website/scripts/screenshots/capture.py

#!/usr/bin/env python3
"""Take real screenshots of a running netOrk instance for the website.

Normally that instance is the local demo copy from scripts/demo (anonymized
production data), which this script logs into on its own:

    capture.py --list-devices          # prints IDs to pick for --var
    capture.py --var ap=<id> --var server=<id> [--only name ...]

Against a real instance, log in by hand and cover what must not be seen:

    NETORK_URL=https://... capture.py --login
    NETORK_URL=https://... capture.py --mask --var ...

While capturing, every request to the API that is not a GET is aborted, so
taking screenshots cannot change anything on the instance.
"""

import argparse
import json
import os
import re
import sys
import urllib.error
import urllib.parse
import urllib.request
from pathlib import Path

from playwright.sync_api import Page, sync_playwright

from shots import SHOTS

BASE = os.environ.get("NETORK_URL", "http://127.0.0.1:5173").rstrip("/")
LOCAL = re.match(r"https?://(127\.0\.0\.1|localhost)[:/]", BASE + "/") is not None
# The demo instance's admin (see scripts/demo/anonymize.py).
USER = os.environ.get("NETORK_USER", "netork")
PASSWORD = os.environ.get("NETORK_PASSWORD", "netork-demo")
STATE = Path(os.environ.get(
    "NETORK_STATE", Path.home() / ".cache" / "netork-screenshots" / "state.json"))
# One term per line: site names, customer names, domains ... never committed.
MASK_FILE = Path(os.environ.get(
    "NETORK_MASK_FILE", Path.home() / ".config" / "netork-screenshots" / "mask.txt"))
OUT = Path(__file__).resolve().parents[2] / "public" / "screenshots"

VIEWPORT = {"width": 1600, "height": 1000}

# Any IPv4 address that is not RFC 1918, loopback or link-local.
PUBLIC_IPV4 = re.compile(
    r"\b(?!10\.)(?!127\.)(?!169\.254\.)(?!192\.168\.)(?!172\.(?:1[6-9]|2\d|3[01])\.)"
    r"(?:25[0-5]|2[0-4]\d|1?\d?\d)(?:\.(?:25[0-5]|2[0-4]\d|1?\d?\d)){3}\b")
EMAIL = re.compile(r"[\w.+-]+@[\w-]+\.[\w.-]+")


def mask_terms() -> list[str]:
    if not MASK_FILE.exists():
        return []
    return [t.strip() for t in MASK_FILE.read_text().splitlines()
            if t.strip() and not t.startswith("#")]


def login() -> None:
    STATE.parent.mkdir(parents=True, exist_ok=True)
    with sync_playwright() as p:
        browser = p.chromium.launch(headless=False)
        ctx = browser.new_context(ignore_https_errors=True, viewport=VIEWPORT)
        page = ctx.new_page()
        page.goto(f"{BASE}/login")
        print("Log in in the browser window (10 minutes) ...", flush=True)
        page.wait_for_function(
            "() => localStorage.getItem('token') && !location.pathname.startsWith('/login')",
            timeout=600_000)
        ctx.storage_state(path=STATE)
        STATE.chmod(0o600)
        browser.close()
    print(f"Session saved to {STATE}")


def token() -> str:
    if LOCAL:
        body = urllib.parse.urlencode({"username": USER, "password": PASSWORD}).encode()
        try:
            with urllib.request.urlopen(f"{BASE}/api/v1/auth/token", body) as res:
                return json.load(res)["access_token"]
        except urllib.error.URLError as e:
            sys.exit(f"Login as {USER} at {BASE} failed: {e} (is scripts/demo/up.sh start running?)")
    state = json.loads(STATE.read_text())
    for origin in state.get("origins", []):
        for item in origin.get("localStorage", []):
            if item["name"] == "token":
                return item["value"]
    sys.exit("No token in the saved session; run --login first.")


def list_devices() -> None:
    with sync_playwright() as p:
        req = p.request.new_context(
            base_url=BASE, ignore_https_errors=True,
            extra_http_headers={"Authorization": f"Bearer {token()}"})
        res = req.get("/api/v1/devices/")
        if not res.ok:
            sys.exit(f"{res.status}: {res.text()[:200]} (session expired? run --login)")
        for d in res.json():
            print(f"{d.get('id')}  {d.get('driver') or '-':18} "
                  f"{d.get('device_type') or '-':20} {d.get('hostname')}")


def settle(page: Page) -> None:
    """Wait until the page has finished loading its data."""
    try:
        page.wait_for_load_state("networkidle", timeout=15_000)
    except Exception:
        pass  # pages that poll never go fully idle
    try:
        page.wait_for_function(
            "() => !document.querySelector('.animate-spin, .animate-pulse')", timeout=15_000)
    except Exception:
        print("  still loading after 15 s, taking the shot anyway")
    page.wait_for_timeout(800)


def capture(variables: dict[str, str], only: set[str], mask: bool) -> None:
    OUT.mkdir(parents=True, exist_ok=True)
    terms = mask_terms()
    tok = token() if LOCAL else None
    blocked: list[str] = []

    def guard(route):
        if route.request.method in ("GET", "HEAD", "OPTIONS"):
            route.continue_()
        else:
            blocked.append(f"{route.request.method} {route.request.url}")
            route.abort()

    with sync_playwright() as p:
        browser = p.chromium.launch()
        ctx = browser.new_context(
            storage_state=None if LOCAL else STATE, ignore_https_errors=True,
            viewport=VIEWPORT, device_scale_factor=2, color_scheme="dark")
        if tok:
            ctx.add_init_script(f"localStorage.setItem('token', {json.dumps(tok)})")
        ctx.route("**/api/**", guard)
        page = ctx.new_page()
        for shot in SHOTS:
            if only and shot.name not in only:
                continue
            try:
                path = shot.path.format(**variables)
            except KeyError as e:
                print(f"skip {shot.name}: needs --var {e.args[0]}=<id>")
                continue
            print(f"{shot.name}: {path}")
            page.goto(f"{BASE}{path}")
            if page.url.rstrip("/").endswith("/login"):
                sys.exit("Session expired; run --login again.")
            page.wait_for_selector(shot.wait_for, timeout=20_000)
            settle(page)
            # The release notes dialog after an upgrade; dismissing it only
            # writes localStorage in this throwaway browser context.
            got_it = page.get_by_role("button", name="Got it")
            if got_it.is_visible():
                got_it.click()
                page.wait_for_timeout(300)
            masks = [page.locator(s) for s in shot.mask]
            if mask:
                masks += [page.get_by_text(PUBLIC_IPV4), page.get_by_text(EMAIL)]
                masks += [page.get_by_text(t) for t in terms]
            clip = None
            if shot.height:
                clip = {"x": 0, "y": 0, "width": VIEWPORT["width"], "height": shot.height}
            page.screenshot(path=OUT / f"{shot.name}.png", full_page=shot.full_page,
                            clip=clip, mask=masks, mask_color="#334155", animations="disabled")
        browser.close()
    if blocked:
        print("Blocked non-GET requests (nothing was sent):")
        for b in sorted(set(blocked)):
            print(f"  {b}")


def main() -> None:
    ap = argparse.ArgumentParser(description=__doc__,
                                 formatter_class=argparse.RawDescriptionHelpFormatter)
    ap.add_argument("--login", action="store_true", help="log in and save the session")
    ap.add_argument("--list-devices", action="store_true", help="print device IDs")
    ap.add_argument("--var", action="append", default=[], metavar="NAME=VALUE",
                    help="fill a {placeholder} in the shot paths")
    ap.add_argument("--only", nargs="*", default=[], help="only these shot names")
    ap.add_argument("--mask", action="store_true",
                    help="cover public IPs, e-mails and the mask-file terms (real instances)")
    args = ap.parse_args()
    if args.login:
        login()
    elif args.list_devices:
        list_devices()
    else:
        capture(dict(v.split("=", 1) for v in args.var), set(args.only), args.mask)


if __name__ == "__main__":
    main()