Just put it in your python path and add it into MIDDLEWARE_CLASSES.
I know that there are a couple of snippets on this site that do something similar to this, but none of them quite suited me.
I wanted something that would indent and space the SQL so that I could differentiate it from the other output from the development server.
Also, I wanted something that would output total query execution time, which my solution does. I just hope that it's useful for someone else, too!
UPDATE: Now this should no longer get upset when running it on windows.
Displays hotshot profiling for any view.
http://yoursite.com/yourview/?prof
Add the "prof" key to query string by appending ?prof (or &prof=)
and you'll see the profiling results in your browser.
It's set up to only be available in django's debug mode,
but you really shouldn't add this middleware to any production configuration.
* Only tested on Linux
* You should probably use this one instead: http://djangosnippets.org/snippets/2126/
Intall
--------
In your settings.py, set it in MIDDLEWARE_CLASSES. Default all the profile files will be save in ./profile folder(if there is no this directory, it'll automatically create one), and you can set a PROFILE_DATA_DIR option in settings.py, so that the profile files can be saved in this folder.
How does it works
-------------------
Record every request in a profile file. As you can see in the code:
profname = "%s.prof" % (request.path.strip("/").replace('/', '.'))
so if you request an url multi-times, only the last request will be saved, because previous profile files will be overriden. And you can find a commentted line, it'll save each request in different file according the time, so if you like that you can change the code.
# profname = "%s.%.3f.prof" % (request.path.strip("/").replace('/', '.'), time.time())
and if you want to see the output, you can run below code, but you should change the filename value:
import hotshot, hotshot.stats
stats = hotshot.stats.load(filename)
#stats.strip_dirs()
#stats.sort_stats('time', 'calls')
stats.print_stats()