#!/usr/bin/env python3

"""
The experiment log files should be in the same directory as this script. The fingerprint files should be in ./devices.
"""

import glob
import json
import statistics

from lib.utils import fuzzy_intersection

high_risk = ("DNS", "open port")


def ll_to_ht(ll):
    return {(x[0], x[1]) for x in ll}


def get_jaccard(l, r):
    l, r = ll_to_ht(l), ll_to_ht(r)
    intersection = fuzzy_intersection(l, r)
    union = len(l) + len(r) - intersection
    return (intersection / union) * 100.0


def get_vanilla_jaccard(l, r):
    l, r = ll_to_ht(l), ll_to_ht(r)
    intersection = len(l.intersection(r))
    union = len(l.union(r))
    return (intersection / union) * 100.0


dts = {}


for device_file in glob.glob("devices/*"):
    with open(device_file, "r") as device_type_file:
        uuid = device_file.replace("\\", "/").split("/")[-1]
        dts[uuid] = json.load(device_type_file)

for fn in glob.glob("experiment-log*.txt"):
    with open(fn, "r") as f:
        parts = f.read().split("--- next run ---")
        id_line = parts[0].split("Recording for device type:")[1].splitlines()[0]
        rec_uuid = id_line.split("uuid=")[1].split(",")[0]
        rec_ip = id_line.split("at ")[1]

        parts = parts[1:]

        scores_no_high_risk, count, scores = [], [], []

        next_best = ()

        for p in parts:
            last_record = p.split("Local device at {} has characteristics".format(rec_ip))[-1].splitlines()[0].rsplit(",", 1)[0]
            last_record = json.loads(last_record)

            # get_jaccard = get_vanilla_jaccard

            scores.append(get_jaccard(last_record, dts[rec_uuid]["characteristics"]))
            scores_no_high_risk.append(get_jaccard([l for l in last_record if l[0] not in high_risk], [l for l in dts[rec_uuid]["characteristics"] if l[0] not in high_risk]))

            next_best = max([(get_jaccard(d["characteristics"], last_record), d["name"]) for d in dts.values() if
                             d["name"] != dts[rec_uuid]["name"]], key=lambda x: x[0])

            next_best_high_risk = max([(get_jaccard([l for l in d["characteristics"] if l[0] not in high_risk], [l for l in last_record if l[0] not in high_risk]), d["name"]) for d in dts.values() if
                                       d["name"] != dts[rec_uuid]["name"]], key=lambda x: x[0])

            count = []
            types = set([r[0] for r in last_record] + [r[0] for r in dts[rec_uuid]["characteristics"]])
            for r in types:
                count.append((r, len([c for c in last_record if c[0] == r])))

         print("Device {}:\tScores {}\t({} without high risk),\tmean {}\t({}),\tvariance {}\t({}),\tstdev {}\t({}),\tnext best: {}\tdist: {}\t({}),\tcount: {}".format(
             dts[rec_uuid]["name"],
             scores,
             scores_no_high_risk,
             statistics.mean(scores),
             statistics.mean(scores_no_high_risk),
             statistics.variance(scores),
             statistics.variance(scores_no_high_risk),
             statistics.stdev(scores),
             statistics.stdev(scores_no_high_risk),
             next_best,
             statistics.mean(scores)-next_best[0],
             next_best_high_risk,
             count))