Extract uMatrix rules from Local Extension Settings directory

extract.py

#!/usr/bin/env python3
"""
Extract uMatrix settings from Chrome's chrome.storage.local LevelDB.
"""
import plyvel
import json
import sys
import os

DB_PATH = sys.argv[1] if len(sys.argv) > 1 else os.path.expanduser(".")

OUTPUT_MAP = {
    "userMatrix":         "umatrix-rules.txt",       # My rules
    "userSettings":       "umatrix-settings.json",   # General settings
    "rawSettings":        "umatrix-raw-settings.json",
    "liveHostsFiles":     "umatrix-hosts-meta.json",
    "selectedHostsFiles": "umatrix-selected-hosts.json",
    "externalHostsFiles": "umatrix-external-hosts.json",
    "userHosts":          "umatrix-user-hosts.txt",
    "userRecipes":        "umatrix-user-recipes.txt",
}

def decode_value(raw_bytes):
    try:
        return json.loads(raw_bytes)
    except (json.JSONDecodeError, UnicodeDecodeError):
        return raw_bytes.decode("utf-8", errors="ignore")

def clean_text(data):
    if isinstance(data, str):
        text = data.replace("\\n", "\n")
        text = text.replace("\\t", "\t")
        text = text.replace("\\r", "")
        text = text.replace("\\\\", "\\")
        text = text.replace("\x00", "").replace("\r", "")
        lines = [line.rstrip() for line in text.split("\n")]
        return "\n".join(lines)
    elif isinstance(data, list):
        return "\n".join(str(x) for x in data)
    elif isinstance(data, dict):
        return json.dumps(data, indent=2)
    else:
        return str(data)

def main():
    if not os.path.isdir(DB_PATH):
        print(f"Error: '{DB_PATH}' is not a directory.")
        sys.exit(1)

    db = plyvel.DB(DB_PATH, create_if_missing=False)
    found = set()

    for key, value in db:
        key_str = key.decode("utf-8", errors="ignore")
        data = decode_value(value)

        if key_str in OUTPUT_MAP:
            out_file = OUTPUT_MAP[key_str]
            cleaned = clean_text(data)
            with open(out_file, "w", encoding="utf-8") as f:
                f.write(cleaned)
            lines = cleaned.count("\n") if "\n" in cleaned else 0
            print(f"  [+] {key_str:25s} -> {out_file}  ({lines} lines)")
            found.add(key_str)
        else:
            preview = str(data)[:120].replace("\n", " ")
            print(f"  [-] {key_str:25s} : {preview}...")

    db.close()
    print()
    if "userMatrix" in found:
        print("Import into uMatrix:")
        print("  Dashboard -> My rules -> Import from file -> select umatrix-rules.txt")
    else:
        print("No userMatrix found. Check the [-] lines for actual key names.")

if __name__ == "__main__":
    main()