Blame for src/agentcore/libcal.py

"""LibCal client + formatting for live opening hours and events.

Rendering emits class names only (lb-hours-*, lb-event*); see rendering.py for
the styling contract.
"""
import asyncio
import html as _html
import json
import logging
import re
import time
from datetime import datetime, timezone, timedelta
from typing import Any, Dict, Optional
from urllib.parse import urlsplit

import httpx

from src.config import get_settings, LIBBEE_VERSION

from src.agentcore.constants import LIBRARY_HOURS_URL
from src.agentcore.utils import _escape
from src.services.http_client import make_client

logger = logging.getLogger(__name__)

_libcal_hours_cache: Dict[str, Any] = {}
_LIBCAL_HOURS_TTL = 60 * 60
_LIBCAL_BASE = "https://kustar.libcal.com"
_LIBCAL_LOC_MAIN = 6386
_LIBCAL_LOC_HABSHAN = 25028
_LIBCAL_IID = 3495
_LIBCAL_EVENTS_CAL = 8398
_UAE_TIMEZONE = timezone(timedelta(hours=4))
_EVENTS_URL = "https://kustar.libcal.com/calendar/events"
_libcal_tokens: Dict[str, Dict[str, Any]] = {
    "hours": {"value": "", "expires_at": 0.0},
    "events": {"value": "", "expires_at": 0.0},
}
_libcal_token_lock = asyncio.Lock()


async def _libcal_access_token(client: httpx.AsyncClient, events: bool = False) -> str:
    """Fetch and briefly cache a LibCal OAuth token without logging secrets."""
    settings = get_settings()
    main_client_id = str(settings.libcal_client_id or "").strip()
    main_client_secret = str(settings.libcal_client_secret or "").strip()
    events_client_id = str(settings.libcal_events_client_id or "").strip()
    events_client_secret = str(settings.libcal_events_client_secret or "").strip()

    if events and bool(events_client_id) != bool(events_client_secret):
        raise RuntimeError(
            "LIBCAL_EVENTS_CLIENT_ID and LIBCAL_EVENTS_CLIENT_SECRET must be configured together"
        )

    if events and events_client_id and events_client_secret:
        client_id = events_client_id
        client_secret = events_client_secret
        cache_key = (
            "hours"
            if (events_client_id, events_client_secret)
            == (main_client_id, main_client_secret)
            else "events"
        )
    else:
        client_id = main_client_id
        client_secret = main_client_secret
        cache_key = "hours"

    if not client_id or not client_secret:
        raise RuntimeError("LibCal credentials are not configured in the HF Space")

    cached = _libcal_tokens[cache_key]
    now = time.time()
    if cached["value"] and now < cached["expires_at"]:
        return str(cached["value"])

    async with _libcal_token_lock:
        cached = _libcal_tokens[cache_key]
        now = time.time()
        if cached["value"] and now < cached["expires_at"]:
            return str(cached["value"])
        response = await client.post(
            f"{_LIBCAL_BASE}/api/1.1/oauth/token",
            data={"client_id": client_id, "client_secret": client_secret, "grant_type": "client_credentials"},
            headers={"Content-Type": "application/x-www-form-urlencoded"},
        )
        response.raise_for_status()
        data = response.json()
        token = str(data.get("access_token") or "")
        if not token:
            raise RuntimeError("LibCal authentication returned no access token")
        cached["value"] = token
        cached["expires_at"] = now + max(5, int(data.get("expires_in") or 3600) - 60)
        return token


async def _libcal_fetch_hours() -> Optional[dict]:
    now = time.time()
    uae_now = datetime.now(_UAE_TIMEZONE)
    date_from = uae_now.strftime("%Y-%m-%d")
    cached_at = _libcal_hours_cache.get("cached_at", 0)
    if (_libcal_hours_cache.get("data") and now - cached_at < _LIBCAL_HOURS_TTL
            and _libcal_hours_cache.get("date_from") == date_from):
        return _libcal_hours_cache["data"]
    date_to   = (uae_now + timedelta(days=6)).strftime("%Y-%m-%d")
    try:
        async with make_client(timeout=12) as client:
            token = await _libcal_access_token(client)

            async def fetch_location(location_id: int) -> dict:
                response = await client.get(
                    f"{_LIBCAL_BASE}/api/1.1/hours/{location_id}",
                    params={"from": date_from, "to": date_to},
                    headers={"Authorization": f"Bearer {token}"},
                )
                response.raise_for_status()
                value = response.json()
                value = value[0] if isinstance(value, list) and value else value
                if not isinstance(value, dict) or not isinstance(value.get("dates"), dict):
                    raise ValueError("LibCal hours response has no date schedule")
                return value

            values = await asyncio.gather(
                fetch_location(_LIBCAL_LOC_MAIN),
                fetch_location(_LIBCAL_LOC_HABSHAN),
                return_exceptions=True,
            )
            data = {}
            for location, value in zip(("main_campus", "habshan"), values):
                if isinstance(value, dict):
                    data[location] = value
                else:
                    logger.warning("LibCal hours unavailable for %s (%s)", location, type(value).__name__)
            if not data:
                return None
            _libcal_hours_cache["data"] = data
            _libcal_hours_cache["cached_at"] = now
            _libcal_hours_cache["date_from"] = date_from
            logger.info("LibCal: hours refreshed directly")
            return data
    except Exception as exc:
        logger.warning("LibCal hours fetch failed: %s", exc)
        return None


def _parse_libcal_time(raw: str) -> str:
    if not raw:
        return raw
    raw = str(raw).strip().replace("–", "-").replace("–", "-")
    raw = re.sub(r"\s+", "", raw)
    raw = re.sub(r"(?i)(am|pm)$", lambda m: " " + m.group(1).upper(), raw)
    return raw


def _format_libcal_day(day_data: dict) -> str:
    if not isinstance(day_data, dict) or not day_data:
        return "See hours page"
    status = str(day_data.get("status", "")).lower().strip()
    note = str(day_data.get("note") or "").strip()

    def with_note(value: str) -> str:
        return f"{value} — {note}" if note else value

    if status in ("closed", "0", "false"):
        return with_note("Closed")
    if status == "24hours":
        return with_note("Open 24 hours")
    # Ramadan and other split schedules may have more than one opening interval.
    for key in ("hours", "times"):
        intervals = day_data.get(key)
        if not isinstance(intervals, list):
            continue
        rendered = []
        for interval in intervals:
            if not isinstance(interval, dict):
                continue
            opens = interval.get("from") or interval.get("opens") or ""
            closes = interval.get("to") or interval.get("closes") or ""
            if opens and closes:
                rendered.append(f"{_parse_libcal_time(opens)} – {_parse_libcal_time(closes)}")
        if rendered:
            return with_note("; ".join(rendered))
    opens = day_data.get("opens") or day_data.get("open") or ""
    closes = day_data.get("closes") or day_data.get("close") or ""
    if opens and closes:
        return with_note(f"{_parse_libcal_time(opens)} – {_parse_libcal_time(closes)}")
    if status == "open":
        return with_note("Open — see hours page for times")
    return with_note("See hours page")


_HOURS_CONTACTS = (
    '<div class="lb-muted lb-hours-contacts">'
    'Main Campus Library · E Building · <a href="tel:+97123124604">+971 2 312 4604</a><br>'
    'Habshan Library · Building 5, SAN Campus · <a href="tel:+97123123160">+971 2 312 3160</a>'
    '</div>'
)


async def _hours_answer() -> str:
    uae_now = datetime.now(_UAE_TIMEZONE)
    today_str = uae_now.strftime("%Y-%m-%d")
    day_names = ["Monday","Tuesday","Wednesday","Thursday","Friday","Saturday","Sunday"]
    today_name = day_names[uae_now.weekday()]
    hours_data = await _libcal_fetch_hours()
    if not hours_data:
        return (
            "Here are the <strong>Khalifa University Library</strong> hours:<br><br>"
            "⚠️ Live hours are currently unavailable. Please check the official schedule:<br>"
            f'<a href="{LIBRARY_HOURS_URL}" target="_blank"><strong>{LIBRARY_HOURS_URL}</strong></a>'
            + _HOURS_CONTACTS
        )
    location_labels = {
        "main_campus": "Main Campus Library",
        "habshan":     "Habshan Library (SAN Campus)",
    }

    def _state(hours_str: str) -> str:
        return "closed" if hours_str.lower().startswith("closed") else "open"

    def _today_line(loc_data: dict) -> str:
        dates = loc_data.get("dates", {})
        today_entry = dates.get(today_str, {})
        if not today_entry:
            return f'<em>See <a href="{LIBRARY_HOURS_URL}" target="_blank">live hours</a> for today</em>'
        hours_str = _format_libcal_day(today_entry)
        return (
            f'<div class="lb-hours-today-line">📅 <strong>Today ({today_name}):</strong> '
            f'<span class="lb-hours-status lb-hours-status--{_state(hours_str)}">{_escape(hours_str)}</span></div>'
        )

    def _week_table(loc_data: dict) -> str:
        from datetime import date as _date
        dates = loc_data.get("dates", {})
        rows = ""
        for date_key in sorted(dates.keys())[:7]:
            try:
                day_label = _date.fromisoformat(date_key).strftime("%a %d %b")
            except ValueError:
                day_label = _escape(str(date_key))
            hours_str = _format_libcal_day(dates[date_key])
            row_class = ' class="lb-hours-today"' if date_key == today_str else ""
            rows += (
                f"<tr{row_class}>"
                f'<td class="lb-hours-day">{day_label}</td>'
                f'<td class="lb-hours-time lb-hours-status--{_state(hours_str)}">{_escape(hours_str)}</td>'
                "</tr>"
            )
        return f'<table class="lb-hours-table"><tbody>{rows}</tbody></table>' if rows else ""

    answer = "Here are the <strong>Khalifa University Library</strong> live hours (UAE time, UTC+4):"
    for key, label in location_labels.items():
        loc = hours_data.get(key)
        answer += f'<div class="lb-hours-location"><div class="lb-hours-heading">📍 {label}</div>'
        if not loc:
            answer += "Live hours unavailable; please check the official schedule below.</div>"
            continue
        answer += _today_line(loc) + _week_table(loc) + "</div>"
    answer += (
        '<div class="lb-note">⚠️ <strong>Hours may change during Ramadan, public holidays, and exam periods.</strong> '
        f'Always confirm at <a href="{LIBRARY_HOURS_URL}" target="_blank">{LIBRARY_HOURS_URL}</a></div>'
        + _HOURS_CONTACTS
    )
    return answer


_libcal_events_cache: Dict[str, Any] = {}


_LIBCAL_EVENTS_TTL = 30 * 60

_TAG_RE = re.compile(r"<[^>]+>")
_BLOCK_TAG_RE = re.compile(r"</?(?:p|br|div|li|h\d)\b[^>]*>", re.IGNORECASE)


def _plain_text(value: str) -> str:
    """LibCal descriptions are rich-text HTML; reduce them to clean prose."""
    text = _TAG_RE.sub("", _BLOCK_TAG_RE.sub(" ", value or ""))
    text = _html.unescape(text).replace("\xa0", " ")
    return re.sub(r"\s+", " ", text).strip()


def _truncate_words(text: str, limit: int = 200) -> str:
    if len(text) <= limit:
        return text
    return text[:limit].rsplit(" ", 1)[0].rstrip(" ,;:.-–") + "…"


def _event_public_url(e: dict) -> str:
    """LibCal 1.1 returns url as {"public": ..., "admin": ...}; never expose admin."""
    url = e.get("url")
    if isinstance(url, dict):
        url = url.get("public") or ""
    url = str(url or "").strip()
    if not url and e.get("id"):
        url = f"{_LIBCAL_BASE}/event/{e['id']}"
    return url


def _event_end(ev: dict) -> Optional[datetime]:
    raw = str(ev.get("end") or ev.get("start") or "")
    try:
        dt = datetime.fromisoformat(raw.replace("Z", "+00:00"))
    except ValueError:
        return None
    if ev.get("allDay"):
        return datetime.combine(dt.date(), datetime.max.time(), _UAE_TIMEZONE)
    return dt.replace(tzinfo=_UAE_TIMEZONE) if dt.tzinfo is None else dt


async def _libcal_fetch_events(days: int = 30, limit: int = 10) -> Optional[list]:
    now = time.time()
    uae_now = datetime.now(_UAE_TIMEZONE)
    date_from = uae_now.strftime("%Y-%m-%d")
    days, limit = max(1, min(int(days), 90)), max(1, min(int(limit), 50))
    cache_key = f"{date_from}_{days}_{limit}"
    cached = _libcal_events_cache.get(cache_key, {})
    if "data" in cached and now - cached.get("cached_at", 0) < _LIBCAL_EVENTS_TTL:
        return cached["data"]
    # Remove expired entries so day-based keys cannot accumulate indefinitely.
    for key in list(_libcal_events_cache):
        if now - _libcal_events_cache[key].get("cached_at", 0) >= _LIBCAL_EVENTS_TTL:
            _libcal_events_cache.pop(key, None)
    try:
        async with make_client(timeout=12) as client:
            token = await _libcal_access_token(client, events=True)
            response = await client.get(
                f"{_LIBCAL_BASE}/api/1.1/events",
                params={
                    "iid": _LIBCAL_IID,
                    "cal_id": _LIBCAL_EVENTS_CAL,
                    "date": uae_now.strftime("%Y-%m-%d"),
                    "days": days,
                    "limit": limit,
                },
                headers={"Authorization": f"Bearer {token}"},
            )
            response.raise_for_status()
            data = response.json()
            events_raw = data if isinstance(data, list) else data.get("events", [])
            def _safe_str(val) -> str:
                if val is None:
                    return ""
                if isinstance(val, str):
                    return val
                if isinstance(val, dict):
                    return str(val.get("name") or val.get("title") or "")
                if isinstance(val, list):
                    return ", ".join(_safe_str(v) for v in val if v)
                return str(val)
            events = []
            for e in events_raw:
                if not isinstance(e, dict):
                    continue
                cats = e.get("category") or e.get("categories") or []
                if isinstance(cats, list):
                    cat_str = ", ".join(_safe_str(c) for c in cats)
                else:
                    cat_str = _safe_str(cats)
                events.append({
                    "id":           e.get("id", ""),
                    "title":        _plain_text(_safe_str(e.get("title") or "Untitled Event")),
                    "start":        _safe_str(e.get("start") or ""),
                    "end":          _safe_str(e.get("end") or ""),
                    "allDay":       _libcal_bool(e.get("allday", e.get("allDay", False))),
                    "location":     _plain_text(_safe_str(e.get("location") or "")),
                    "description":  _plain_text(_safe_str(e.get("description") or e.get("summary") or ""))[:600],
                    "url":          _event_public_url(e),
                    "registration": _libcal_bool(e.get("registration")) or _libcal_bool(e.get("seats")),
                    "campus":       _safe_str(e.get("campus") or ""),
                    "category":     cat_str,
                })
            _libcal_events_cache[cache_key] = {"data": events, "cached_at": now}
            logger.info("LibCal: %d events fetched directly", len(events))
            return events
    except Exception as exc:
        logger.warning("LibCal events fetch failed: %s", exc)
        return None


def _libcal_bool(value: Any) -> bool:
    """API values can be booleans, numbers, or the strings '0'/'false'."""
    if isinstance(value, str):
        return value.strip().lower() not in ("", "0", "false", "no", "none", "null")
    return bool(value)


def _format_event_datetime(start: str, end: str, all_day: bool, with_tz: bool = True) -> str:
    if not start:
        return ""
    suffix = " (UAE, UTC+4)" if with_tz else ""
    try:
        s = datetime.fromisoformat(start.replace("Z", "+00:00"))
        if all_day:
            # All-day events express a calendar date, not an instant to shift.
            return s.strftime("%a %d %b") + " · All day"
        # Naive institution-local times are interpreted as UAE time, never server time.
        s = s.replace(tzinfo=_UAE_TIMEZONE) if s.tzinfo is None else s.astimezone(_UAE_TIMEZONE)
        date_str = s.strftime("%a %d %b")
        time_str = s.strftime("%I:%M %p").lstrip("0")
        if end:
            e = datetime.fromisoformat(end.replace("Z", "+00:00"))
            e = e.replace(tzinfo=_UAE_TIMEZONE) if e.tzinfo is None else e.astimezone(_UAE_TIMEZONE)
            end_str = e.strftime("%I:%M %p").lstrip("0")
            if e.date() != s.date():
                end_str = e.strftime("%a %d %b") + " · " + end_str
            return f"{date_str} · {time_str} – {end_str}{suffix}"
        return f"{date_str} · {time_str}{suffix}"
    except (ValueError, TypeError):
        return "Time unavailable — check the event page"


def _render_event(ev: dict, now: datetime) -> str:
    url = str(ev.get("url") or "")
    try:
        parsed = urlsplit(url)
        if parsed.scheme != "https" or not parsed.netloc:
            url = _EVENTS_URL
    except ValueError:
        url = _EVENTS_URL
    url = _escape(url)
    title = _escape(str(ev.get("title") or "Untitled event"))
    when = _format_event_datetime(
        str(ev.get("start") or ""), str(ev.get("end") or ""), bool(ev.get("allDay")), with_tz=False
    )
    today = ' <span class="lb-badge">Today</span>' if when.startswith(now.strftime("%a %d %b")) else ""
    location = _escape(str(ev.get("location") or ""))
    desc = _escape(_truncate_words(str(ev.get("description") or "")))
    parts = [
        '<div class="lb-event">'
        f'<div class="lb-event-title"><a href="{url}" target="_blank">{title}</a>{today}</div>'
    ]
    if when:
        parts.append(f'<div class="lb-event-meta">🕐 {_escape(when)}</div>')
    if location:
        parts.append(f'<div class="lb-event-meta">📍 {location}</div>')
    if desc:
        parts.append(f'<div class="lb-event-desc">{desc}</div>')
    if ev.get("registration"):
        parts.append(f'<a class="lb-btn" href="{url}" target="_blank">Register →</a>')
    parts.append("</div>")
    return "".join(parts)


async def _events_answer(days: int = 30) -> str:
    events = await _libcal_fetch_events(days=days, limit=15)
    if events is None:
        return (
            "<strong>📅 Khalifa University Library Events</strong><br><br>"
            "Live event data is currently unavailable. Browse all upcoming events here:<br>"
            f'<a href="{_EVENTS_URL}" target="_blank">{_EVENTS_URL}</a>'
        )
    now = datetime.now(_UAE_TIMEZONE)
    # LibCal's date parameter returns the whole first day, including sessions that have ended.
    upcoming = [ev for ev in events if (_event_end(ev) or now) >= now]
    if not upcoming:
        return (
            "<strong>📅 Khalifa University Library Events</strong><br><br>"
            f"No upcoming events in the next {days} days.<br><br>"
            f'Browse the full calendar: <a href="{_EVENTS_URL}" target="_blank">{_EVENTS_URL}</a>'
        )
    items = "".join(_render_event(ev, now) for ev in upcoming[:8])
    return (
        '<div class="lb-events">'
        "<strong>📅 Upcoming library events &amp; workshops</strong>"
        f'<div class="lb-muted">Next {days} days · all times are UAE time (UTC+4)</div>'
        f"{items}"
        f'<a class="lb-events-all" href="{_EVENTS_URL}" target="_blank">View the full events calendar →</a>'
        "</div>"
    )