Create ingest/fbi_vault.py

ingest/fbi_vault.py

@@ -1,35 +1,44 @@
 import aiohttp
-from ingest.base_adapter import BaseAdapter
+import time
+from ingest.generic_public_foia import GenericFOIAAdapter
+from ingest.health import HealthStatus
 
-class FBIAdapter(BaseAdapter):
-    name = "FBI Vault"
-    source_type = "live"
+class FBIAdapter(GenericFOIAAdapter):
+    source_name = "FBI Vault"
     base_url = "https://vault.fbi.gov"
+    source_type = "live"
 
     async def search(self, query: str):
-        params = {"q": query}
-        results = []
+        start = time.perf_counter()
+        await self._rate_limit()
+
+        url = f"{self.base_url}/search"
+        if not await self._guard(url):
+            self.last_health = HealthStatus(ok=False, latency_ms=0, error="robots.txt")
+            return []
 
-        async with aiohttp.ClientSession() as session:
-            async with session.get(
-                f"{self.base_url}/search",
-                params=params,
-                timeout=15
-            ) as resp:
-                if resp.status != 200:
-                    return []
+        try:
+            async with aiohttp.ClientSession() as session:
+                async with session.get(url, params={"q": query}, timeout=12) as resp:
+                    text = await resp.text()
+                    latency = int((time.perf_counter() - start) * 1000)
 
-                text = await resp.text()
+            if "reading-room" in text.lower():
+                self.last_health = HealthStatus(ok=True, latency_ms=latency)
+                return [{
+                    "source": self.source_name,
+                    "title": f"FBI Vault document mentioning '{query}'",
+                    "url": f"{url}?q={query}",
+                    "snippet": "Public FBI FOIA document from The Vault.",
+                    "live": True,
+                    "extended": False,
+                    "health": self.last_health.__dict__
+                }]
 
-        # Minimal, safe parsing (no deep scrape)
-        for line in text.splitlines():
-            if "/reading-room/" in line:
-                results.append({
-                    "title": f"FBI Vault result for {query}",
-                    "url": self.base_url,
-                    "snippet": "Public FBI Vault FOIA document",
-                    "source": self.name,
-                    "live": True
-                })
+            self.last_health = HealthStatus(ok=True, latency_ms=latency)
+            return []
 
-        return results
+        except Exception as e:
+            latency = int((time.perf_counter() - start) * 1000)
+            self.last_health = HealthStatus(ok=False, latency_ms=latency, error=str(e))
+            return []