jangevaare/ilmpythontools

Tools to simulate and perform inference of individual level infectious disease transmission models in python

ilmpythontools's Issues

Package organization

Properly organize modules into a package..

AttributeError: exp

Find what's causing this...

edb1 = infect(pop, 20, 1, edb1, 2)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-29-f20974850e7a> in <module>()
----> 1 edb1 = infect(pop, 20, 1, edb1, 2)

<ipython-input-2-4fb6aa0cdbfb> in infect(pop, alpha, beta, event_db, time)
     50     infectious = find_infectious(pop, event_db, time)
     51     susceptible = find_susceptible(pop, event_db, time)
---> 52     prob = infect_prob(pop, alpha, beta, infectious, susceptible)
     53     new_infections = susceptible.index[np.where(prob < np.random.uniform(0, 1, size=prob.size))]
     54     return event_db.append(pd.DataFrame({"time":np.repeat(time, new_infections.size), 

<ipython-input-2-4fb6aa0cdbfb> in infect_prob(pop, alpha, beta, infectious, susceptible)
     44 def infect_prob(pop, alpha, beta, infectious, susceptible):
     45     """Determine infection probabilities for each susceptible individual following ILM framework."""
---> 46     return 1-np.exp(np.multiply(-alpha, kappa(pop, beta, infectious, susceptible)))
     47 
     48 def infect(pop, alpha, beta, event_db, time):

AttributeError: exp 

Add visualization functionality

Either utilize matplotlib or pandas for this... animation creation would be useful as well.

Epidemic curves

Add functionality to plot epidemic curves from any event_db

MCMC for SIR models

Finish development of MCMC algorithm for SIR models - applicable to constant and random recovery periods

Report time and other statistics during/after MCMC

time per iteration, time remaining, functionality for producing updating trace plots, status bar, etc. would all serve as enhancements.

Catch when an attempt is made to generate events for a time which has already occurred

Currently it is possible to generate new infections at any time. Must alter this to either raise an assertion error (easier), or to rewrite the event database when permission is granted (more difficult).

Full examples

Provide full sets of examples for all package functionality in ipython notebooks