fix polygons

Signed-off-by: zyphlar <zyphlar@gmail.com>
This commit is contained in:
zyphlar
2026-01-04 16:38:53 -08:00
parent e9406c0f36
commit 9e45e04d03
3 changed files with 177 additions and 44 deletions

View File

@@ -33,43 +33,169 @@ except ImportError:
print("Error: duckdb is required. Install with: pip install duckdb", file=sys.stderr)
sys.exit(1)
try:
from shapely.geometry import Point, Polygon, MultiPolygon
from shapely.strtree import STRtree
except ImportError:
print("Error: shapely is required. Install with: pip install shapely", file=sys.stderr)
sys.exit(1)
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)
def load_country_polygons(polygons_file: Path) -> Dict[str, any]:
def parse_poly_file(poly_path: Path) -> MultiPolygon:
"""
Load country polygons from packed_polygons.bin file.
Parse an Osmosis .poly file and return a Shapely MultiPolygon.
This is a placeholder - actual implementation would need to parse the binary format.
For now, we'll use a simpler approach with DuckDB spatial functions.
.poly format:
Line 1: Region name
Section N: (numbered 1, 2, 3...)
lon lat (pairs of coordinates)
...
END
"""
# TODO: Implement actual polygon loading from packed_polygons.bin
# For MVP, we can use a simplified approach or require pre-processed country boundaries
logger.warning("Country polygon loading not yet implemented - using fallback method")
return {}
with open(poly_path, 'r', encoding='utf-8') as f:
lines = f.readlines()
polygons = []
current_coords = []
in_section = False
def determine_country_from_coords(lat: float, lon: float, conn: duckdb.DuckDBPyConnection) -> str:
"""
Determine which country a coordinate belongs to.
for line in lines[1:]: # Skip first line (region name)
line = line.strip()
This uses a simple approach for MVP - can be enhanced later.
Returns country name or "Unknown" if not found.
"""
# Simplified country detection for MVP
# TODO: Use actual country polygons for accurate spatial join
if not line:
continue
# For now, return a simplified country code based on rough lat/lon bounds
# This is just for initial testing - real implementation needs proper spatial join
if 40 < lat < 52 and -5 < lon < 10:
return "France"
elif 45 < lat < 48 and 5 < lon < 11:
return "Switzerland"
elif 43 < lat < 44 and 7 < lon < 8:
return "Monaco"
if line.upper() == 'END':
if current_coords:
# Close the polygon if needed
if current_coords[0] != current_coords[-1]:
current_coords.append(current_coords[0])
# Create polygon (need at least 3 points + closing point)
if len(current_coords) >= 4:
try:
poly = Polygon(current_coords)
# If polygon is invalid, try to fix it
if not poly.is_valid:
# Try buffer(0) trick to fix self-intersections
poly = poly.buffer(0)
# Only accept if it's now valid and is a Polygon or MultiPolygon
if poly.is_valid and not poly.is_empty:
if poly.geom_type == 'Polygon':
polygons.append(poly)
elif poly.geom_type == 'MultiPolygon':
# Split multipolygon into individual polygons
polygons.extend(poly.geoms)
else:
logger.debug(f"Skipping invalid section in {poly_path.name}")
except Exception as e:
logger.debug(f"Error creating polygon in {poly_path.name}: {e}")
current_coords = []
in_section = False
continue
# Try to parse as section number
try:
int(line)
in_section = True
continue
except ValueError:
pass
# Parse coordinate pair
if in_section:
parts = line.split()
if len(parts) >= 2:
try:
lon = float(parts[0])
lat = float(parts[1])
current_coords.append((lon, lat))
except ValueError:
pass
if not polygons:
logger.warning(f"No valid polygons found in {poly_path.name}")
return None
if len(polygons) == 1:
return MultiPolygon([polygons[0]])
else:
return "Unknown"
return MultiPolygon(polygons)
def load_country_polygons(borders_dir: Path) -> Dict[str, MultiPolygon]:
"""
Load all .poly files from the borders directory.
Returns a dict mapping region name (without .poly extension) to MultiPolygon.
"""
logger.info(f"Loading .poly files from {borders_dir}")
poly_files = list(borders_dir.glob("*.poly"))
logger.info(f"Found {len(poly_files)} .poly files")
polygons = {}
for poly_file in poly_files:
region_name = poly_file.stem # Filename without .poly extension
try:
multi_polygon = parse_poly_file(poly_file)
if multi_polygon:
polygons[region_name] = multi_polygon
except Exception as e:
logger.error(f"Error parsing {poly_file.name}: {e}")
continue
logger.info(f"Successfully loaded {len(polygons)} region polygons")
return polygons
class RegionFinder:
"""
Efficient spatial index for finding which region a point belongs to.
Uses Shapely's STRtree for fast spatial queries.
"""
def __init__(self, regions: Dict[str, MultiPolygon]):
logger.info("Building spatial index for region lookup...")
self.regions = regions
self.region_names = []
self.geometries = []
for region_name, multi_polygon in regions.items():
self.region_names.append(region_name)
self.geometries.append(multi_polygon)
# Build R-tree spatial index for fast lookups
self.tree = STRtree(self.geometries)
logger.info(f"Spatial index built with {len(self.geometries)} regions")
def find_region(self, lat: float, lon: float) -> str:
"""
Find which region a coordinate belongs to.
Returns region name or None if not found.
"""
point = Point(lon, lat) # Note: Shapely uses (x, y) = (lon, lat)
# Query the spatial index for candidate polygons
candidates = self.tree.query(point)
# Check each candidate to see if point is actually inside
for idx in candidates:
if self.geometries[idx].contains(point):
return self.region_names[idx]
return None
def write_binary_file(output_path: Path, points: List[Tuple[float, float, str]]):
@@ -109,12 +235,21 @@ def write_binary_file(output_path: Path, points: List[Tuple[float, float, str]])
logger.info(f"Wrote {point_count} points to {output_path}")
def process_parquet_streaming(parquet_url: str, output_dir: Path, batch_size: int = 100000):
def process_parquet_streaming(parquet_url: str, output_dir: Path, borders_dir: Path, batch_size: int = 100000):
"""
Stream the Panoramax parquet file and write per-country binary files.
Uses DuckDB to stream the large parquet file without loading it entirely into memory.
Uses .poly files from borders_dir to categorize points into regions.
"""
# Load region polygons and build spatial index
regions = load_country_polygons(borders_dir)
if not regions:
logger.error("No regions loaded - cannot process panoramax data")
return
region_finder = RegionFinder(regions)
conn = duckdb.connect(database=':memory:')
# Enable httpfs extension for remote file access
@@ -175,12 +310,12 @@ def process_parquet_streaming(parquet_url: str, output_dir: Path, batch_size: in
for row in batch:
lat, lon, image_id = row
# Determine country
country = determine_country_from_coords(lat, lon, conn)
# Find which region this point belongs to
region = region_finder.find_region(lat, lon)
# Skip unknown countries for now (or save to separate file)
if country != "Unknown":
country_points[country].append((lat, lon, str(image_id)))
# Only add points that fall within a defined region
if region:
country_points[region].append((lat, lon, str(image_id)))
# Periodically write to disk to avoid memory issues
if batch_count % 10 == 0:
@@ -229,9 +364,10 @@ def main():
)
parser.add_argument(
'--polygons',
'--borders-dir',
type=Path,
help='Path to packed_polygons.bin file (optional, for accurate country detection)'
default=Path(__file__).parent.parent.parent.parent / 'data' / 'borders',
help='Path to directory containing .poly border files (default: <repo>/data/borders)'
)
parser.add_argument(
@@ -246,19 +382,19 @@ def main():
logger.info("Panoramax Preprocessor starting")
logger.info(f"Input: {args.input}")
logger.info(f"Output directory: {args.output}")
logger.info(f"Borders directory: {args.borders_dir}")
logger.info(f"Batch size: {args.batch_size}")
if args.polygons:
logger.info(f"Country polygons: {args.polygons}")
# TODO: Load and use country polygons for accurate spatial join
else:
logger.warning("No country polygons provided - using simplified country detection")
# Verify borders directory exists
if not args.borders_dir.exists():
logger.error(f"Borders directory not found: {args.borders_dir}")
sys.exit(1)
# Create output directory
args.output.mkdir(parents=True, exist_ok=True)
# Process the parquet file
process_parquet_streaming(args.input, args.output, args.batch_size)
process_parquet_streaming(args.input, args.output, args.borders_dir, args.batch_size)
logger.info("Panoramax preprocessing complete!")