Spaces:
Sleeping
Sleeping
Create ingest/adapters/state.py
Browse files- ingest/adapters/state.py +25 -0
ingest/adapters/state.py
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import aiohttp
|
| 2 |
+
from ingest.generic_public_foia import GenericFOIAAdapter
|
| 3 |
+
|
| 4 |
+
class StateDeptAdapter(GenericFOIAAdapter):
|
| 5 |
+
source_name = "State Department"
|
| 6 |
+
base_url = "https://foia.state.gov/Search/Results.aspx"
|
| 7 |
+
|
| 8 |
+
async def search(self, query: str):
|
| 9 |
+
if not self.robots_allowed():
|
| 10 |
+
return []
|
| 11 |
+
|
| 12 |
+
await self._rate_limit()
|
| 13 |
+
|
| 14 |
+
params = {"searchText": query}
|
| 15 |
+
|
| 16 |
+
async with aiohttp.ClientSession() as session:
|
| 17 |
+
async with session.get(self.base_url, params=params) as resp:
|
| 18 |
+
html = await resp.text()
|
| 19 |
+
|
| 20 |
+
return [{
|
| 21 |
+
"source": self.source_name,
|
| 22 |
+
"title": "State Dept FOIA Search",
|
| 23 |
+
"url": self.base_url,
|
| 24 |
+
"snippet": f"Searched for '{query}' in State FOIA database"
|
| 25 |
+
}]
|