"""Parser for Pioneer Heated Storage (facility 002691).

The storage-options page uses a tabbed layout (GenerateBlocks tabs).
The "All UNITS" tab panel lists every unit as a card with:
  - size headline  (e.g. "5' x 10'")
  - square / cubic feet
  - short description
  - a "Request" button (no prices published)

Prices are not displayed; the site asks customers to call for
availability.  We extract size and description and leave price as None.
"""

from __future__ import annotations

import re

from bs4 import BeautifulSoup

from src.parsers.base import BaseParser, ParseResult, UnitResult


class Facility002691Parser(BaseParser):
    """Extract storage units from Pioneer Heated Storage."""

    platform = "custom_facility_002691"

    # Matches sizes like  5' x 10'  or  10′ x 20′
    _SIZE_RE = re.compile(
        r"(\d+)\s*['\u2032\u2019]?\s*[xX\u00d7]\s*(\d+)\s*['\u2032\u2019]?"
    )

    def parse(self, html: str, url: str = "") -> ParseResult:
        soup = BeautifulSoup(html, "lxml")
        result = ParseResult(
            platform=self.platform,
            parser_name=self.__class__.__name__,
        )

        # Locate the first tabpanel (the "All UNITS" tab) to avoid
        # counting duplicates from the per-category tabs.
        all_units_panel = soup.find(attrs={"role": "tabpanel"})
        container = all_units_panel if all_units_panel else soup

        # Each unit size lives in a <p> with class gb-headline-text
        seen: set[str] = set()
        for headline in container.find_all(
            "p", class_=lambda c: c and "gb-headline-text" in c
        ):
            text = headline.get_text(strip=True)
            m = self._SIZE_RE.search(text)
            if not m:
                continue

            size_text = f"{m.group(1)}x{m.group(2)}"
            if size_text in seen:
                continue
            seen.add(size_text)

            unit = UnitResult()
            unit.size = text  # preserve original formatting
            w, ln, sq = self.normalize_size(size_text)
            if w is not None:
                unit.metadata = {"width": w, "length": ln, "sqft": sq}

            # Walk sibling paragraphs for sqft / description context
            parent_col = headline.find_parent(
                "div", class_=lambda c: c and "gb-grid-wrapper" in c
            )
            if parent_col:
                paras = parent_col.find_all("p")
                descs = []
                for p in paras:
                    p_text = p.get_text(strip=True)
                    if p_text and p_text != text:
                        descs.append(p_text)
                if descs:
                    unit.description = " | ".join(descs)

            # Price is not published on this site
            unit.price = None

            result.units.append(unit)

        if not result.units:
            result.warnings.append(
                "No units found -- page structure may have changed"
            )

        return result