planner.py

# planner.py
# Ronald L. Rivest (with Karim Husayn Karimi)
# August 1, 2017
# python3

"""
Routines to work with multi.py on post-election audits.
Specifically, to produce an "audit plan" for the next stage,
given what has been done already, and the results obtained
from the previous stage.
"""


##############################################################################
# Compute audit plan for next stage

def compute_plan(e):
    """ 
    Compute a sampling plan for the next stage.
    Put in e.plan_tp[e.stage_time] a dict of target sample sizes 
    keyed by pbcid. 
    Only input is contest statuses, pbcid audit rates, pbcid current
    sample size, and pcbid size.
    """

    # for now, use simple strategy of looking at more ballots
    # only in those paper ballot collections that are still being audited
    e.plan_tp[e.stage_time] = e.sn_tp[e.stage_time].copy()
    pbcids_to_adjust = set()
    for mid in e.mids:
        cid = e.cid_m[mid]
        for pbcid in e.possible_pbcid_c[cid]:
            if e.status_tm[e.stage_time][mid] == "Open":
                pbcids_to_adjust.add(pbcid)
    for pbcid in pbcids_to_adjust:
        # if contest still being audited do as much as you can without
        # exceeding size of paper ballot collection
        # CHECK: is e.rn_p[pbcid] right number to use here?
        e.plan_tp[e.stage_time][pbcid] = \
             min(e.sn_tp[e.stage_time][pbcid] + e.max_audit_rate_p[pbcid],
                 e.rn_p[pbcid])
    return