Spaces:
Sleeping
Sleeping
Create ingest/extended_live_adapter.py
Browse files
ingest/extended_live_adapter.py
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import aiohttp
|
| 2 |
+
from ingest.generic_public_foia import GenericFOIAAdapter
|
| 3 |
+
|
| 4 |
+
class ExtendedLiveAdapter(GenericFOIAAdapter):
|
| 5 |
+
"""
|
| 6 |
+
Extended live adapter.
|
| 7 |
+
Only active when Extended Features are enabled.
|
| 8 |
+
Uses conservative public search / keyword presence.
|
| 9 |
+
"""
|
| 10 |
+
|
| 11 |
+
source_type = "live"
|
| 12 |
+
extended_only = True # UI gate
|
| 13 |
+
|
| 14 |
+
async def _safe_search(self, search_url: str, query: str):
|
| 15 |
+
await self._rate_limit()
|
| 16 |
+
|
| 17 |
+
try:
|
| 18 |
+
async with aiohttp.ClientSession() as session:
|
| 19 |
+
async with session.get(
|
| 20 |
+
search_url,
|
| 21 |
+
params={"q": query},
|
| 22 |
+
timeout=12,
|
| 23 |
+
headers={"User-Agent": "HF-FOIA-Search/1.0"}
|
| 24 |
+
) as resp:
|
| 25 |
+
if resp.status != 200:
|
| 26 |
+
return []
|
| 27 |
+
|
| 28 |
+
text = await resp.text()
|
| 29 |
+
|
| 30 |
+
# conservative presence check
|
| 31 |
+
if query.lower() in text.lower():
|
| 32 |
+
return [{
|
| 33 |
+
"source": self.source_name,
|
| 34 |
+
"title": f"{self.source_name} public document mentioning '{query}'",
|
| 35 |
+
"url": search_url,
|
| 36 |
+
"snippet": "Public FOIA-released material (extended live mode).",
|
| 37 |
+
"live": True,
|
| 38 |
+
"extended": True
|
| 39 |
+
}]
|
| 40 |
+
except Exception:
|
| 41 |
+
return []
|
| 42 |
+
|
| 43 |
+
return []
|