gimme-job/gimme_job/adapters/indeed.py

"""Indeed job site adapter."""
from __future__ import annotations

import time
from typing import TYPE_CHECKING

from loguru import logger

from gimme_job.adapters.base import ManifestDrivenAdapter
from gimme_job.adapters.registry import register
from gimme_job.models.dto import JobPostingCandidate, RawJobCard, SearchQuery

if TYPE_CHECKING:
    from playwright.sync_api import Page

    from gimme_job.models.manifest import SiteManifest


@register("indeed")
class IndeedAdapter(ManifestDrivenAdapter):
    """Indeed-specific adapter with custom card extraction and URL normalization."""

    def prepare(self, page: "Page", config: "SiteManifest") -> None:
        """Navigate to Indeed search URL with fromage=1 for last 24h."""
        url = config.build_start_url(
            type("Q", (), {
                "keywords_urlencoded": page.__class__.__name__,  # placeholder
            })()
        )
        # Actually build URL using the query stored in the manifest
        pass  # URL is built in orchestrator before adapter.prepare() is called

    def apply_search(self, page: "Page", query: SearchQuery) -> None:
        """Indeed uses URL params — no form input needed."""
        pass

    def apply_filters(self, page: "Page", query: SearchQuery) -> None:
        """Indeed date filter is already in the URL (fromage=1). Nothing to click."""
        pass

    def collect_cards(self, page: "Page", config: "SiteManifest") -> list[RawJobCard]:
        """Extract Indeed job cards."""
        cards = []

        # Try container selectors
        container_selectors = config.extract.container_selectors or [
            "div.job_seen_beacon",
            "div[data-jk]",
            "td.resultContent",
        ]

        container_elements = []
        for selector in container_selectors:
            try:
                elements = page.query_selector_all(selector)
                if elements:
                    container_elements = elements
                    logger.debug(f"[indeed] Container: '{selector}' ({len(elements)} items)")
                    break
            except Exception:
                continue

        if not container_elements:
            logger.warning("[indeed] No job cards found")
            return cards

        for el in container_elements:
            try:
                card = self._extract_card(el)
                if card and card.title:
                    cards.append(card)
            except Exception as e:
                logger.debug(f"[indeed] Card extraction error: {e}")

        logger.info(f"[indeed] Extracted {len(cards)} cards")
        return cards

    def _extract_card(self, el) -> RawJobCard | None:
        def get_text(*selectors: str) -> str | None:
            for sel in selectors:
                try:
                    child = el.query_selector(sel)
                    if child:
                        text = child.inner_text()
                        if text and text.strip():
                            return text.strip()
                except Exception:
                    continue
            return None

        def get_attr(selector: str, attr: str) -> str | None:
            try:
                child = el.query_selector(selector)
                if child:
                    val = child.get_attribute(attr)
                    return val.strip() if val else None
            except Exception:
                return None

        title = get_text(
            "h2.jobTitle a span[id^='jobTitle']",
            "h2.jobTitle span",
            "h2 a span",
            "h2[class*='jobTitle'] a",
            "h2",
        )
        if not title:
            return None

        company = get_text(
            "span[data-testid='company-name']",
            ".company",
            "span.companyName",
        )
        location = get_text(
            "div[data-testid='text-location']",
            ".companyLocation",
        )
        posted_text = get_text(
            "span[data-testid='myJobsStateDate']",
            "span.date",
        )
        salary = get_text(
            "div.salary-snippet-container",
            "div[data-testid='attribute_snippet_testid']",
        )

        # URL: prefer the job card link
        url = get_attr("h2.jobTitle a", "href") or get_attr("a[data-jk]", "href")
        if url and url.startswith("/"):
            url = "https://www.indeed.com" + url
        if url:
            url = self._clean_indeed_url(url)

        raw_text = None
        try:
            raw_text = el.inner_text()
        except Exception:
            pass

        return RawJobCard(
            title=title,
            company=company,
            location=location,
            posted_text=posted_text,
            url=url,
            salary_text=salary,
            raw_text=raw_text,
        )

    def _clean_indeed_url(self, url: str) -> str:
        """Strip Indeed tracking parameters, keep the /viewjob?jk= part."""
        from urllib.parse import parse_qs, urlencode, urlparse, urlunparse
        try:
            parsed = urlparse(url)
            if parsed.path == "/rc/clk" or "/clk" in parsed.path:
                # Extract jk parameter and build clean URL
                params = parse_qs(parsed.query)
                jk = params.get("jk", [None])[0]
                if jk:
                    return f"https://www.indeed.com/viewjob?jk={jk}"
            return url
        except Exception:
            return url

    def normalize(self, raw: RawJobCard) -> JobPostingCandidate:
        return super().normalize(raw)