From b457599fd4577a65dc1fba61288574e1ba3c6204 Mon Sep 17 00:00:00 2001
From: CREDO23 <bakerathierry@gmail.com>
Date: Tue, 14 Jul 2026 21:49:23 +0200
Subject: [PATCH] feat: indeed.scrape executor

---
 .../capabilities/indeed/scrape/executor.py    | 55 +++++++++++++++++++
 1 file changed, 55 insertions(+)
 create mode 100644 surfsense_backend/app/capabilities/indeed/scrape/executor.py

diff --git a/surfsense_backend/app/capabilities/indeed/scrape/executor.py b/surfsense_backend/app/capabilities/indeed/scrape/executor.py
new file mode 100644
index 000000000..60df11972
--- /dev/null
+++ b/surfsense_backend/app/capabilities/indeed/scrape/executor.py
@@ -0,0 +1,55 @@
+"""``indeed.scrape`` executor: verb input → scraper → Indeed job items."""
+
+from __future__ import annotations
+
+from collections.abc import Awaitable, Callable
+
+from app.capabilities.core import Executor
+from app.capabilities.core.progress import emit_progress
+from app.capabilities.indeed.scrape.schemas import ScrapeInput, ScrapeOutput
+from app.exceptions import ForbiddenError
+from app.proprietary.platforms.indeed_jobs import (
+    IndeedAccessBlockedError,
+    IndeedScrapeInput,
+    scrape_indeed,
+)
+
+ScrapeFn = Callable[..., Awaitable[list[dict]]]
+
+
+def build_scrape_executor(scrape_fn: ScrapeFn | None = None) -> Executor:
+    """Bind the executor to a scraper fn (defaults to the proprietary actor)."""
+    scrape_fn = scrape_fn or scrape_indeed
+
+    async def execute(payload: ScrapeInput) -> ScrapeOutput:
+        actor_input = IndeedScrapeInput(
+            startUrls=[{"url": url} for url in payload.urls],
+            queries=payload.search_queries,
+            country=payload.country,
+            location=payload.location,
+            radius=payload.radius,
+            jobType=payload.job_type,
+            level=payload.level,
+            remote=payload.remote,
+            fromDays=payload.from_days,
+            sort=payload.sort,
+            maxItems=payload.max_items,
+            maxItemsPerQuery=payload.max_items_per_query,
+        )
+        emit_progress(
+            "starting", "Resolving Indeed targets", total=payload.max_items, unit="job"
+        )
+        try:
+            items = await scrape_fn(actor_input, limit=payload.max_items)
+        except IndeedAccessBlockedError as exc:
+            # Anonymous-only scraper; a hard block can't be retried with creds.
+            raise ForbiddenError(
+                f"Indeed refused anonymous access: {exc}",
+                code="INDEED_ACCESS_BLOCKED",
+            ) from exc
+        emit_progress(
+            "done", f"Scraped {len(items)} job(s)", current=len(items), unit="job"
+        )
+        return ScrapeOutput(items=items)
+
+    return execute