sciencefair2023/analyze.py

import sys
import requests
from math import radians, cos, sin, acos, asin, sqrt
from pathlib import Path
from os import chdir
from datetime import datetime, timezone
from dateutil import tz
import pickle

class IPCache:
    picklefile_name = 'cached_ips.pkl'

    def __init__(self, /):
        with Path('.') as p:
            self.creation_dir = p.resolve()

    def get(self, ip, /):
        if ip in self.cache:
            return self.cache[ip]
        addr = f'http://ip-api.com/json/{ip}'
        response = requests.get(addr)
        if not response.ok:
            raise RuntimeError(f'request for ip failed with {response.status_code}')
        resulting_dict = eval(response.content)
        if resulting_dict['status'] == 'fail':
            raise RuntimeError(f'ip was invalid')
        # the given timezone is like, 'Australia/Sydney'. we need to convert to
        # a datetime.timezone type
        timezone_str = resulting_dict['timezone']
        tzfile = tz.gettz(timezone_str)
        as_timedelta = tzfile.utcoffset(datetime.utcnow())
        as_timezone_type = timezone(as_timedelta)

        self.cache[ip] = ((resulting_dict['lat'], resulting_dict['lon']), as_timezone_type)
        return self.cache[ip]

    def __enter__(self, filename=picklefile_name, /):
        chdir(self.creation_dir)
        with Path(filename) as p:
            if p.exists():
                with open(filename, 'rb') as f:
                    cache = pickle.load(f)
            else:
                cache = {}
        self.filename = filename
        self.cache = cache
        return self

    def __exit__(self, err_type, err_value, traceback, /):
        chdir(self.creation_dir)
        with open(self.filename, 'wb') as f:
            pickle.dump(self.cache, f)

def filter_logs(filename: str):
    with open(filename, 'r') as f:
        for line in f:
            # IP,YYMMDD hhmmss TIMEZONE,STATUS,METHOD,FILE
            ip, timethings, status, method, filepath = line.split(',', 4)
            file = '/'.join(filepath.split('/')[4:]).strip()
            if (
                status != '200'
                or method != 'GET'
            ): continue
            yield ip, timethings, status, method, file

def latlon_distance(p1, p2) -> float:
    # black magic do not touch. use the haversine formula to find the distance
    lat1, lon1 = p1
    lat2, lon2 = p2
    lon1 = radians(lon1)
    lat1 = radians(lat1)
    lon2 = radians(lon2)
    lat2 = radians(lat2)

    dlon = lon2 - lon1
    dlat = lat2 - lat1
    a = sin(dlat / 2)**2 + cos(lat1) * cos(lat2) * sin(dlon / 2)**2

    c = 2 * asin(sqrt(a))
    earth_radius_km = 6371

    return c * earth_radius_km

def analyze_server(server: Path) -> None:
    if not server.is_dir():
        return
    serverip = get_server_ip(server.name)

    result = ''

    filename = f'{server.name}/access.log'
    self_latlon, _ = ip_cache.get(serverip)
    for ip, timethings, status, method, file in filter_logs(filename):
        # get_ip_latlon_tz() returns the latitude, longitude, and timezone
        # of an ip
        # TODO: cache results in a picklefile
        latlon, timezone = ip_cache.get(ip)

        # convert `timethings` to a datetime object
        time_of_hit = datetime.strptime(timethings, log_date_format)
        # convert from its default timezone in GMT to the timezone of the requester
        localtime = time_of_hit.replace(tzinfo=timezone.utc).astimezone(timezone)
        # convert this time back to a string for logging purposes, without timezone
        localtime_str = ' '.join(localtime.strftime(log_date_format).split(' ')[:2])

        distance = latlon_distance(self_latlon, latlon)
        date, time = timethings.split(' ')[:2]
        result += (
            f'{server.name},{ip},{distance:.5f},{date},{time},'
            + f'{localtime_str},{method},"{file}"\n'
        )

    return result

def get_server_ip(servername: str) -> str:
    # associate servers with ips
    if servername == 'nova':
        return '184.73.25.153'
    elif servername == 'singapore':
        return '18.139.108.77'
    elif servername == 'sydney':
        return '54.206.216.118'
    elif servername == 'dublin':
        return '54.194.92.137'
    elif servername == 'brazil':
        return '18.228.245.48'

    raise ValueError(f'{servername} is not a known server')

def main(args: list) -> int:
    if len(args) == 1:
        # no log dirs provided
        print('no logdir provided')
        return 1

    outfile = 'analysis.csv'
    start_dir = Path('.').resolve()

    with open(outfile, 'w') as f:
        for logdir in args[1:]:
            chdir(logdir)
            serverdir = Path('.')
            for subdir in serverdir.iterdir():
                csv_lines = analyze_server(subdir)
                f.write(csv_lines)
            chdir(start_dir)

    return 0

log_date_format = r'%y/%m/%d %H:%M:%S %z'

if __name__ == '__main__':
    with IPCache() as ip_cache:
        sys.exit(main(sys.argv))