US2KR/us2kr/weekly.py

"""Monday-morning weekly recap: aggregate the stored week into one briefing.

No US session closes into Monday KST (the US Sunday never trades), so the Monday slot carries a
recap of the previous week's US sessions instead. Everything is built from what the daily pipeline
already stored - the `quotes` table for week-over-week changes and the daily reports for the
narrative - so no collector runs.

The result is a regular ``BriefingResult`` with ``kind="weekly"`` whose payload reuses
``MarketPayload``: ``quotes`` carry weekly changes (close = Friday close, prev_close = the previous
week's close), ``history_notes`` carry the daily briefing headlines, ``session_date`` is the week's
last US session and ``krx_date`` the Monday. That keeps the store, every deliverer and the
dashboard working unchanged.
"""

from __future__ import annotations

import json
import logging
from datetime import date, datetime, timedelta
from zoneinfo import ZoneInfo

from us2kr.config import KST, Settings
from us2kr.llm.base import LLMError
from us2kr.llm.prompts import WEEKLY_SYSTEM_PROMPT, build_weekly_user_message
from us2kr.market_time import MarketClock
from us2kr.models import BriefingResult, MarketPayload, Quote, Report, Verification
from us2kr.pipeline import SKIP_ALREADY_SENT, _deliver, _skip_result
from us2kr.render import render_weekly_markdown
from us2kr.store import Store
from us2kr.summarize import build_provider, summarize_weekly_payload, verify_and_revise
from us2kr.symbols import display_name, group_of

log = logging.getLogger(__name__)

WEEKDAY_KO = ("월", "화", "수", "목", "금", "토", "일")


def _now() -> datetime:
    return datetime.now(ZoneInfo(KST))


def most_recent_monday(today: date) -> date:
    """Today if it is a Monday, otherwise the latest past Monday."""
    return today - timedelta(days=today.weekday())


def week_sessions_for(monday: date, clock: MarketClock) -> list[date]:
    """The previous week's US sessions (Mon..Fri before `monday`), holidays excluded."""
    return [
        d
        for d in (monday - timedelta(days=offset) for offset in range(7, 2, -1))
        if clock.is_us_session(d)
    ]


def _weekly_quote(symbol: str, rows: list) -> Quote:
    """Collapse one symbol's session rows (oldest first) into a week-over-week Quote."""
    first, last = rows[0], rows[-1]
    base = first["prev_close"] if first["prev_close"] is not None else first["close"]
    end = last["close"]
    change_pct = None
    change_abs = None
    if end is not None and base:
        change_pct = (end / base - 1.0) * 100.0
        change_abs = end - base
    highs = [r["day_high"] for r in rows if r["day_high"] is not None]
    lows = [r["day_low"] for r in rows if r["day_low"] is not None]
    return Quote(
        symbol=symbol,
        name=display_name(symbol),
        group=group_of(symbol) or "etc",
        close=end,
        prev_close=base,
        change_pct=change_pct,
        change_abs=change_abs,
        day_high=max(highs) if highs else None,
        day_low=min(lows) if lows else None,
        source="store",
    )


def _daily_headline_notes(store: Store, monday: date, gaps: list[str]) -> list[str]:
    """'미 {세션}({요일}): {헤드라인}' lines from the week's stored daily briefings."""
    notes: list[str] = []
    for row in store.reports_between(monday - timedelta(days=6), monday - timedelta(days=1)):
        try:
            data = json.loads(row["result_json"] or "{}")
        except json.JSONDecodeError:
            gaps.append(f"daily report {row['krx_date']}: unreadable")
            continue
        session = row["session_date"] or data.get("payload", {}).get("session_date") or "?"
        try:
            weekday = WEEKDAY_KO[date.fromisoformat(session).weekday()]
            label = f"미 {session}({weekday})"
        except ValueError:
            label = f"미 {session}"
        report = data.get("report") or {}
        headline = (report.get("headline") or "").strip()
        notes.append(f"{label}: {headline}" if headline else f"{label}: (요약 없음)")
    return notes


def build_weekly_payload(
    monday: date, settings: Settings, clock: MarketClock, store: Store
) -> tuple[MarketPayload, list[date]]:
    """Assemble the weekly MarketPayload for the Monday `monday` from stored data only."""
    sessions = week_sessions_for(monday, clock)
    gaps: list[str] = []

    quotes: list[Quote] = []
    covered: set[str] = set()
    if sessions:
        by_symbol: dict[str, list] = {}
        for row in store.load_quotes_between(sessions[0], sessions[-1]):
            by_symbol.setdefault(row["symbol"], []).append(row)
            covered.add(row["session_date"])
        quotes = [_weekly_quote(symbol, rows) for symbol, rows in sorted(by_symbol.items())]
    missing = [s for s in sessions if s.isoformat() not in covered]
    if not quotes:
        gaps.append("weekly quotes: no stored sessions for last week")
    elif missing:
        gaps.append("weekly quotes: missing sessions " + ", ".join(s.isoformat() for s in missing))

    history_notes = _daily_headline_notes(store, monday, gaps)
    if not history_notes:
        gaps.append("daily headlines: no stored briefings for last week")

    last_session = sessions[-1] if sessions else monday - timedelta(days=3)
    payload = MarketPayload(
        session_date=last_session,
        us_close_time=clock.us_session_close(last_session),
        generated_at=_now(),
        krx_date=monday,
        krx_open_today=clock.is_krx_session(monday),
        quotes=quotes,
        history_notes=history_notes,
        data_gaps=gaps,
    )
    log.info(
        "weekly payload for %s: %d sessions, %d symbols, %d headlines, %d gaps",
        monday,
        len(sessions),
        len(quotes),
        len(history_notes),
        len(gaps),
    )
    return payload, sessions


def _summarize_weekly(
    payload: MarketPayload, sessions: list[date], settings: Settings
) -> tuple[Report | None, str | None, str | None]:
    """Return (report, model_used, error); any failure degrades instead of aborting (pipeline policy)."""
    try:
        provider = build_provider(settings)
        report, model_used = summarize_weekly_payload(payload, sessions, provider)
        return report, model_used, None
    except LLMError as exc:
        log.error("weekly LLM summarisation failed, degrading: %s", exc)
        return None, None, str(exc)
    except Exception as exc:
        log.exception("unexpected weekly LLM failure, degrading")
        return None, None, f"{type(exc).__name__}: {exc}"


def _verify_weekly(
    payload: MarketPayload, sessions: list[date], report: Report, settings: Settings
) -> tuple[Report, Verification | None]:
    try:
        provider = build_provider(settings)
        return verify_and_revise(
            payload,
            report,
            provider,
            data_message=build_weekly_user_message(payload, sessions),
            system_prompt=WEEKLY_SYSTEM_PROMPT,
        )
    except Exception as exc:  # noqa: BLE001 - verification must never block the recap
        log.warning("weekly verification failed, delivering unverified: %s: %s", type(exc).__name__, exc)
        return report, None


def run_weekly(
    monday: date | None,
    settings: Settings,
    *,
    dry_run: bool = False,
    force: bool = False,
    no_llm: bool = False,
) -> BriefingResult:
    """Produce and deliver the weekly recap for ``monday`` (default: the most recent Monday, KST)."""
    clock = MarketClock(settings.run_offset_minutes, extra_krx_holidays=settings.krx_extra_holidays)
    store = Store(settings.db_path)
    try:
        now = _now()
        if monday is None:
            monday = most_recent_monday(now.date())
        if monday.weekday() != 0:
            raise ValueError(f"{monday} is not a Monday - the weekly recap is keyed by Mondays (KST)")
        if monday > now.date():
            raise ValueError(f"{monday} is in the future")

        payload, sessions = build_weekly_payload(monday, settings, clock, store)
        if not dry_run and not force and store.has_report(monday, kind="weekly"):
            log.info("weekly recap for %s already sent - skipping (use --force to resend)", monday)
            return _skip_result(payload, SKIP_ALREADY_SENT)

        report: Report | None = None
        model_used: str | None = None
        error: str | None = None
        verification: Verification | None = None
        degraded = False
        if no_llm:
            degraded = True
            log.info("--no-llm: skipping weekly summarisation")
        else:
            report, model_used, error = _summarize_weekly(payload, sessions, settings)
            degraded = report is None
            if report is not None and settings.verify_enabled:
                report, verification = _verify_weekly(payload, sessions, report, settings)

        markdown = render_weekly_markdown(payload, report, sessions)
        result = BriefingResult(
            payload=payload,
            report=report,
            markdown=markdown,
            model_used=model_used,
            degraded=degraded,
            error=error,
            verification=verification,
            kind="weekly",
        )
        if dry_run:
            print(markdown)
            return result

        report_id = store.save_report(result, kind="weekly")
        log.info("weekly report %d saved for %s (degraded=%s)", report_id, monday, degraded)
        _deliver(result, settings, store, report_id)
        return result
    finally:
        store.close()