path: root/jeeves/news.py

"""
Sopel plugin: jeeves news and weather

Weather (via wttr.in - no API key required):
  !weather <city>         - current conditions
  !forecast <city>        - 3-day forecast

News (via RSS feeds):
  !headlines [label] [n]  - latest headlines (label: global / mena / egypt)
  !news                   - show auto-broadcast state for this channel
  !news on|off            - enable/disable auto-broadcast (op/owner)

Broadcast interval and count are configured globally/per-channel via !set:
  !set news_interval <min>    - minutes between broadcasts (default 90)
  !set news_count <n>         - headlines per broadcast cycle (default 1)
"""

import os as _os, sys as _sys
_d = _os.path.dirname(_os.path.abspath(__file__))
if _d not in _sys.path:
    _sys.path.insert(0, _d)
import jv_core as jv

import json
import logging
import time
import urllib.error
import urllib.parse
import urllib.request
import xml.etree.ElementTree as ET

from sopel import plugin

log = logging.getLogger(__name__)

_MAX_COUNT = 10


def setup(bot):
    jv.ensure_setup(bot)


def _fetch_url(url, timeout=10):
    req = urllib.request.Request(url, headers={"User-Agent": "jeeves-irc-bot/1.0"})
    with urllib.request.urlopen(req, timeout=timeout) as resp:
        return resp.read().decode("utf-8", errors="replace")


def _city_param(city):
    return urllib.parse.quote(city.replace(" ", "+"))


def _clean_url(url):
    try:
        p = urllib.parse.urlparse(url)
        return urllib.parse.urlunparse(p._replace(query="", fragment=""))
    except Exception:
        return url


_COUNTRY_CODES = {
    "Afghanistan": "AF", "Albania": "AL", "Algeria": "DZ", "Argentina": "AR",
    "Australia": "AU", "Austria": "AT", "Bahrain": "BH", "Bangladesh": "BD",
    "Belgium": "BE", "Brazil": "BR", "Canada": "CA", "Chile": "CL",
    "China": "CN", "Colombia": "CO", "Croatia": "HR", "Czech Republic": "CZ",
    "Denmark": "DK", "Egypt": "EG", "Ethiopia": "ET", "Finland": "FI",
    "France": "FR", "Germany": "DE", "Ghana": "GH", "Greece": "GR",
    "Hungary": "HU", "India": "IN", "Indonesia": "ID", "Iran": "IR",
    "Iraq": "IQ", "Ireland": "IE", "Israel": "IL", "Italy": "IT",
    "Japan": "JP", "Jordan": "JO", "Kenya": "KE", "Kuwait": "KW",
    "Lebanon": "LB", "Libya": "LY", "Malaysia": "MY", "Mexico": "MX",
    "Morocco": "MA", "Netherlands": "NL", "New Zealand": "NZ", "Nigeria": "NG",
    "Norway": "NO", "Oman": "OM", "Pakistan": "PK", "Palestine": "PS",
    "Peru": "PE", "Philippines": "PH", "Poland": "PL", "Portugal": "PT",
    "Qatar": "QA", "Romania": "RO", "Russia": "RU", "Saudi Arabia": "SA",
    "Senegal": "SN", "Serbia": "RS", "Singapore": "SG", "South Africa": "ZA",
    "South Korea": "KR", "Spain": "ES", "Sudan": "SD", "Sweden": "SE",
    "Switzerland": "CH", "Syria": "SY", "Thailand": "TH", "Tunisia": "TN",
    "Turkey": "TR", "Ukraine": "UA", "United Arab Emirates": "AE",
    "United Kingdom": "GB", "United States of America": "US",
    "United States": "US", "Venezuela": "VE", "Vietnam": "VN", "Yemen": "YE",
}


def _format_location(city_name, region, country):
    code = _COUNTRY_CODES.get(country, country)
    if region and region.lower() not in city_name.lower():
        return f"{city_name}, {region}, {code}"
    return f"{city_name}, {code}"


def _fetch_weather_json(city):
    url = f"https://wttr.in/{_city_param(city)}?format=j1"
    try:
        raw = _fetch_url(url)
    except urllib.error.HTTPError as exc:
        if exc.code in (404, 500):
            return None, f"city not found: '{city}'"
        return None, f"weather service error ({exc.code})"
    except Exception as exc:
        return None, f"weather lookup failed: {exc}"
    try:
        return json.loads(raw), None
    except Exception:
        return None, f"city not found: '{city}'"


@plugin.command("weather")
def cmd_weather(bot, trigger):
    city = (trigger.group(2) or "").strip()
    if not city:
        bot.say("usage: !weather <city>", trigger.sender)
        return
    data, err = _fetch_weather_json(city)
    if err:
        bot.say(err, trigger.sender)
        return
    try:
        area      = data["nearest_area"][0]
        city_name = area["areaName"][0]["value"]
        region    = area["region"][0]["value"]
        country   = area["country"][0]["value"]
        cur       = data["current_condition"][0]
        desc      = cur["weatherDesc"][0]["value"]
        temp_c    = cur["temp_C"]
        feels_c   = cur["FeelsLikeC"]
        humidity  = cur["humidity"]
        location  = _format_location(city_name, region, country)
        bot.say(
            f"[Weather] {location}: {desc}, {temp_c}°C (feels like {feels_c}°C), humidity {humidity}%",
            trigger.sender,
        )
    except (KeyError, IndexError):
        bot.say(f"city not found: '{city}'", trigger.sender)


@plugin.command("forecast")
def cmd_forecast(bot, trigger):
    city = (trigger.group(2) or "").strip()
    if not city:
        bot.say("usage: !forecast <city>", trigger.sender)
        return
    data, err = _fetch_weather_json(city)
    if err:
        bot.say(err, trigger.sender)
        return
    try:
        area      = data["nearest_area"][0]
        city_name = area["areaName"][0]["value"]
        region    = area["region"][0]["value"]
        country   = area["country"][0]["value"]
        location  = _format_location(city_name, region, country)
        days      = data["weather"]
        labels    = ["Today", "Tomorrow", "Day after"]
        parts     = []
        for day, label in zip(days[:3], labels):
            desc  = day["hourly"][4]["weatherDesc"][0]["value"]
            max_c = day["maxtempC"]
            min_c = day["mintempC"]
            parts.append(f"{label}: {desc} {min_c}–{max_c}°C")
        bot.say(f"[Forecast] {location}: {' | '.join(parts)}", trigger.sender)
    except (KeyError, IndexError):
        bot.say(f"city not found: '{city}'", trigger.sender)


# --- RSS ---

def _fetch_feed(url):
    raw  = _fetch_url(url, timeout=15)
    root = ET.fromstring(raw)
    items = []

    for item in root.iter("item"):
        title = (item.findtext("title") or "").strip()
        link  = _clean_url((item.findtext("link") or "").strip())
        guid  = (item.findtext("guid") or link).strip()
        if title and guid:
            items.append({"title": title, "link": link, "guid": guid})

    if not items:
        ns = {"a": "http://www.w3.org/2005/Atom"}
        for entry in root.findall("a:entry", ns):
            title_el = entry.find("a:title", ns)
            link_el  = entry.find("a:link", ns)
            id_el    = entry.find("a:id", ns)
            title = (title_el.text or "").strip() if title_el is not None else ""
            link  = _clean_url((link_el.get("href") or "").strip()) if link_el is not None else ""
            guid  = (id_el.text or "").strip() if id_el is not None else link
            if title and guid:
                items.append({"title": title, "link": link, "guid": guid})

    return items


def _fresh_headlines(data, label=None, max_items=3):
    seen  = set(data.get("news_seen", []))
    feeds = [f for f in data.get("news_feeds", []) if f.get("enabled", True)]
    if label:
        feeds = [f for f in feeds if f.get("label", "").lower() == label.lower()]

    results = []
    for feed in feeds:
        if len(results) >= max_items:
            break
        try:
            for item in _fetch_feed(feed["url"]):
                if item["guid"] not in seen and len(results) < max_items:
                    results.append({"feed": feed["name"], **item})
        except Exception as exc:
            log.warning("jeeves/news: failed to fetch %s: %s", feed.get("id", "?"), exc)

    return results


def _mark_seen(data, items):
    for item in items:
        if item["guid"] not in data["news_seen"]:
            data["news_seen"].append(item["guid"])


@plugin.command("headlines")
def cmd_headlines(bot, trigger):
    args  = (trigger.group(2) or "").strip().split()
    label = None
    count = jv.cfg_val(bot, "news_count")

    for arg in args:
        if arg.lower() in ("global", "mena", "egypt"):
            label = arg.lower()
        else:
            try:
                count = min(int(arg), _MAX_COUNT)
                if count < 1:
                    raise ValueError
            except ValueError:
                bot.say("usage: !headlines [global|mena|egypt] [count]", trigger.sender)
                return

    with bot.memory["jv_lock"]:
        data  = bot.memory["jv_db"]
        feeds = list(data.get("news_feeds", []))
        seen  = set(data.get("news_seen", []))

    items = _fresh_headlines({"news_feeds": feeds, "news_seen": list(seen)}, label=label, max_items=count)
    if not items:
        bot.say("no fresh headlines available right now", trigger.sender)
        return

    with bot.memory["jv_lock"]:
        _mark_seen(bot.memory["jv_db"], items)
        jv.save(bot)

    for item in items:
        bot.say(f"[{item['feed']}] {item['title']} | {item['link']}", trigger.sender)


@plugin.command("news")
def cmd_news(bot, trigger):
    args    = (trigger.group(2) or "").strip().split()
    arg     = args[0].lower() if args else ""
    in_chan = str(trigger.sender).startswith("#")
    channel = str(trigger.sender) if in_chan else None

    if arg in ("on", "off"):
        if not jv.is_authorized(bot, trigger):
            return
        if not in_chan:
            bot.notice("use !news on|off in a channel", trigger.nick)
            return
        with bot.memory["jv_lock"]:
            jv.chan_db(bot, channel)["news"]["enabled"] = (arg == "on")
            jv.save(bot)
        bot.say(f"auto-broadcast turned {arg} for {channel}", channel)
        return

    if in_chan:
        ch       = jv.chan_db(bot, channel)
        enabled  = ch["news"]["enabled"]
        last     = ch["news"]["last_broadcast"]
        interval = jv.cfg_val(bot, "news_interval", channel)
        count    = jv.cfg_val(bot, "news_count",    channel)
        if last:
            ago      = int(time.time() - last)
            last_str = f"{ago // 60}m ago" if ago >= 60 else f"{ago}s ago"
        else:
            last_str = "never"
        state = "on" if enabled else "off"
        bot.say(
            f"auto-broadcast: {state} | every {interval} min | {count} headline(s)/cycle | last: {last_str}",
            channel,
        )
    else:
        bot.notice("usage: !news [on|off]  (use in a channel)", trigger.nick)


@plugin.interval(60)
def _broadcast_tick(bot):
    if "jv_db" not in bot.memory:
        return
    now = time.time()

    with bot.memory["jv_lock"]:
        data    = bot.memory["jv_db"]
        feeds   = list(data.get("news_feeds", []))
        seen    = set(data.get("news_seen", []))
        pending = []
        for chan_name in list(bot.channels):
            channel = str(chan_name)
            ch = jv.chan_db(bot, channel)
            if not ch["news"]["enabled"]:
                continue
            interval_sec = jv.cfg_val(bot, "news_interval", channel) * 60
            if now - ch["news"]["last_broadcast"] < interval_sec:
                continue
            pending.append((channel, jv.cfg_val(bot, "news_count", channel)))

    for channel, b_count in pending:
        active_feeds = [f for f in feeds if f.get("enabled", True)]
        results = []
        for feed in active_feeds:
            if len(results) >= b_count:
                break
            try:
                for item in _fetch_feed(feed["url"]):
                    if item["guid"] not in seen and len(results) < b_count:
                        results.append({"feed": feed["name"], **item})
            except Exception as exc:
                log.warning("jeeves/news tick: %s: %s", feed.get("id", "?"), exc)

        if not results:
            continue

        with bot.memory["jv_lock"]:
            _mark_seen(bot.memory["jv_db"], results)
            jv.chan_db(bot, channel)["news"]["last_broadcast"] = now
            jv.save(bot)
            seen.update(item["guid"] for item in results)

        for item in results:
            bot.say(f"[{item['feed']}] {item['title']} | {item['link']}", channel)