ActiveState Code

a webserver which serves files from a specified directory and transforms files containing reStructuredText to HTML on the fly.

Comments

1. 1. At 10:49 p.m. on 9 may 2005, Michele Simionato said:

   is there any caching here? I really like this recipe. However, I wonder what happens if 1000 users call the same page: is the rendering being executed 1000 times or is there some caching mechanism built in the Resource class? If not, I guess I would need to change the render method in something like this:

   def render(self, request):
       try:
           return self.html
       except AttributeError:
           self.html = publish_string(self.rst, writer_name = 'html')
           return self.html
   

   Also, what happens for long pages such that the rendering takes a long time? Should I execute the rendering in a separate thread? I really don't know what Resource does under the hood.

                 Michele Simionato
   

2. 2. At 11:43 p.m. on 9 may 2005, Michele Simionato said:

   thinking a bit more ... After some experiments, I see that twisted does not cache the resources by default. Actually, each time I click on the link the a new resource object is created and the .rst file is re-read. Assuming the .rst file does not change often, this is a waste. A solution would be to memoize resource creation with a Memoize metaclass:

   class Memoize(type):
       @memoize # a memoize decorator, there are many in the cookbook
       def __call__(cls, *args):
           return super(Memoize, cls).__call__(*args)
   
   class ReStructured(Resource, object):
       __metaclass__ = Memoize
       def __init__(self, filename, *args):
           self.rst = open(filename).read()
           self.html = publish_string(self.rst, writer_name = 'html')
       def render(self, request):
           return self.html
   

   Now one should add an "update" method to reinitialize the resource object when the .rst source file changes on the filesystem, but that is easy.

3. 3. At 5:52 p.m. on 12 may 2005, Andrew Bennetts said:

   Twisted's Resource doesn't thread. So if a .rst file is large enough, it'll block all other activity in that Twisted server.

   The simplest way to fix that would be:

   from twisted.internet.threads import deferToThread
   from twisted.python import log
   from twisted.web.server import NOT_DONE_YET
   ...
   
       def render(self, request):
           d = deferToThread(publish_string, self.rst, writer_name='html')
           d.addCallback(request.write)
           d.addCallback(lambda _: request.finish())
           d.addErrback(log.err)
           return NOT_DONE_YET
   

4. 4. At 10:01 a.m. on 16 may 2005, Markus Gritsch said:

   Snakelets version. This is a version of this recipe for the Snakelets web application server.

   • Download and extract Snakelets from http://snakelets.sourceforge.net/

   • Create a new directory in the 'webapps' folder named 'rest'

   • Create a file named '__init__.py' containing the code below in the 'rest' folder

   • Run Snakelets 'serv.py' script

   • Point your browser at http://localhost:9080/rest/

     from snakeserver.snakelet import Snakelet from docutils.core import publish_string

     class ReStructured(Snakelet): def serve(self, request, response): filepath = self.getWebApp().getFileSystemPath() + '/..' + \ request.getRequestURL().replace('%20', ' ').rsplit('.', 1)[0] + '.txt' f = open(filepath, 'rt'); cont = f.read(); f.close() response.getOutput().write(publish_string(cont, writer_name = 'html'))

     name='Snakelets reStructuredText Webapp' docroot='.' snakelets= { '.txt': ReStructured, # reStructuredText files have this extension '.html': ReStructured # to be able to navigate in the generated page }

     def dirListAllower(path): return True

Sign in to comment

Tags

• web