summarize-locs

#! /usr/bin/python3

"""Extract summary information from all of the location matrices named
on the command line.  Output is in CSV format and is written to stdout."""

import os
import sys
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "lib")))

import ageo
import argparse
import collections
import csv
import datetime
import glob
import multiprocessing
import time

_time_0 = time.monotonic()
def progress(message, *args):
    global _time_0
    sys.stderr.write(
        ("{}: " + message + "\n").format(
            datetime.timedelta(seconds = time.monotonic() - _time_0),
            *args))

def produce_summary_row(fname):
    loc  = ageo.Location.load(fname)
    cen  = loc.centroid
    rep  = loc.rep_pt
    area = loc.area
    ann  = loc.annotations

    if ann["proxied"]:
        addr = ann["proxy_addr"]
        epr  = ann["estimated_proxy_rtt"]
        prov = ann["proxy_provider"]
    else:
        addr = ann["client_addr"]
        epr  = 0
        prov = "none"

    return [addr, prov, epr, rep[0], rep[1], cen[0], cen[1], area]

def main():
    ap = argparse.ArgumentParser(description=__doc__)
    ap.add_argument("files", nargs="+",
                    help=".h5-format location matrices to summarize.")
    args = ap.parse_args()

    with multiprocessing.Pool() as pool, sys.stdout as outf:
        wr = csv.writer(outf, dialect="unix", quoting=csv.QUOTE_MINIMAL)
        wr.writerow((
            "ip", "provider", "epr",
            "longitude", "latitude", "c.longitude", "c.latitude",
            "area"))

        n_todo = len(args.files)
        n = 0
        for row in pool.imap_unordered(produce_summary_row, args.files):
            n += 1
            if row[5] > 0:
                wr.writerow(row)
            progress("{}/{}: {}/{}: {:.4f},{:.4f}x{:.3f}".format(
                n, n_todo, row[1], row[0], row[4], row[3], row[7]/(1000*1000)))

main()