"""Resolve + fetch the real OHIP source files from ontario.ca.

The ministry embeds a date in each filename (e.g.
`moh-ohip-fee-schedule-master-text-2026-06-02.txt`) that changes on every
update, so we scrape the landing page to discover the *current* URLs rather
than hard-coding them. Downloads are content-hashed so the refresh job can
skip re-parsing/re-embedding when nothing changed.

Compliance note: these are PUBLIC reference documents (fee schedules), never
patient data. For a strictly air-gapped clinic, set OHIP_LOCAL_* paths and the
downloader is bypassed entirely.
"""
from __future__ import annotations

import hashlib
import logging
import re
from dataclasses import dataclass
from pathlib import Path

import httpx

from .config import settings

logger = logging.getLogger(__name__)

# Link to the fixed-width Physician Fee Schedule Master, "Text format".
_FSM_RE = re.compile(
    r'href="([^"]*fee-schedule-master-text-[^"]*\.txt)"', re.IGNORECASE
)
# Link to the Physician Schedule of Benefits PDF (the big descriptive doc).
_SOB_RE = re.compile(
    r'href="([^"]*moh-schedule-benefit-[^"]*\.pdf)"', re.IGNORECASE
)
# Trailing YYYY-MM-DD in the filename, used to pick the most recent file.
_DATE_RE = re.compile(r"(\d{4}-\d{2}-\d{2})")


def _latest(urls: list[str]) -> str | None:
    """Pick the URL whose embedded YYYY-MM-DD date is newest."""
    if not urls:
        return None

    def key(u: str) -> str:
        m = _DATE_RE.findall(u)
        return m[-1] if m else "0000-00-00"

    return max(urls, key=key)


@dataclass
class SourceFile:
    path: Path
    sha256: str
    changed: bool
    url: str | None = None


def _data_dir() -> Path:
    d = Path(settings.ohip_data_dir)
    d.mkdir(parents=True, exist_ok=True)
    return d


def _sha256(data: bytes) -> str:
    return hashlib.sha256(data).hexdigest()


def _absolutize(url: str) -> str:
    if url.startswith("http"):
        return url
    if url.startswith("/"):
        return f"https://www.ontario.ca{url}"
    return f"https://www.ontario.ca/{url}"


def _discover_urls() -> tuple[str | None, str | None]:
    """Scrape the landing page for the current FSM + SoB URLs."""
    fsm_url, sob_urls = _discover_all()
    return fsm_url, (sob_urls[0] if sob_urls else None)


def _discover_all() -> tuple[str | None, list[str]]:
    """Return the latest FSM URL and ALL SoB PDF URLs (newest first)."""
    with httpx.Client(timeout=30.0, follow_redirects=True) as client:
        html = client.get(settings.ohip_source_page).text
    fsm_url = _latest([_absolutize(u) for u in _FSM_RE.findall(html)])

    def date_key(u: str) -> str:
        m = _DATE_RE.findall(u)
        return m[-1] if m else "0000-00-00"

    sob_urls = sorted(
        {_absolutize(u) for u in _SOB_RE.findall(html)}, key=date_key, reverse=True
    )
    logger.info("Discovered FSM=%s and %d SoB PDF(s)", fsm_url, len(sob_urls))
    return fsm_url, sob_urls


def _write_if_changed(name: str, data: bytes, url: str | None) -> SourceFile:
    dest = _data_dir() / name
    digest = _sha256(data)
    hash_file = _data_dir() / f"{name}.sha256"
    previous = hash_file.read_text().strip() if hash_file.exists() else None
    changed = digest != previous
    if changed:
        dest.write_bytes(data)
        hash_file.write_text(digest)
        logger.info("Updated %s (%d bytes, sha256=%s…)", name, len(data), digest[:12])
    else:
        logger.info("%s unchanged (sha256=%s…)", name, digest[:12])
    return SourceFile(path=dest, sha256=digest, changed=changed, url=url)


def _download(url: str) -> bytes:
    with httpx.Client(timeout=120.0, follow_redirects=True) as client:
        resp = client.get(url)
        resp.raise_for_status()
        return resp.content


def fetch_fsm() -> SourceFile:
    """Return the current FSM text file (downloaded or local override)."""
    if settings.ohip_local_fsm_path:
        data = Path(settings.ohip_local_fsm_path).read_bytes()
        return _write_if_changed("fsm.txt", data, url=None)
    if not settings.allow_network_download:
        raise RuntimeError(
            "Network download disabled and no OHIP_LOCAL_FSM_PATH provided."
        )
    fsm_url, _ = _discover_urls()
    if not fsm_url:
        raise RuntimeError("Could not locate the FSM text URL on the ministry page.")
    return _write_if_changed("fsm.txt", _download(fsm_url), url=fsm_url)


def fetch_sob_pdfs() -> list[SourceFile]:
    """Return ALL published Schedule of Benefits PDFs, newest first.

    Description coverage varies between editions (a code listed with an inline
    description in one year may appear only in a fee matrix the next), so we
    parse every edition and merge — newest wins, older fills gaps. Fees always
    come from the current FSM, so an older description text is still accurate
    for identification/embedding purposes.
    """
    if settings.ohip_local_sob_pdf_path:
        data = Path(settings.ohip_local_sob_pdf_path).read_bytes()
        return [_write_if_changed("sob.pdf", data, url=None)]
    if not settings.allow_network_download:
        logger.warning("Skipping SoB PDFs: network disabled and no local path set.")
        return []
    _, sob_urls = _discover_all()
    if not sob_urls:
        logger.warning("Could not locate any SoB PDF URLs; descriptions will be sparse.")
        return []
    files: list[SourceFile] = []
    for idx, url in enumerate(sob_urls):
        files.append(_write_if_changed(f"sob_{idx}.pdf", _download(url), url=url))
    return files