> print item.getDoc()
I give my pledge, as a human
to save, and faithfully to defend from waste
the natural resources of my planet.

>>> item.finish([])
WorkItemFinished(u'rfinal')
ActivityFinished(Activity(u'Publication.rfinal'))
Transition(Activity(u'Publication.rfinal'),
           Activity(u'Publication.publish'))
ActivityStarted(Activity(u'Publication.publish'))
Published
WorkItemFinished(u'publish')
ActivityFinished(Activity(u'Publication.publish'))
ProcessFinished(Process(u'Publication'))

>>> context.decision
True

Descriptions

Most process elements can have names and descriptions.

>>> pd.__name__
u'Publication'

>>> pd.description
u'This is the sample process'

>>> pd.applications['prepare'].__name__
u'Prepare'

>>> pd.applications['prepare'].description
u'Prepare the initial draft'

>>> pd.activities['tech1'].__name__
u'Technical Review 1'

>>> pd.activities['tech1'].description
u'This is the first Technical Review.'

>>> pd.participants['tech1'].__name__
u'Technical Reviewer 1'

>>> pd.participants['tech1'].description
u'He is a smart guy.'

>>> sorted([item.__name__ for item in pd.transitions])
    [u'Transition', u'Transition', u'Transition', u'Transition',
    u'Transition', u'Transition', u'Transition', u'Transition',
    u'Transition', u'Transition', u'Transition to Tech Review 1',
    u'Transition to Tech Review 2']

>>> sorted([item.description for item in pd.transitions])
    [None, None, None, None, None, None, None, None, None, None, None,
    u'Use this transition if there are editorial changes required.']

CHANGES

3.5.0 (2009-07-24)

• Update tests to latest package versions.

3.4.0 (2007-11-02)

• Initial release independent of the main Zope tree.