Yes, the API (which I will add a link to in the google code site wiki) specifies getters and setters for the image file used as a colormap and the image file used for point-plotting
Very cool looking! I am going to consider using this for an upcoming project.
One of my most significant criteria for using a library is its API and documentation. Maybe I'm not looking in the right places, but I had trouble finding API.txt in your repo; consider including a link to that from your project page. Also, it would be helpful if you provided an example script in each language on your website as well.
It seems quite slow. I am using Python 2.6.4 on Linux, slow processor (~2GHz). I out-commented the testOpacityFrequency and _compare defs in the test/pil.py. Now I added about 10000 points and it is calculating for half an hour already. With 1000 points it tool 5 minutes.
Is that normal performance or did I do something wrong?
edit: after 36 minutes it threw "AttributeError: 'LoadTest' object has no attribute '_compare'", oops that was my fault. Trying again with less points.
edit2: It is swapping... I have 2GB RAM. Using an 8192x8192 image and 20 points. :-(
One question though: where does the screenshot of the item which has been clicked on come from? As far as I understand your lib does not provide any means to create a screenshot? Is this correct? Would it be possible to add this? If not could you explain how you would make sure the coordinates and the screenshot at any given time match?
Just curious and interested in using your library :)
Coordinate <-> image matching seems to be the job of the coder. I have not looked at the code yet but I am sure it does not do screenshots. What it probably does is this:
-You provide an image.
-You provide x,y coordinates.
-The tool/library calculates a heatmap from the relationship between coordinates
-The tool overlays the heatmap 1:1 to the image
The Google screenshot and clicking demo is just an example of a possible implementation. You can use heatmaps for many many other things so this surely does not limit itself (nor prefer) to website click tracking. ;-)
Exactly, it expects you to provide the image. The coordinates are currently expected to be absolute and pixel-specific, where 55,11 is 55 pixels inward and 11 downward.
If there was sufficient interest, I could implement relative positioning as well, where they resolve to a percentage of the source image.
I like it, although I'm having a hard time thinking of a serious case where I can use it other than for example making another CrazyEgg, which I don't think I have the time for.
TINY request - add a reset button to the example. I guess it's saved per session or something, but I spammed a bunch of clicks in one spot which dominates the page a bit.
Very cool! This will be awesome for usability testing.
I like the name too, in fact I liked it enough to create a jQuery templating plugin with the same name a few months ago ;)http://github.com/fitzgen/tempest/
This comes 3 weeks to late for me but I will definitely rewrite something to use this instead of my half-assed workaround. Will message you with hopefully tasty screenshots once I am done.
In fact, it was only one bug (checked on the email I send them): it didn't track clicks on images. So any sort of "Sign Up" button weren't tracked, manking the service worthless to me.