#!/usr/bin/env bash
# Download the third-party inputs that this repository does not redistribute.
#   raw/tg/   TeleGeography Submarine Cable Map (cables.py, network.py, build.py, export.py)
#   raw/wn/   WonderNetwork 2020-07-19/20 ping archive (analyze_wn.py)
#   raw/bgp/  RouteViews prefix2as, RIPE NCC AS names, PeeringDB (asmap.py)
# The survey used the TeleGeography data as served on 2026-10-23; raw/tg.sha256
# lists that snapshot. Later downloads may differ and change results slightly.
set -euo pipefail
cd "$(dirname "$0")"
UA="cable-latency-survey/1.0 (+https://pingstats.aethercosmology.com)"
get() { curl -sSf -A "$UA" --retry 3 -o "$1" "$2"; }
mkdir -p raw/tg/cables raw/wn raw/bgp

TG=https://www.submarinecablemap.com/api/v3
get raw/tg/cable-geo.json "$TG/cable/cable-geo.json"
get raw/tg/landing-point-geo.json "$TG/landing-point/landing-point-geo.json"
python3 - "$TG" "$UA" <<'PY'
import json, sys, time, urllib.request
tg, ua = sys.argv[1], sys.argv[2]
ids = sorted({f["properties"]["id"] for f in json.load(open("raw/tg/cable-geo.json"))["features"]})
for n, i in enumerate(ids, 1):
    req = urllib.request.Request(f"{tg}/cable/{i}.json", headers={"user-agent": ua})
    open(f"raw/tg/cables/{i}.json", "wb").write(urllib.request.urlopen(req, timeout=60).read())
    time.sleep(0.2)
print(f"TeleGeography: {len(ids)} cables")
PY
get raw/tg/tga.json "$TG/cable/tasman-global-access-tga-cable.json"
(cd raw && sha256sum -c --quiet tg.sha256 2>/dev/null && echo "TeleGeography data identical to the 2026-10-23 snapshot") \
  || echo "note: TeleGeography data differs from the 2026-10-23 snapshot (raw/tg.sha256)"

S3=https://wp-public-data.s3.amazonaws.com/pings
get raw/wn/pings-2020-07-19-2020-07-20.csv.gz "$S3/pings-2020-07-19-2020-07-20.csv.gz"
get raw/wn/servers-2020-07-19.csv "$S3/servers-2020-07-19.csv"

get raw/bgp/routeviews-rv2-20261023-0200.pfx2as.gz \
  https://publicdata.caida.org/datasets/routing/routeviews-prefix2as/2026/10/routeviews-rv2-20261023-0200.pfx2as.gz
get raw/bgp/asn.txt https://ftp.ripe.net/ripe/asnames/asn.txt
for t in ixpfx ixlan ix; do get "raw/bgp/pdb_$t.json" "https://www.peeringdb.com/api/$t"; sleep 4; done
echo "done; PeeringDB and AS names are current, not the 2026-10-27 copies used for data/asmap.json"