Fix too many polygons issue

This commit is contained in:
Nitzan Pomerantz
2025-10-27 23:58:57 +02:00
parent a234e809bf
commit 3478426006
5 changed files with 42 additions and 12 deletions
+4
View File
@@ -366,6 +366,9 @@ GOVMAP_REQUESTS_PER_SECOND=5.0
GOVMAP_DEFAULT_RADIUS=50 GOVMAP_DEFAULT_RADIUS=50
GOVMAP_DEFAULT_YEARS_BACK=2 GOVMAP_DEFAULT_YEARS_BACK=2
GOVMAP_DEFAULT_DEAL_LIMIT=100 GOVMAP_DEFAULT_DEAL_LIMIT=100
# Performance
GOVMAP_MAX_POLYGONS=10 # Max polygons to query per search (limits API calls)
``` ```
### Tuning Guidelines ### Tuning Guidelines
@@ -379,6 +382,7 @@ GOVMAP_DEFAULT_DEAL_LIMIT=100
- `MAX_RETRIES=2` - `MAX_RETRIES=2`
- `RETRY_MAX_WAIT=5` - `RETRY_MAX_WAIT=5`
- `REQUESTS_PER_SECOND=10.0` - `REQUESTS_PER_SECOND=10.0`
- `MAX_POLYGONS=5` (faster, fewer results)
**Conservative (Shared API):** **Conservative (Shared API):**
- `MAX_RETRIES=3` - `MAX_RETRIES=3`
+3
View File
@@ -272,6 +272,9 @@ GOVMAP_REQUESTS_PER_SECOND=5.0
GOVMAP_DEFAULT_RADIUS=50 GOVMAP_DEFAULT_RADIUS=50
GOVMAP_DEFAULT_YEARS_BACK=2 GOVMAP_DEFAULT_YEARS_BACK=2
GOVMAP_DEFAULT_DEAL_LIMIT=100 GOVMAP_DEFAULT_DEAL_LIMIT=100
# Performance
GOVMAP_MAX_POLYGONS=10 # Limit polygons per search (reduces API calls, improves speed)
``` ```
## Development Roadmap ## Development Roadmap
+7
View File
@@ -57,6 +57,11 @@ class GovmapConfig:
default_deal_limit: int = field( default_deal_limit: int = field(
default_factory=lambda: int(os.getenv("GOVMAP_DEFAULT_DEAL_LIMIT", "100")) default_factory=lambda: int(os.getenv("GOVMAP_DEFAULT_DEAL_LIMIT", "100"))
) )
# Performance optimization
max_polygons_to_query: int = field(
default_factory=lambda: int(os.getenv("GOVMAP_MAX_POLYGONS", "10"))
)
# User agent # User agent
user_agent: str = field( user_agent: str = field(
@@ -90,6 +95,8 @@ class GovmapConfig:
raise ValueError("default_years_back must be positive") raise ValueError("default_years_back must be positive")
if self.default_deal_limit <= 0: if self.default_deal_limit <= 0:
raise ValueError("default_deal_limit must be positive") raise ValueError("default_deal_limit must be positive")
if self.max_polygons_to_query <= 0:
raise ValueError("max_polygons_to_query must be positive")
if not self.base_url: if not self.base_url:
raise ValueError("base_url cannot be empty") raise ValueError("base_url cannot be empty")
if not self.user_agent: if not self.user_agent:
+23 -8
View File
@@ -604,6 +604,12 @@ class GovmapClient:
logger.info(f"Found {len(polygon_ids)} unique polygon IDs") logger.info(f"Found {len(polygon_ids)} unique polygon IDs")
# Limit polygons to query (performance optimization)
max_polygons = self.config.max_polygons_to_query
if len(polygon_ids) > max_polygons:
polygon_ids = list(polygon_ids)[:max_polygons]
logger.info(f"Limited to {max_polygons} polygons for performance")
# Step 3: Calculate date range # Step 3: Calculate date range
end_date = datetime.now() end_date = datetime.now()
start_date = end_date - timedelta(days=years_back * 365) start_date = end_date - timedelta(days=years_back * 365)
@@ -618,6 +624,12 @@ class GovmapClient:
seen_deals = set() # For deduplication seen_deals = set() # For deduplication
for polygon_id in polygon_ids: for polygon_id in polygon_ids:
# Early termination if we have enough deals
total_collected = len(building_deals) + len(street_deals) + len(neighborhood_deals)
if total_collected >= max_deals:
logger.info(f"Collected {total_collected} deals, stopping polygon queries")
break
try: try:
# Get street deals first (higher priority) # Get street deals first (higher priority)
current_street_deals = self.get_street_deals( current_street_deals = self.get_street_deals(
@@ -628,14 +640,17 @@ class GovmapClient:
deal_type=deal_type, deal_type=deal_type,
) )
# Get neighborhood deals (lower priority) # Get neighborhood deals (lower priority) - optional for performance
current_neighborhood_deals = self.get_neighborhood_deals( current_neighborhood_deals = []
polygon_id, # Skip neighborhood deals if we have enough street deals
limit=max_deals // 4, # Allocate less to neighborhood deals if len(street_deals) < max_deals // 2:
start_date=start_date_str, current_neighborhood_deals = self.get_neighborhood_deals(
end_date=end_date_str, polygon_id,
deal_type=deal_type, limit=max_deals // 4, # Allocate less to neighborhood deals
) start_date=start_date_str,
end_date=end_date_str,
deal_type=deal_type,
)
# Process street deals and separate building deals # Process street deals and separate building deals
for deal in current_street_deals: for deal in current_street_deals:
+5 -4
View File
@@ -55,10 +55,11 @@ def validate_coordinates(point: Tuple[float, float]) -> Tuple[float, float]:
raise ValueError("Coordinates must be numeric values") raise ValueError("Coordinates must be numeric values")
# Basic validation for Israeli coordinates (ITM projection) # Basic validation for Israeli coordinates (ITM projection)
if not (0 < lon < 400000): # Rough bounds for Israeli ITM longitude # ITM bounds for Israel: X (longitude) ~150,000-300,000, Y (latitude) ~3,500,000-4,000,000
logger.warning(f"Longitude {lon} may be outside Israeli bounds") if not (150000 <= lon <= 300000): # ITM longitude bounds for Israel
if not (0 < lat < 1400000): # Rough bounds for Israeli ITM latitude logger.warning(f"Longitude {lon} appears to be outside Israeli ITM bounds (150,000-300,000)")
logger.warning(f"Latitude {lat} may be outside Israeli bounds") if not (3500000 <= lat <= 4000000): # ITM latitude bounds for Israel
logger.warning(f"Latitude {lat} appears to be outside Israeli ITM bounds (3,500,000-4,000,000)")
return (lon, lat) return (lon, lat)