Medios-Macina/Provider/vimm.py

"""Vimm provider skeleton (lxml + HTTPClient).

This is a lightweight, resilient provider implementation intended as a
starting point for implementing a full Vimm (vimm.net) provider.

It prefers server-rendered HTML parsing via lxml and uses the repo's
`HTTPClient` helper for robust HTTP calls (timeouts/retries).

Selectors in `search()` are intentionally permissive heuristics; update the
XPaths to match the real site HTML when you have an actual fixture.
"""

from __future__ import annotations

import re
import sys
from typing import Any, Dict, List, Optional
from urllib.parse import urljoin, quote_plus
from lxml import html as lxml_html

from API.HTTP import HTTPClient
from ProviderCore.base import Provider, SearchResult
from SYS.logger import log, debug


class Vimm(Provider):
    """Provider for vimm.net vault listings (skeleton).

    - Uses lxml for parsing
    - No authentication required
    """

    URL = ("https://vimm.net/vault/",)
    URL_DOMAINS = ("vimm.net",)

    def validate(self) -> bool:
        # This provider has no required config; consider more checks if needed.
        return True

    def _parse_size_bytes(self, size_str: str) -> Optional[int]:
        if not size_str:
            return None
        try:
            s = str(size_str or "").strip().replace(",", "")
            m = re.search(r"(?P<val>[\d\.]+)\s*(?P<unit>[KMGT]?B)?", s, flags=re.I)
            if not m:
                return None
            val = float(m.group("val"))
            unit = (m.group("unit") or "B").upper()
            mul = {
                "B": 1,
                "KB": 1024,
                "MB": 1024 ** 2,
                "GB": 1024 ** 3,
                "TB": 1024 ** 4,
            }.get(unit, 1)
            return int(val * mul)
        except Exception:
            return None

    def search(
        self,
        query: str,
        limit: int = 50,
        filters: Optional[Dict[str, Any]] = None,
        **kwargs: Any,
    ) -> List[SearchResult]:
        q = (query or "").strip()
        if not q:
            return []

        # Build search/list URL
        base = "https://vimm.net/vault/"
        url = f"{base}?p=list&q={quote_plus(q)}"

        try:
            with HTTPClient(timeout=20.0) as client:
                resp = client.get(url)
                content = resp.content
        except Exception as exc:
            log(f"[vimm] HTTP fetch failed: {exc}", file=sys.stderr)
            return []

        try:
            doc = lxml_html.fromstring(content)
        except Exception as exc:
            log(f"[vimm] HTML parse failed: {exc}", file=sys.stderr)
            return []

        results: List[SearchResult] = []

        # Candidate XPaths for list items (tweak to match real DOM)
        container_xpaths = [
            '//div[contains(@class,"list-item")]',
            '//div[contains(@class,"result")]',
            '//li[contains(@class,"item")]',
            '//tr[contains(@class,"result")]',
            '//article',
        ]

        nodes = []
        for xp in container_xpaths:
            try:
                found = doc.xpath(xp)
                if found:
                    nodes = found
                    debug(f"[vimm] using xpath {xp} -> {len(found)} nodes")
                    break
            except Exception:
                continue

        # Fallback: try generic anchors under a list area
        if not nodes:
            try:
                nodes = doc.xpath('//div[contains(@id,"list")]/div') or doc.xpath('//div[contains(@class,"results")]/div')
            except Exception:
                nodes = []

        for n in (nodes or [])[: max(1, int(limit))]:
            try:
                # Prefer explicit title anchors
                title = None
                href = None
                try:
                    # a few heuristic searches for a meaningful anchor
                    a = (n.xpath('.//a[contains(@class,"title")]') or
                         n.xpath('.//h2/a') or
                         n.xpath('.//a[contains(@href,"/vault/")]') or
                         n.xpath('.//a'))
                    if a:
                        a0 = a[0]
                        title = a0.text_content().strip()
                        href = a0.get('href')
                except Exception:
                    title = None
                    href = None

                if not title:
                    title = (n.text_content() or "").strip()

                path = urljoin(base, href) if href else ""

                # Extract size & platform heuristics
                size_text = ""
                try:
                    s = n.xpath('.//*[contains(@class,"size")]/text()') or n.xpath('.//span[contains(text(),"MB") or contains(text(),"GB")]/text()')
                    if s:
                        size_text = str(s[0]).strip()
                except Exception:
                    size_text = ""

                size_bytes = self._parse_size_bytes(size_text)

                platform = ""
                try:
                    p = n.xpath('.//*[contains(@class,"platform")]/text()')
                    if p:
                        platform = str(p[0]).strip()
                except Exception:
                    platform = ""

                columns = []
                if platform:
                    columns.append(("Platform", platform))
                if size_text:
                    columns.append(("Size", size_text))

                results.append(
                    SearchResult(
                        table="vimm",
                        title=str(title or "").strip(),
                        path=str(path or ""),
                        detail="",
                        annotations=[],
                        media_kind="file",
                        size_bytes=size_bytes,
                        tag={"vimm"},
                        columns=columns,
                        full_metadata={"raw": lxml_html.tostring(n, encoding="unicode")},
                    )
                )
            except Exception:
                continue

        return results[: max(0, int(limit))]