Traceroute AS parsing

gistfile1.txt

# Run "traceroute -n google.com | tee traceroute_output.txt" in your terminal
# Paste your traceroute/tracert output between the triple quotes and run this cell.
TR_TEXT = """
 <PASTE YOUR TRACEROUTE OUTPUT HERE>
"""

import re, requests, ipaddress, sys

def extract_hops(text: str):
    """
    Works for macOS/Linux `traceroute` and Windows `tracert`.
    Returns an ordered list of hop IPs (one per hop), skipping '*' lines.
    """
    hops = []
    for line in text.splitlines():
        line = line.strip()
        if not line:
            continue
        # Grab the first IPv4 address on the line (good proxy for the hop's responding router)
        m = re.search(r'(\d{1,3}(?:\.\d{1,3}){3})', line)
        if not m:
            continue
        ip = m.group(1)
        # Skip placeholder '*' hops
        if ip == '*':
            continue
        # Keep order by hop (don't dedupe globally yet; traceroute already has 1 IP per hop line typically)
        hops.append(ip)
    # Some traceroute variants may echo the same IP on multiple adjacent lines; collapse consecutive duplicates.
    deduped = []
    for ip in hops:
        if not deduped or deduped[-1] != ip:
            deduped.append(ip)
    return deduped

def ip_class(ip):
    try:
        addr = ipaddress.ip_address(ip)
        if addr.is_private: return "Private"
        if addr.is_loopback: return "Loopback"
        if addr.is_link_local: return "Link-local"
        if addr.is_reserved: return "Reserved"
        if addr.is_multicast: return "Multicast"
        if ip.startswith("100.64.") or ip.startswith("100.65.") or ip.startswith("100."):  # CGNAT (approx)
            # ipaddress has .is_private True for CGNAT; keep label explicit for readability:
            return "CGNAT/Private"
        return "Public"
    except ValueError:
        return "Unknown"

def ripe_asn_lookup(ip):
    """
    Query RIPEstat for announcing ASN and prefix for an IP.
    Returns dict with asn, holder, prefix, rir (or minimal info if not found).
    """
    try:
        r = requests.get(
            "https://stat.ripe.net/data/prefix-overview/data.json",
            params={"resource": ip},
            timeout=10,
        )
        r.raise_for_status()
        d = r.json().get("data", {})
        asns = d.get("asns", []) or []
        if asns:
            a = asns[0]
            return {
                "asn": f"AS{a.get('asn')}",
                "holder": a.get("holder") or "",
                "prefix": d.get("prefix") or "",
                "rir": d.get("rir") or "",
            }
        # Fallback: still return prefix if present
        return {
            "asn": "",
            "holder": "",
            "prefix": d.get("prefix") or "",
            "rir": d.get("rir") or "",
        }
    except Exception:
        return {"asn": "", "holder": "", "prefix": "", "rir": ""}

# Process
hops = extract_hops(TR_TEXT)

# Build rows with ASN mapping (cache lookups)
cache = {}
rows = []
for i, ip in enumerate(hops, start=1):
    kind = ip_class(ip)
    if kind != "Public":
        rows.append({"hop": i, "ip": ip, "type": kind, "asn": "", "holder": "", "prefix": "", "rir": ""})
        continue
    if ip not in cache:
        cache[ip] = ripe_asn_lookup(ip)
    info = cache[ip]
    rows.append({
        "hop": i, "ip": ip, "type": kind,
        "asn": info.get("asn",""), "holder": info.get("holder",""),
        "prefix": info.get("prefix",""), "rir": info.get("rir","")
    })

# Pretty print as Markdown table
def to_markdown_table(rows):
    headers = ["Hop", "IP", "Type", "ASN", "Holder", "Prefix", "RIR"]
    out = ["| " + " | ".join(headers) + " |", "|" + "|".join(["---"]*len(headers)) + "|"]
    for r in rows:
        out.append("| " + " | ".join(str(r.get(k.lower(), r.get(k, ""))) for k in ["hop","ip","type","asn","holder","prefix","rir"]) + " |")
    return "\n".join(out)

md = to_markdown_table(rows)
print(md)

# If pandas is available, also show a nice table
try:
    import pandas as pd
    df = pd.DataFrame(rows, columns=["hop","ip","type","asn","holder","prefix","rir"])
    from IPython.display import display
    display(df)
except Exception as e:
    print("\n(pandas not available or display failed; Markdown table printed above.)")