fix(ks2): fire-and-forget instead of polling to avoid socket timeout

Kestra's HTTP client socket read timeout is shorter than any reasonable
wait for a full geocoded migration. POST /api/admin/reimport-ks2 returns
immediately with {status:started}; the backend runs the job in a thread.
Check GET /api/admin/reimport-ks2/status or watch the UI for schools.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

integrator/flows/ks2.yml

@@ -10,7 +10,7 @@ tasks:
     uri: http://integrator:8001/run/ks2?action=load
     method: POST
     allowFailed: false
-    timeout: PT2H    # polls backend every 30s; geocoding 15k schools takes up to 1h
+    timeout: PT30S   # fire-and-forget; backend runs migration in background
 
 errors:
   - id: notify-failure

integrator/scripts/sources/ks2.py

@@ -10,13 +10,10 @@ The CSV files must already be present in the data volume under
   /data/{year}/england_ks2final.csv
 (populated at deploy time from the repo's data/ directory).
 """
-import time
 import requests
 from config import BACKEND_URL, ADMIN_API_KEY
 
 HEADERS = {"X-API-Key": ADMIN_API_KEY}
-POLL_INTERVAL = 30   # seconds between status checks
-MAX_WAIT = 7200      # 2 hours
 
 
 def download():
@@ -26,33 +23,19 @@ def download():
 
 
 def load():
-    """Trigger full KS2 re-import and poll until complete."""
+    """Trigger KS2 re-import on the backend and return immediately.
-    start_url = f"{BACKEND_URL}/api/admin/reimport-ks2?geocode=true"
-    status_url = f"{BACKEND_URL}/api/admin/reimport-ks2/status"
 
-    print(f"POST {start_url}")
+    The migration (including geocoding) runs as a background thread on the
-    resp = requests.post(start_url, headers=HEADERS, timeout=30)
+    backend and can take up to an hour. Poll GET /api/admin/reimport-ks2/status
+    to check progress, or simply wait for schools to appear in the UI.
+    """
+    url = f"{BACKEND_URL}/api/admin/reimport-ks2?geocode=true"
+    print(f"POST {url}")
+    resp = requests.post(url, headers=HEADERS, timeout=30)
     resp.raise_for_status()
-    print(f"Started: {resp.json()}")
+    result = resp.json()
-
+    print(f"Result: {result}")
-    print(f"Polling {status_url} every {POLL_INTERVAL}s (max {MAX_WAIT // 60} min)...")
+    return result
-    elapsed = 0
-    while elapsed < MAX_WAIT:
-        time.sleep(POLL_INTERVAL)
-        elapsed += POLL_INTERVAL
-
-        sr = requests.get(status_url, headers=HEADERS, timeout=15)
-        sr.raise_for_status()
-        state = sr.json()
-        print(f"  [{elapsed // 60}m] {state}")
-
-        if state.get("done"):
-            print("Re-import complete.")
-            return state
-        if not state.get("running"):
-            raise RuntimeError(f"Re-import stopped unexpectedly: {state}")
-
-    raise TimeoutError(f"KS2 re-import did not complete within {MAX_WAIT // 60} minutes")
 
 
 if __name__ == "__main__":