netsight.async 1.1.0

Provides a base view for running asynchronous processes from Zope.

Introduction

netsight.async provides a base browser view for the Zope Web Framework which enables browser requests to run in the background whilst progress of the request is returned to the browser.

Usage

Basic

First, subclass the BaseAsyncView class. Where you might ordinarily write code in the __call__ method of a view class, to perform some process, instead, place it in the __run__ method.

>>> import time
>>> from netsight.async.browser.BaseAsyncView import BaseAsyncView
>>> class MyView(BaseAsyncView):
...
...    def __run__(self, *args, **kwargs):
...         time.sleep(30)
...         return "Hello world!"
...
>>>

When you call this view from the browser, it will display whatever output is configured as normal. When you perform a POST request to the view, however, the __run__ method will be called in the background as if it were the __call__ method of the view class. Meanwhile, a page displaying a spinner will be returned to the browser and will poll at 5 second intervals until the process defined in __run__ completes.

Once the __run__ method has completed, the browser will redirect to the result.

Example timeline:

1. User visits ‘/myview’ and is shown some form.

2. User submits the form, the process is started in the background.

3. The user is redirected to ‘/myview/processing?process_id=abcde-f01234’ which shows a spinner.

4. The user’s current page polls for process status for up to 30 seconds via AJAX if possible, otherwise by page refresh.

5. Once the process has completed, the user is redirected to ‘/myview/result?process_id=abcde-f01234’ and shown “Hello world!”

Using with page templates

If you have configured a browser view for your view class with a page template file specified in ZCML, this will be shown by default when the view is first called. If the view is POSTed to, then the process will be kick-started. You can change the initial template and the conditions under which the process is started by overriding the initial_page and the run_process methods of the view.

>>> from Products.Five.browser.pagetemplatefile import \
...      ViewPageTemplateFile
>>>
>>> class MyView(BaseAsyncView):
...
...     def run_process(self):
...         return 'run' in self.request.form
...
...     initial_page = \
...         ViewPageTemplateFile('templates/my_template.pt')
...
>>>

Or you could use a method:

>>> class MyView(BaseAsyncView):
...
...     def initial_page(self, *args, **kwargs):
...         return 'Hello world!'
...
>>>

You can also override the page returned to the browser once the process has been started by overriding the processing_page method.

If you want to call a template defined in ZCML from your __run__ method, you may pass a True value named no_process` to the call method if your ``run_process method would ordinarily start the process again.

>>> class MyView(BaseAsyncView):
...
...     def __run__(self, *args, **kwargs):
...         return self.__call__(message="Hello world",
...                              no_process=True)
...
>>>

Checking the status & retrieving the result

Once you have kicked off your __run__ method, the resulting response will redirect to the processing view, with a unique ID for the newly started process given as a GET variable, process_id.

This process ID can be used to retrieve information on the status of the process and its result.

Calling the completed method of the view with the process ID will return either a True or False completion state, or a number representing a percentage completion out of 100 (more on recording progress later). If the optional argument, output_json is set to some value which evaluates to True, the method returns a JSON object with the single key, completed containing the same True, False or numeric value.

If your process died before it completed, it will raise an error, or if JSON output is chosen, it will return a completed value of the string, ‘ERROR’.

To retrieve the result of the __run__ method once it has completed, call the result method of the view with the process ID.

If your process died before it completed, this too will raise an error, or if JSON output is chosen, it will return a completed value of the string, ‘ERROR’.

If the process has not yet completed when result is called, None will be returned.

Setting process progress from your task

If you want your task to return some measure of completion you can call the set_progress method with the process ID and some numeric value.

>>> class MyView(BaseAsyncView):
...
...    def __run__(self, process_id=None, *args, **kwargs):
...         time.sleep(15)
...         self.set_progress(process_id, 50)
...         time.sleep(15)
...         return "Hello world!"
...
>>>

When your task completes without raising an exception, the progress is automatically set to 100 so there is no need to set this before the method returns.

Installation

Simply add netsight.async to the eggs section of your buildout configuration. If you also plan on using the stock ‘processing’ page, you may also need to add it to the zcml section.

[buildout]
eggs = ...
       netsight.async
zcml = ...
       netsight.async

Limitations

Because running the new process cannot be done using existing threads from the Zope pool, for the duration of the asynchronous process, an extra thread is created by the Zope process, beyond the normal thread limit. This also means an extra connection is opened to the ZODB beyond the normal connection limit which may cause a warning to be shown in either the console or log files.

Once the __run__ method has started, it cannot be stopped by the user in any way. This a feature that subclasses may implement if they choose, but would be dangerous to implement in this package without knowledge of what the background task was doing & what cleanup may be required.

To be improved

Currently processes are stored in memory of a particular Python instance. This introduces the following issues:

• If the user never retrieves the results from the __run__ method, they are stored in the ZODB permanently.

Dependencies

• Python>=2.4.0

• zope.component>=3.4.0

• zope.i18n>=3.4.0

• zope.i18nmessageid>=3.4.0

• zope.publisher

• Zope>=2.8.0

The default processing page template depends on a main template being, provided, much like the one provided by Products.CMFPlone, however this may be overridden by your own view, as discussed above.

Contributions

You can find the source code for this project at:

http://github.com/netsight/netsight.async

This product needs translations! There are only 2 strings to do, so this is a really quick and easy way to contribute to an open-source project.

Any bug fixes, new features & documentation improvements are welcome, just submit a pull request on github.

Changelog

1.1.0 - (2011-10-27)

• Made Python 2.4 compatible.

• Process progress/result storage is now held in the ZODB, rather than in a Python instance’s memory.

1.0.1 - (2011-10-19)

• Fixed package manifest.

• Fixed markup on processing page.

1.0.0 - (2011-10-19)

Initial release.

License

Copyright 2011 Netsight Internet Solutions Limited

Licensed 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.