agenda/agenda/data.py

"""Agenda data."""

import asyncio
import collections
import itertools
import os
import typing
from datetime import date, datetime, timedelta

import dateutil.rrule
import dateutil.tz
import flask
import holidays
import isodate  # type: ignore
import lxml
import pytz
import yaml

from . import (
    accommodation,
    birthday,
    calendar,
    conference,
    domains,
    economist,
    fx,
    gwr,
    hn,
    meetup,
    stock_market,
    subscription,
    sun,
    thespacedevs,
    travel,
    uk_holiday,
    uk_tz,
    waste_schedule,
)
from .types import Event, Holiday

StrDict = dict[str, typing.Any]

here = dateutil.tz.tzlocal()

# deadline to file tax return
# credit card expiry dates
# morzine ski lifts
# chalet availablity calendar

# starlink visible


def timezone_transition(
    start: datetime, end: datetime, key: str, tz_name: str
) -> list[Event]:
    """Clocks changes."""
    tz = pytz.timezone(tz_name)
    return [
        Event(name=key, date=pytz.utc.localize(t).astimezone(tz))
        for t in tz._utc_transition_times  # type: ignore
        if start <= t <= end
    ]


def us_holidays(start_date: date, end_date: date) -> list[Holiday]:
    """Get US holidays."""
    found: list[Holiday] = []
    for year in range(start_date.year, end_date.year + 1):
        hols = holidays.country_holidays("US", years=year, language="en")
        found += [
            Holiday(date=hol_date, name=title, country="us")
            for hol_date, title in hols.items()
            if start_date < hol_date < end_date
        ]

    extra = []
    for h in found:
        if h.name != "Thanksgiving":
            continue
        extra += [
            Holiday(date=h.date + timedelta(days=1), name="Black Friday", country="us"),
            Holiday(date=h.date + timedelta(days=4), name="Cyber Monday", country="us"),
        ]

    return found + extra


def get_nyse_holidays(
    start_date: date, end_date: date, us_hols: list[Holiday]
) -> list[Event]:
    """NYSE holidays."""
    known_us_hols = {(h.date, h.name) for h in us_hols}
    found: list[Event] = []
    rename = {"Thanksgiving Day": "Thanksgiving"}
    for year in range(start_date.year, end_date.year + 1):
        hols = holidays.financial_holidays("NYSE", years=year)
        found += [
            Event(
                name="holiday",
                date=hol_date,
                title=rename.get(title, title),
            )
            for hol_date, title in hols.items()
            if start_date < hol_date < end_date
        ]
    found = [hol for hol in found if (hol.date, hol.title) not in known_us_hols]
    for hol in found:
        assert hol.title
        hol.title += " (NYSE)"
    return found


def get_holidays(country: str, start_date: date, end_date: date) -> list[Holiday]:
    """Get holidays."""
    found: list[Holiday] = []
    for year in range(start_date.year, end_date.year + 1):
        hols = holidays.country_holidays(country.upper(), years=year, language="en_US")
        found += [
            Holiday(
                date=hol_date,
                name=title,
                country=country.lower(),
            )
            for hol_date, title in hols.items()
            if start_date < hol_date < end_date
        ]

    return found


def midnight(d: date) -> datetime:
    """Convert from date to midnight on that day."""
    return datetime.combine(d, datetime.min.time())


def dates_from_rrule(
    rrule: str, start: date, end: date
) -> typing.Sequence[datetime | date]:
    """Generate events from an RRULE between start_date and end_date."""
    all_day = not any(param in rrule for param in ["BYHOUR", "BYMINUTE", "BYSECOND"])

    return [
        i.date() if all_day else uk_tz.localize(i)
        for i in dateutil.rrule.rrulestr(rrule, dtstart=midnight(start)).between(
            midnight(start), midnight(end)
        )
    ]


async def waste_collection_events(data_dir: str) -> list[Event]:
    """Waste colllection events."""
    postcode = "BS48 3HG"
    uprn = "24071046"

    html = await waste_schedule.get_html(data_dir, postcode, uprn)
    root = lxml.html.fromstring(html)
    events = waste_schedule.parse(root)
    return events


async def bristol_waste_collection_events(
    data_dir: str, start_date: date
) -> list[Event]:
    """Waste colllection events."""
    uprn = "358335"

    return await waste_schedule.get_bristol_gov_uk(start_date, data_dir, uprn)


def combine_holidays(holidays: list[Holiday]) -> list[Event]:
    """Combine UK and US holidays with the same date and title."""

    all_countries = {h.country for h in holidays}

    standard_name = {
        (1, 1): "New Year's Day",
        (1, 6): "Epiphany",
        (12, 8): "Immaculate conception",
        (12, 25): "Christmas Day",
        (12, 26): "Boxing Day",
        (5, 1): "Labour Day",
    }

    combined: collections.defaultdict[
        tuple[date, str], set[str]
    ] = collections.defaultdict(set)

    for h in holidays:
        assert isinstance(h.name, str) and isinstance(h.date, date)

        event_key = (h.date, standard_name.get((h.date.month, h.date.day), h.name))
        combined[event_key].add(h.country)

    events: list[Event] = []
    for (d, name), countries in combined.items():
        if len(countries) == len(all_countries):
            country_list = ""
        elif len(countries) < len(all_countries) / 2:
            country_list = ", ".join(sorted(country.upper() for country in countries))
        else:
            country_list = "not " + ", ".join(
                sorted(country.upper() for country in all_countries - set(countries))
            )

        e = Event(
            name="holiday",
            date=d,
            title=f"{name} ({country_list})" if country_list else name,
        )
        events.append(e)

    return events


def get_yaml_event_date_field(item: dict[str, str]) -> str:
    """Event date field name."""
    return (
        "end_date"
        if item["name"] == "travel_insurance"
        else ("start_date" if "start_date" in item else "date")
    )


def get_yaml_event_end_date_field(item: dict[str, str]) -> str:
    """Event date field name."""
    return (
        "end_date"
        if item["name"] == "travel_insurance"
        else ("start_date" if "start_date" in item else "date")
    )


def read_events_yaml(data_dir: str, start: date, end: date) -> list[Event]:
    """Read eventes from YAML file."""
    events: list[Event] = []
    for item in yaml.safe_load(open(os.path.join(data_dir, "events.yaml"))):
        duration = (
            isodate.parse_duration(item["duration"]) if "duration" in item else None
        )
        dates = (
            dates_from_rrule(item["rrule"], start, end)
            if "rrule" in item
            else [item[get_yaml_event_date_field(item)]]
        )
        for dt in dates:
            e = Event(
                name=item["name"],
                date=dt,
                end_date=(
                    dt + duration
                    if duration
                    else (
                        item.get("end_date")
                        if item["name"] != "travel_insurance"
                        else None
                    )
                ),
                title=item.get("title"),
                url=item.get("url"),
            )
            events.append(e)
    return events


def find_markets_during_stay(
    accommodation_events: list[Event], markets: list[Event]
) -> list[Event]:
    """Market events that happen during accommodation stays."""
    overlapping_markets = []
    for market in markets:
        for e in accommodation_events:
            # Check if the market date is within the accommodation dates.
            if e.as_date <= market.as_date <= e.end_as_date:
                overlapping_markets.append(market)
                break  # Breaks the inner loop if overlap is found.
    return overlapping_markets


def find_gaps(events: list[Event], min_gap_days: int = 3) -> list[StrDict]:
    """Gaps of at least `min_gap_days` between events in a list of events."""
    # Sort events by start date

    gaps: list[tuple[date, date]] = []
    previous_event_end = None

    by_start_date = {
        d: list(on_day)
        for d, on_day in itertools.groupby(events, key=lambda e: e.as_date)
    }

    by_end_date = {
        d: list(on_day)
        for d, on_day in itertools.groupby(events, key=lambda e: e.end_as_date)
    }

    for event in events:
        # Use start date for current event
        start_date = event.as_date

        # If previous event exists, calculate the gap
        if previous_event_end:
            gap_days = (start_date - previous_event_end).days
            if gap_days >= (min_gap_days + 2):
                start_end = (
                    previous_event_end + timedelta(days=1),
                    start_date - timedelta(days=1),
                )
                gaps.append(start_end)

        # Update previous event end date
        end = event.end_as_date
        if not previous_event_end or end > previous_event_end:
            previous_event_end = end

    return [
        {
            "start": gap_start,
            "end": gap_end,
            "after": by_start_date[gap_end + timedelta(days=1)],
            "before": by_end_date[gap_start - timedelta(days=1)],
        }
        for gap_start, gap_end in gaps
    ]


def busy_event(e: Event) -> bool:
    """Busy."""
    if e.name not in {
        "event",
        "accommodation",
        "conference",
        "dodainville",
        "transport",
        "meetup",
    }:
        return False

    if e.title in ("IA UK board meeting", "Mill Road Winter Fair"):
        return False

    if e.name == "conference" and not e.going:
        return False
    if not e.title:
        return True
    if e.title == "LHG Run Club" or "Third Thursday Social" in e.title:
        return False

    lc_title = e.title.lower()
    return "rebels" not in lc_title and "south west data social" not in lc_title


async def get_data(
    now: datetime, config: flask.config.Config
) -> typing.Mapping[str, str | object]:
    """Get data to display on agenda dashboard."""
    data_dir = config["DATA_DIR"]

    rocket_dir = os.path.join(data_dir, "thespacedevs")
    today = now.date()
    two_weeks_ago = today - timedelta(weeks=2)
    last_week = today - timedelta(weeks=1)
    last_year = today - timedelta(days=365)
    next_year = today + timedelta(days=365)

    minus_365 = now - timedelta(days=365)
    plus_365 = now + timedelta(days=365)

    (
        gbpusd,
        gwr_advance_tickets,
        bank_holiday,
        rockets,
        backwell_bins,
        bristol_bins,
    ) = await asyncio.gather(
        fx.get_gbpusd(config),
        gwr.advance_ticket_date(data_dir),
        uk_holiday.bank_holiday_list(last_year, next_year, data_dir),
        thespacedevs.get_launches(rocket_dir, limit=40),
        waste_collection_events(data_dir),
        bristol_waste_collection_events(data_dir, today),
    )

    reply: dict[str, typing.Any] = {
        "now": now,
        "gbpusd": gbpusd,
        "stock_markets": stock_market.open_and_close(),
        "rockets": rockets,
        "gwr_advance_tickets": gwr_advance_tickets,
    }

    my_data = config["PERSONAL_DATA"]
    events = (
        [
            Event(name="mothers_day", date=uk_holiday.get_mothers_day(today)),
        ]
        + timezone_transition(minus_365, plus_365, "uk_clock_change", "Europe/London")
        + timezone_transition(
            minus_365, plus_365, "us_clock_change", "America/New_York"
        )
    )

    if gwr_advance_tickets:
        events.append(Event(name="gwr_advance_tickets", date=gwr_advance_tickets))

    us_hols = us_holidays(last_year, next_year)

    holidays: list[Holiday] = bank_holiday + us_hols
    for country in (
        "at",
        "be",
        "br",
        "ch",
        "cz",
        "de",
        "dk",
        "ee",
        "es",
        "fi",
        "fr",
        "gr",
        "it",
        "ke",
        "nl",
        "pl",
    ):
        holidays += get_holidays(country, last_year, next_year)

    events += get_nyse_holidays(last_year, next_year, us_hols)

    accommodation_events = accommodation.get_events(
        os.path.join(my_data, "accommodation.yaml")
    )

    events += combine_holidays(holidays)
    events += birthday.get_birthdays(last_year, os.path.join(my_data, "entities.yaml"))
    events += accommodation_events
    events += travel.all_events(my_data)
    events += conference.get_list(os.path.join(my_data, "conferences.yaml"))
    events += backwell_bins + bristol_bins
    events += read_events_yaml(my_data, last_year, next_year)
    events += subscription.get_events(os.path.join(my_data, "subscriptions.yaml"))
    events += economist.publication_dates(last_week, next_year)
    events += meetup.get_events(my_data)
    events += hn.whoishiring(last_year, next_year)

    events += domains.renewal_dates(my_data)

    # hide markets that happen while away
    markets = [e for e in events if e.name == "market"]
    going = [e for e in events if e.going]

    overlapping_markets = find_markets_during_stay(
        accommodation_events + going, markets
    )
    for market in overlapping_markets:
        events.remove(market)

    for launch in rockets:
        dt = None

        if launch["net_precision"] == "Day":
            dt = datetime.strptime(launch["net"], "%Y-%m-%dT00:00:00Z").date()
        elif launch["t0_time"]:
            dt = pytz.utc.localize(
                datetime.strptime(launch["net"], "%Y-%m-%dT%H:%M:%SZ")
            )

        if not dt:
            continue

        rocket_name = f'🚀{launch["rocket"]}: {launch["mission_name"] or "[no mission]"}'
        e = Event(name="rocket", date=dt, title=rocket_name)
        events.append(e)

    events += [Event(name="today", date=today)]

    busy_events = [
        e
        for e in sorted(events, key=lambda e: e.as_date)
        if e.as_date > today
        and e.as_date < (today + timedelta(days=365 * 2))
        and busy_event(e)
    ]

    gaps = find_gaps(busy_events)

    events += [
        Event(name="gap", date=gap["start"], end_date=gap["end"]) for gap in gaps
    ]

    # Sort events by their datetime; the "today" event is prioritised
    # at the top of the list for today. This is achieved by sorting first by
    # the datetime attribute, and then ensuring that events with the name
    # "today" are ordered before others on the same date.
    events.sort(key=lambda e: (e.as_datetime, e.name != "today"))

    reply["gaps"] = gaps

    observer = sun.bristol()
    reply["sunrise"] = sun.sunrise(observer)
    reply["sunset"] = sun.sunset(observer)
    reply["events"] = events
    reply["last_week"] = last_week
    reply["two_weeks_ago"] = two_weeks_ago

    reply["fullcalendar_events"] = calendar.build_events(events)

    return reply