"""Apple Machine Learning Research — RSS feed → `paper` entities (title, date, abstract, URL) + ANNOUNCEMENT events.

Each research page is followed deterministically (no LLM) to pick up the arXiv / PDF links and the author list when present.
"""
from __future__ import annotations

import re

from aiatlas.registry import org_ref
from aiatlas.sdk.connector import BaseConnector, Parsed, RunContext
from aiatlas.sdk.extract.html import node_text
from aiatlas.sdk.facts import EntityRef, Facts, Target
from aiatlas.sdk.fetch import FetchResult

from ._common import announcement_events

FEED = "https://machinelearning.apple.com/rss.xml"
ARXIV = re.compile(r"arxiv\.org/(?:abs|pdf)/(\d{4}\.\d{4,5})(?:v\d+)?")


class AppleMLConnector(BaseConnector):
    name = "apple_ml"
    label = "Apple Machine Learning Research — feed"
    description = "Apple Machine Learning Research RSS: research items as paper entities (+ arXiv links from the research pages) and events."
    source_key = "machinelearning.apple.com"
    version = "1"
    parser_version = "1"
    interval_seconds = 21600
    min_interval_seconds = 7200
    rate_per_min = 10
    tier = 1
    priority = 1
    expected_min_records = 5
    concurrency = 2

    async def discover(self, ctx: RunContext) -> list[Target]:
        return [Target(url=FEED, doc_type="feed", key="feed", min_bytes=500)]

    async def extract(self, ctx: RunContext, target: Target, res: FetchResult, parsed: Parsed) -> Facts:
        facts = Facts()
        org = org_ref("apple")
        facts.entities.append(org)
        if parsed.kind == "feed":
            for it in parsed.feed_items:
                if not it.url or not it.title:
                    continue
                paper = facts.entity("paper", it.title, identifiers={"url": it.url.rstrip("/")}, organization=org)
                facts.claim(paper, "published_at", it.published_at.date().isoformat() if it.published_at else None)
                facts.claim(paper, "abstract", (it.summary or "")[:2000] or None)
                facts.claim(paper, "authors", it.authors or None)
                facts.claim(paper, "paper_url", it.url)
                facts.relate(paper, "published_by", org)
                facts.follow(it.url, doc_type="paper_page", entity=paper, key=f"paper:{it.url.rstrip('/').rsplit('/', 1)[-1]}", priority=2, min_bytes=1000)
            announcement_events(facts, org, parsed.feed_items, source_name="machinelearning.apple.com", follow=False, importance_default=1)
            facts.document_title, facts.document_entity = "Apple Machine Learning Research", org
        elif target.doc_type == "paper_page" and parsed.html:
            url = (res.final_url or res.url).rstrip("/")
            title = next((t for lvl, t in parsed.html.headings if lvl == 1), None) or (parsed.html.title or "").split(" - Apple")[0].strip()
            paper = target.entity or (EntityRef(entity_type="paper", name=title, identifiers={"url": url}, organization=org) if title else None)
            if paper is not None:
                self._paper_page(facts, org, paper, parsed)
        return facts

    def _paper_page(self, facts: Facts, org: EntityRef, paper: EntityRef, parsed: Parsed) -> None:
        html = parsed.html
        assert html
        if paper not in facts.entities:
            facts.entities.append(paper)
            facts.relate(paper, "published_by", org)
        facts.document_entity = paper
        facts.document_title = html.title
        for href, _text in html.links:
            m = ARXIV.search(href)
            if m:
                paper.identifiers.setdefault("arxiv", m.group(1))
                facts.claim(paper, "arxiv_id", m.group(1))
                facts.claim(paper, "pdf_url", f"https://arxiv.org/pdf/{m.group(1)}")
                break
        else:
            pdf = next((h for h, _t in html.links if h.lower().endswith(".pdf")), None)
            facts.claim(paper, "pdf_url", pdf)
        # <p><span class="a11y">Authors</span>Name‡, Name†**, …</p> — footnote marks (affiliations) stripped
        authors: list[str] = []
        for p in html.css("p"):
            span = p.css_first("span.a11y")
            if span is None or node_text(span).lower() != "authors":
                continue
            text = node_text(p)[len(node_text(span)):]
            for raw in text.split(","):
                name_ = re.sub(r"[‡†§*¶#0-9]+", "", raw).strip()
                if 3 <= len(name_) <= 60 and name_ not in authors:
                    authors.append(name_)
            break
        if authors:
            facts.claim(paper, "authors", authors[:40])
        # publication date comes from the feed (day precision); the page only states the month



CONNECTORS = [AppleMLConnector]