Next's dev server intermittently answers a 500 while recompiling a route: it reads a build manifest mid-write and cannot parse it. A single one of those during the locate poll was fatal, so a blip in the pipeline killed a run partway through an auth flow on Windows. 5xx responses and dropped connections are now a distinct TransientError, retried until the step's own timeout. A 4xx still fails immediately — those are verdicts about the request, not blips. If the errors persist all the way to the timeout, the message says so rather than blaming a missing extension. Error bodies are also summarised. A dev-server 500 replies with a full HTML page, and printing it raw buried the one line that said what went wrong under kilobytes of script tags. This makes the client tolerant of the fault, which is not the same as fixing it: the real answer on an automation host is to run a production build rather than `next dev`, so those manifests are written once instead of continuously. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
391 lines
16 KiB
Python
391 lines
16 KiB
Python
#!/usr/bin/env python3
|
|
"""Automation runner — the process that makes dashboard buttons do something.
|
|
|
|
Leave this running. It polls the dashboard for queued runs, and when one appears
|
|
it works through that automation's steps, driving the real mouse and keyboard,
|
|
reporting each step back so the page can show progress.
|
|
|
|
python runner.py # against http://localhost:3000
|
|
python runner.py --api http://vps:3000
|
|
|
|
Steps come from the server, so adding a new button means editing
|
|
lib/automations.ts — nothing here needs to change.
|
|
|
|
Stop with Ctrl-C. A run in progress can be halted from the dashboard's Stop
|
|
button; the runner notices between steps.
|
|
"""
|
|
|
|
import argparse
|
|
import os
|
|
import random
|
|
import socket
|
|
import sys
|
|
import threading
|
|
import time
|
|
|
|
import actions
|
|
import focus
|
|
from clicker import Dashboard, DashboardError, NotFoundError
|
|
|
|
POLL_SECONDS = 1.0
|
|
HEARTBEAT_SECONDS = 2.0
|
|
|
|
# Bumped whenever the step vocabulary or the locate protocol changes. Reported in
|
|
# the heartbeat so the dashboard can say "restart your runner" instead of letting
|
|
# a stale process fail on a step type it has never heard of.
|
|
VERSION = "0.13.0"
|
|
|
|
# Shared with the heartbeat thread: whether a run is currently executing.
|
|
_busy = threading.Event()
|
|
|
|
|
|
def heartbeat_loop(dash: Dashboard, opts, stop: threading.Event) -> None:
|
|
"""Check in on our own thread.
|
|
|
|
Deliberately not folded into the main poll loop: a single step can block for
|
|
twenty seconds waiting on the extension, and a runner that goes quiet that
|
|
long would show as offline in the middle of the run it is executing.
|
|
"""
|
|
payload = {"host": socket.gethostname(), "pid": os.getpid(),
|
|
"dryRun": opts.dry_run, "version": VERSION}
|
|
while not stop.is_set():
|
|
try:
|
|
dash._request("/api/autobuyer/runner", "POST", {**payload, "busy": _busy.is_set()})
|
|
except DashboardError:
|
|
pass # the main loop reports connectivity; don't double up
|
|
stop.wait(HEARTBEAT_SECONDS)
|
|
|
|
|
|
def run_one_step(dash: Dashboard, step: dict, opts, rng, report, run_id: int | None = None) -> None:
|
|
"""Dispatch a single step. Raises StepError / DashboardError on failure."""
|
|
if step["action"] == "wait":
|
|
time.sleep(float(step.get("seconds", 1)))
|
|
report(f"waited {step.get('seconds', 1)}s")
|
|
|
|
elif step["action"] == "navigate":
|
|
# No mouse involved: the extension points the tab at the page and waits
|
|
# for it to load. Locating <body> confirms it is really there.
|
|
target = step.get("url", "")
|
|
landed = dash.locate("body", 0, step.get("urlPattern", "") or opts.url,
|
|
opts.timeout, step.get("openUrl", "") or "",
|
|
navigate_url=target)
|
|
# Navigating to a protected page is where an expired session shows up:
|
|
# the site answers with a redirect to its login form.
|
|
actions.check_signed_in(landed, step.get("signedOut", "") or "")
|
|
report(f"tab is on {landed.get('url', target)}")
|
|
|
|
elif step["action"] == "waitFor":
|
|
# A gate, not an action: block until the element exists. Nothing is
|
|
# clicked or typed. Whatever has to make it appear — a person solving a
|
|
# challenge, a slow server, a background job — happens outside this run.
|
|
selector = step["selector"]
|
|
timeout_s = float(step.get("timeoutSeconds", 120))
|
|
deadline = time.time() + timeout_s
|
|
announced = False
|
|
|
|
while True:
|
|
# Long gates must still honour the dashboard's Stop button.
|
|
if run_id is not None and run_status(dash, run_id) == "cancelled":
|
|
raise actions.StepError("cancelled while waiting")
|
|
|
|
try:
|
|
dash.locate(selector, int(step.get("index", 0)),
|
|
step.get("urlPattern", "") or opts.url,
|
|
opts.timeout, step.get("openUrl", "") or "")
|
|
report(f"{selector} appeared")
|
|
break
|
|
except DashboardError:
|
|
pass # not there yet, or the tab is mid-render
|
|
|
|
if time.time() >= deadline:
|
|
raise actions.StepError(
|
|
f"{selector} did not appear within {timeout_s:.0f}s"
|
|
)
|
|
if not announced:
|
|
report(f"waiting for {selector} (up to {timeout_s:.0f}s)")
|
|
announced = True
|
|
time.sleep(2.0)
|
|
|
|
elif step["action"] not in ("click", "type"):
|
|
# Almost always a stale runner: the server defines the step vocabulary,
|
|
# so a step type this process has never heard of means automations.ts has
|
|
# moved on and this file has not been restarted.
|
|
raise actions.StepError(
|
|
f"unknown step action {step['action']!r} — this runner is "
|
|
f"v{VERSION}; restart it to pick up newer step types"
|
|
)
|
|
|
|
else:
|
|
try:
|
|
actions.perform(
|
|
dash,
|
|
step["action"],
|
|
step["selector"],
|
|
index=int(step.get("index", 0)),
|
|
url=step.get("urlPattern", "") or opts.url,
|
|
open_url=step.get("openUrl", "") or "",
|
|
signed_out=step.get("signedOut", "") or "",
|
|
text=step.get("text"),
|
|
clear=bool(step.get("clear")),
|
|
scale=opts.scale,
|
|
timeout=opts.timeout,
|
|
rng=rng,
|
|
activate=not opts.no_activate,
|
|
dry_run=opts.dry_run,
|
|
report=report,
|
|
)
|
|
except NotFoundError:
|
|
# Only absence is skippable. A covered element, an off-screen target
|
|
# or an unreachable extension still fails the run.
|
|
if not step.get("skipIfNotFound"):
|
|
raise
|
|
report(f"{step['selector']} not present — skipped")
|
|
|
|
|
|
def run_status(dash: Dashboard, run_id: int) -> str | None:
|
|
"""Current status, so a wait can notice the dashboard's Stop button."""
|
|
try:
|
|
return dash._request(f"/api/autobuyer/runs?id={run_id}")["run"]["status"]
|
|
except (DashboardError, KeyError):
|
|
return None
|
|
|
|
|
|
def wait_for_sign_in(dash: Dashboard, step: dict, run_id: int, opts, deadline: float) -> bool:
|
|
"""Block until the tab leaves the login page.
|
|
|
|
The runner does not log in — entering credentials is the human's job, and a
|
|
keystroke aimed at the wrong field is not something to risk automating. All
|
|
this does is watch the tab and pick the run back up once you are through.
|
|
"""
|
|
signed_out = step.get("signedOut", "") or ""
|
|
pattern = step.get("urlPattern", "") or opts.url
|
|
open_url = step.get("openUrl", "") or ""
|
|
|
|
while time.time() < deadline:
|
|
if run_status(dash, run_id) == "cancelled":
|
|
return False
|
|
time.sleep(3.0)
|
|
try:
|
|
landed = dash.locate("body", 0, pattern, opts.timeout, open_url)
|
|
except DashboardError:
|
|
continue # tab busy mid-redirect; look again shortly
|
|
if signed_out not in (landed.get("url") or ""):
|
|
return True
|
|
|
|
return False
|
|
|
|
|
|
def signed_in_now(dash: Dashboard, step: dict, opts, settle_seconds: float = 20.0) -> bool:
|
|
"""Did we actually get back in?
|
|
|
|
Auth steps can 'succeed' — every click landing — while the site still rejects
|
|
the sign-in, so this confirms rather than assumes before the failed step is
|
|
retried.
|
|
|
|
It polls rather than checking once: submitting a login form starts a network
|
|
round trip and then a redirect, so the tab is still sitting on the login URL
|
|
for a second or two afterwards. Checking immediately reports a failure for a
|
|
sign-in that is merely still in flight.
|
|
"""
|
|
signed_out = step.get("signedOut", "") or ""
|
|
if not signed_out:
|
|
return True
|
|
|
|
deadline = time.time() + settle_seconds
|
|
while True:
|
|
try:
|
|
landed = dash.locate("body", 0, step.get("urlPattern", "") or opts.url,
|
|
opts.timeout, step.get("openUrl", "") or "")
|
|
if signed_out not in (landed.get("url") or ""):
|
|
return True
|
|
except DashboardError:
|
|
pass # mid-redirect the tab can be briefly un-injectable
|
|
|
|
if time.time() >= deadline:
|
|
return False
|
|
time.sleep(1.0)
|
|
|
|
|
|
def run_steps(dash: Dashboard, run: dict, opts) -> None:
|
|
"""Work through one automation. Reports every step; stops on the first
|
|
failure, because a half-completed purchase flow should not barrel on."""
|
|
run_id = run["id"]
|
|
steps = run["steps"]
|
|
auth_steps = run.get("authSteps") or []
|
|
rng = random.Random(opts.seed) if opts.seed is not None else random.Random()
|
|
|
|
print(f"\n▶ run #{run_id} — {run['label']} ({len(steps)} steps)")
|
|
|
|
for i, step in enumerate(steps):
|
|
label = step.get("label") or f"{step['action']} {step.get('selector', '')}".strip()
|
|
print(f" [{i + 1}/{len(steps)}] {label}")
|
|
|
|
detail_parts: list[str] = []
|
|
|
|
def report(message: str) -> None:
|
|
detail_parts.append(message)
|
|
print(f" {message}")
|
|
|
|
# One retry after a sign-in: the session can only be expired once per step.
|
|
for attempt in range(2):
|
|
try:
|
|
run_one_step(dash, step, opts, rng, report, run_id)
|
|
break
|
|
|
|
except actions.SignedOutError as exc:
|
|
wait_seconds = float(step.get("signedOutWait", 0) or 0)
|
|
|
|
# Scripted auth first, if the firm defines any; otherwise fall
|
|
# back to pausing for a human. Only ever attempted once per step —
|
|
# auth that "succeeds" without signing in would otherwise loop.
|
|
if auth_steps and attempt == 0:
|
|
post_progress(dash, run_id, i, f"signed out — running {len(auth_steps)} auth step(s)",
|
|
False, exc.args[0])
|
|
try:
|
|
for astep in auth_steps:
|
|
alabel = astep.get("label") or f"auth: {astep['action']}"
|
|
print(f" [auth] {alabel}")
|
|
run_one_step(dash, astep, opts, rng, report, run_id)
|
|
except (actions.StepError, DashboardError) as aexc:
|
|
msg = f"auth steps failed: {aexc}"
|
|
post_progress(dash, run_id, i, label, False, msg)
|
|
finish(dash, run_id, msg)
|
|
print(f"✗ run #{run_id} stopped: {msg}", file=sys.stderr)
|
|
return
|
|
|
|
if not signed_in_now(dash, step, opts):
|
|
msg = "auth steps ran but the session is still signed out"
|
|
post_progress(dash, run_id, i, label, False, msg)
|
|
finish(dash, run_id, msg)
|
|
print(f"✗ run #{run_id} stopped: {msg}", file=sys.stderr)
|
|
return
|
|
|
|
report("signed in via auth steps — retrying")
|
|
continue
|
|
|
|
if wait_seconds <= 0 or attempt > 0:
|
|
print(f" FAILED: {exc}", file=sys.stderr)
|
|
post_progress(dash, run_id, i, label, False, str(exc))
|
|
finish(dash, run_id, str(exc))
|
|
print(f"✗ run #{run_id} stopped at step {i + 1}")
|
|
return
|
|
|
|
mins = wait_seconds / 60
|
|
note = f"signed out — waiting up to {mins:.0f} min for you to log in"
|
|
print(f" {note}", file=sys.stderr)
|
|
post_progress(dash, run_id, i, note, False, exc.args[0])
|
|
|
|
if not wait_for_sign_in(dash, step, run_id, opts, time.time() + wait_seconds):
|
|
if run_status(dash, run_id) == "cancelled":
|
|
print(f"■ run #{run_id} cancelled while waiting for sign-in")
|
|
return
|
|
msg = f"still signed out after {mins:.0f} min"
|
|
post_progress(dash, run_id, i, label, False, msg)
|
|
finish(dash, run_id, msg)
|
|
print(f"✗ run #{run_id} gave up waiting for sign-in")
|
|
return
|
|
|
|
report("signed in — retrying the step")
|
|
|
|
except (actions.StepError, DashboardError) as exc:
|
|
print(f" FAILED: {exc}", file=sys.stderr)
|
|
post_progress(dash, run_id, i, label, False, str(exc))
|
|
finish(dash, run_id, str(exc))
|
|
print(f"✗ run #{run_id} stopped at step {i + 1}")
|
|
return
|
|
|
|
status = post_progress(dash, run_id, i + 1, label, True, "; ".join(detail_parts) or None)
|
|
|
|
# The dashboard's Stop button shows up here.
|
|
if status == "cancelled":
|
|
print(f"■ run #{run_id} cancelled from the dashboard")
|
|
return
|
|
|
|
finish(dash, run_id, None)
|
|
print(f"✓ run #{run_id} complete")
|
|
|
|
|
|
def post_progress(dash, run_id, step_index, step, ok, detail):
|
|
try:
|
|
res = dash._request("/api/autobuyer/runs/progress", "POST", {
|
|
"id": run_id, "stepIndex": step_index, "step": step, "ok": ok, "detail": detail,
|
|
})
|
|
return res.get("status")
|
|
except DashboardError as exc:
|
|
print(f" (could not report progress: {exc})", file=sys.stderr)
|
|
return None
|
|
|
|
|
|
def finish(dash, run_id, error):
|
|
try:
|
|
dash._request("/api/autobuyer/runs/progress", "POST",
|
|
{"id": run_id, "finish": True, "error": error})
|
|
except DashboardError as exc:
|
|
print(f" (could not report completion: {exc})", file=sys.stderr)
|
|
|
|
|
|
def main() -> int:
|
|
parser = argparse.ArgumentParser(description=__doc__,
|
|
formatter_class=argparse.RawDescriptionHelpFormatter)
|
|
parser.add_argument("--api", default="http://localhost:3000", help="dashboard URL")
|
|
parser.add_argument("--url", default="", help="fallback Chrome match pattern for steps that omit one")
|
|
parser.add_argument("--timeout", type=float, default=45.0, help="seconds to wait for the extension per step (a cold page load takes time)")
|
|
parser.add_argument("--scale", type=float, default=None, help="CSS-to-desktop pixel ratio")
|
|
parser.add_argument("--seed", type=int, default=None, help="seed the motion RNG (debugging)")
|
|
parser.add_argument("--dry-run", action="store_true", help="move the cursor through the steps but never press or type")
|
|
parser.add_argument("--no-activate", action="store_true", help="do not raise the browser before each step")
|
|
opts = parser.parse_args()
|
|
|
|
try:
|
|
import pyautogui # noqa: F401
|
|
except ImportError:
|
|
print("error: pyautogui is not installed — run: pip install -r requirements.txt", file=sys.stderr)
|
|
return 1
|
|
|
|
# Before anything asks the OS how big the screen is. No-op off Windows.
|
|
dpi = focus.enable_dpi_awareness()
|
|
if dpi:
|
|
print(f" {dpi}")
|
|
|
|
dash = Dashboard(opts.api)
|
|
print(f"Runner watching {opts.api}" + (" [DRY RUN — nothing will be pressed]" if opts.dry_run else ""))
|
|
print("Waiting for a run. Press a button on the AutoBuyer page. Ctrl-C to stop.")
|
|
|
|
stop = threading.Event()
|
|
beat = threading.Thread(target=heartbeat_loop, args=(dash, opts, stop), daemon=True)
|
|
beat.start()
|
|
|
|
idle_warned = False
|
|
while True:
|
|
try:
|
|
claim = dash._request("/api/autobuyer/runs/claim", "POST", {})
|
|
except DashboardError as exc:
|
|
if not idle_warned:
|
|
print(f" ({exc})", file=sys.stderr)
|
|
idle_warned = True
|
|
time.sleep(POLL_SECONDS * 3)
|
|
continue
|
|
|
|
idle_warned = False
|
|
run = claim.get("run")
|
|
if not run:
|
|
time.sleep(POLL_SECONDS)
|
|
continue
|
|
|
|
_busy.set()
|
|
try:
|
|
run_steps(dash, run, opts)
|
|
except Exception as exc: # keep the daemon alive
|
|
print(f"✗ run failed unexpectedly: {exc}", file=sys.stderr)
|
|
finish(dash, run["id"], f"runner error: {exc}")
|
|
finally:
|
|
_busy.clear()
|
|
|
|
|
|
if __name__ == "__main__":
|
|
try:
|
|
sys.exit(main())
|
|
except KeyboardInterrupt:
|
|
print("\nstopped.")
|
|
sys.exit(0)
|