har.py

#!/usr/bin/env python3

import tabulate
from pathlib import Path
import json
import base64
import sys
from urllib.parse import urlparse
import mimetypes

har = json.loads(Path(sys.argv[1]).read_text())

entries = har["log"]["entries"]

t = []
num = 0
for entry in entries:
    mimetype = entry["response"]["content"]["mimeType"]

    # if mimetype in ("text/css", "font/woff2", "image/gif"):
    #     continue

    # size = entry["response"]["content"]["size"]
    # if size == 0:
    #     continue

    url = urlparse(entry["request"]["url"])

    if any(
        map(
            url.hostname.endswith,
            (
                ".google-analytics.com",
                ".snapchat.com",
                ".doubleclick.net",
            ),
        )
    ):
        print(f"skip {url.hostname}")
        continue

    path = Path(url.path)
    filename = path.stem
    response_text = entry["response"]["content"].get("text")
    encoding = entry["response"]["content"].get("encoding")

    shortened_path = url.path
    if len(shortened_path) > 40:
        shortened_path = shortened_path[:40] + "…"

    suffix = mimetypes.guess_extension(mimetype)
    if not suffix:
        suffix = ".bin" if encoding == "base64" else ".txt"
    num += 1
    f = Path(f"r{num:03d}").with_suffix(suffix)

    t.append(
        (
            f,
            url.hostname,
            shortened_path,
            entry["response"]["content"]["size"],
            mimetype,
            len(response_text.encode()) if response_text else "",
            encoding,
        )
    )

    if response_text:

        if encoding == "base64":
            f.write_bytes(base64.b64decode(response_text))
        elif encoding is None:
            f.write_text(response_text)
        else:
            print(f"error encoding {encoding}")
            exit()


print(tabulate.tabulate(t, tablefmt="fancy_outline"))