Currently, when a ``min_count`` filter is used, the logarithmic
distribution doesn't work as well as it could, IMHO.
Imagine max count is 400, and min count is either filtered to be 100 (or
possibly really is 100, even without a filter).
The minimum weight of any tag in the cloud then is:
log(100) * 400 / log(400) = ~307
In a 4-step distribution, this will be font size 2 - size 1 is skipped
completely.
The linear algorithm does not have that problem, since the weights are not
modified.
I'd provide a patch, but unfortunately, my math skills fail me once again.
I managed to modify ``calculate_cloud`` so that everything is shifted down,
e.g.
max_weight = max_weight - min_weight
min_weight = 1
But that may not be the best solution. The change should probably be inside
``_calculate_tag_weight``, if possible.
Original issue reported on code.google.com by elsdoer...@gmail.com on 4 Aug 2008 at 8:19
Original issue reported on code.google.com by
elsdoer...@gmail.com
on 4 Aug 2008 at 8:19