databank/scripts/run_scheduler.py

"""Run the formal Databank scheduler orchestrating spiders, reporters, and analyzers.

Usage (PowerShell):
  # Ensure deps
  # python -m pip install requests pymongo
  # Configure DB if needed
  # $env:DATABANK_DB_URI = "mongodb://localhost:27017"
  # $env:DATABANK_DB_NAME = "databank"
  python scripts/run_scheduler.py
"""

from __future__ import annotations

import os
from typing import Dict, List

from databank.db import MongoDB
from databank.reporter.daily_file import DailyFileReporter
from databank.spiders.get_league_match_list import GetLeagueMatchListSpider
from databank.spiders.base import BaseSpider
from databank.scheduler.orchestrator import DatabankScheduler, TaskProvider
from databank.tasks.providers import league_matchlist_from_mongo
from databank.analytics.simple_counts import PerTokenCounter


def main() -> None:
    """Entry point that builds and runs the Databank scheduler once."""
    uri = os.getenv("DATABANK_DB_URI", "mongodb://localhost:27017")
    name = os.getenv("DATABANK_DB_NAME", "databank")

    db = MongoDB(
        uri=uri,
        name=name,
        indexes={
            "match": [
                {
                    "keys": [("match.matchId", 1)],
                    "unique": True,
                    "name": "uniq_match_matchId",
                }
            ]
        },
    )
    db.connect()
    db.ensure_indexes()

    # Spiders
    get_match = GetLeagueMatchListSpider()
    spiders: List[BaseSpider] = [get_match]

    # Reporters
    reporters = [DailyFileReporter(timezone="utc+8")]

    # Task providers wiring (no caps in production)
    tasks_provider: Dict[BaseSpider, TaskProvider] = {
        get_match: league_matchlist_from_mongo(),
    }

    # Analyzers
    analyzers = [PerTokenCounter()]

    # Orchestrator
    scheduler = DatabankScheduler(
        db=db,
        spiders=spiders,
        reporters=reporters,
        task_providers=tasks_provider,
        analyzers=analyzers,
        interval_s=None,  # set to seconds to loop
    )

    summary = scheduler.run_once()
    print("Scheduler finished. Total persisted:", summary.total_docs)


if __name__ == "__main__":
    main()