Update geocoding intel

drb-c2-core/app/internal/intelligence.py

@@ -8,6 +8,7 @@ Falls back gracefully if the API is unavailable or returns malformed output.
 """
 import asyncio
 import json
+import re
 from typing import Optional
 from app.internal.logger import logger
 from app.internal import firestore as fstore
@@ -37,8 +38,20 @@ System: {system_id}
 Talkgroup: {talkgroup_name}
 {transcript_block}"""
 
-# Nominatim viewbox half-width in degrees (~35 km at mid-latitudes)
-_GEO_DELTA = 0.3
+# Nominatim viewbox half-width in degrees (~11 km at mid-latitudes)
+_GEO_DELTA = 0.1
+
+# node_id → state abbreviation/name from one-time reverse geocode
+_node_state_cache: dict[str, str] = {}
+
+# Strip P25 service suffixes to extract the municipality name from a talkgroup
+_TG_SUFFIX_RE = re.compile(
+    r"\s*\b(police\s*dep(t|artment)?|pd|fire\s*(dep(t|artment)|district)?|"
+    r"ems|rescue|dispatch|fd|tac(tical)?|ops|operations?|command|"
+    r"(fire\s*)?ground|mutual\s*aid|channel|ch\b|car[-\s]to[-\s]car|"
+    r"division|unit)\b.*",
+    re.IGNORECASE,
+)
 
 
 async def extract_tags(
@@ -86,7 +99,11 @@ async def extract_tags(
             node_lat = node_doc.get("lat")
             node_lon = node_doc.get("lon")
             if node_lat is not None and node_lon is not None:
-                location_coords = await _geocode_location(location, node_lat, node_lon)
+                state      = await _get_node_state(node_id, node_lat, node_lon)
+                muni       = _municipality_from_tg(talkgroup_name)
+                hint_parts = [p for p in [muni, state] if p]
+                query      = f"{location}, {', '.join(hint_parts)}" if hint_parts else location
+                location_coords = await _geocode_location(query, node_lat, node_lon)
 
     # Store embedding alongside structured data
     embedding = await asyncio.to_thread(_sync_embed, _embed_text(transcript, incident_type))
@@ -157,6 +174,50 @@ async def _geocode_location(
     return None
 
 
+async def _get_node_state(node_id: str, lat: float, lon: float) -> Optional[str]:
+    """
+    Reverse geocode the node's position once to extract its state.
+    Result is cached for the process lifetime — nodes don't move.
+    """
+    if node_id in _node_state_cache:
+        return _node_state_cache[node_id]
+
+    import httpx
+    headers = {"User-Agent": "DRB-Dispatch/1.0 (public-safety radio monitor)"}
+    try:
+        async with httpx.AsyncClient(timeout=5.0) as client:
+            r = await client.get(
+                "https://nominatim.openstreetmap.org/reverse",
+                params={"lat": lat, "lon": lon, "format": "json", "zoom": 5},
+                headers=headers,
+            )
+            r.raise_for_status()
+            data = r.json()
+            state = data.get("address", {}).get("state", "")
+            if state:
+                _node_state_cache[node_id] = state
+                logger.info(f"Node {node_id} reverse-geocoded to state: {state!r}")
+                return state
+    except Exception as e:
+        logger.warning(f"Node state reverse geocode failed: {e}")
+    return None
+
+
+def _municipality_from_tg(tg_name: Optional[str]) -> Optional[str]:
+    """
+    Extract the municipality name from a talkgroup name.
+    e.g. "Ossining PD" → "Ossining", "Westchester County Fire" → "Westchester County"
+    Returns None for tactical/operational channels with no useful location info.
+    """
+    if not tg_name:
+        return None
+    cleaned = _TG_SUFFIX_RE.sub("", tg_name).strip()
+    # Discard if nothing left, purely numeric, or a short all-caps abbreviation (e.g. "WC", "TAC")
+    if not cleaned or cleaned.isdigit() or (len(cleaned) <= 3 and cleaned.isupper()):
+        return None
+    return cleaned
+
+
 def _build_transcript_block(transcript: str, segments: Optional[list[dict]]) -> str:
     """Format transcript as numbered transmissions if segments are available."""
     if segments and len(segments) > 1: