Tornado, Accessing additional data in callback function?

Question

I have just started a project using Tornado, and asyncmongo.

I have a handler with an async method. Inside I am querying mongo for some words:

@tornado.web.asynchronous
def get(self):
    word = self.get_argument('word', None)
    if not word:
        self.write('{}')
        self.finish()
    self.db.spanish_dict.find({'$or': [{'word': word}, {'stem': word}]},
                              callback=self._on_response)


def _on_response(self, response, error):
   # need to sort response by relevancy

In my callback method I need the original word to sort the mongo results accurately.

I found this post which uses functools.partial to accomplish this, by allowing me to pass additional parameters to the callback method

I was wondering if there are any adverse affects to setting an instance attribute in the get method and accessing it in _on_response ? THank you

@tornado.web.asynchronous
def get(self):
    word = self.get_argument('word', None)
    if not word:
        self.write('{}')
        self.finish()
    self.word = word
    self.db.spanish_dict.find({'$or': [{'word': word}, {'stem': word}]},
                              callback=self._on_response)


def _on_response(self, response, error):
   # need to sort response by relevancy
   # will self.word always be accurate?
   self.word

Answer 1

Use tornado.gen and you sidestep the problem completely

http://www.tornadoweb.org/documentation/gen.html?highlight=tornado.gen#tornado.gen