databank/scripts/reporter_demo.py

"""Reporter demo script.

This script demonstrates how to use DailyFileReporter with UTC+8 timezone
and year/month archived log directories. It simulates a short run with
start/success/error/summary notifications.

Usage (PowerShell):
  pwsh -NoProfile -File ./scripts/reporter_demo.py
or
  python ./scripts/reporter_demo.py
"""

from __future__ import annotations

import random
import time
from pathlib import Path

from databank.reporter.daily_file import DailyFileReporter
from databank.core.models import RunSummary


def main() -> None:
    """Run the reporter demo."""
    rep = DailyFileReporter(log_dir=str(Path.cwd() / "logs"), timezone="utc+8")

    # Simulate a run
    spider_name = "demoSpider"
    seeds = ["tokenA", "tokenB", "tokenC"]
    rep.notify_start(spider_name, seeds)

    # simulate successes
    docs = 0
    for _ in range(2):
        cnt = random.randint(1, 5)
        docs += cnt
        rep.notify_success(spider_name, cnt)
        time.sleep(0.2)

    # simulate an error
    rep.notify_error(spider_name, "Simulated network timeout")

    # Final summary
    summary = RunSummary()
    summary.total_docs = docs
    summary.per_spider[spider_name] = docs
    rep.notify_summary(summary)

    print("Reporter demo finished.")
    print("Logs are written under:", Path.cwd() / "logs")


if __name__ == "__main__":
    main()