I just made a new release of epydemic
.
It’s a fairly incremental set of changes, but dominated by a roughly
7x speed increase that came from replacing the implementation of event
loci with a dedicated data structure rather than using Python’s built-in set
implementation — and replacing one line of code
with about 500 lines. This removed a constraint whose complexity
was linear in the number of edges due to an accident of coding
(i.e., due to my stupidity), and replaced it with something that was
both significantly faster and significantly more scalable.
I also added a discussion of moving from $\mathcal{R}$ values to the probabilities required for simulation, to make things easier for those more comfortable with this style of epidemic description.
The new version can be installed using pip install -U
epydemic
, or from Github.