Fix vsizip extract error

src/dbcp/extract/fips_tables.py

@@ -1,6 +1,10 @@
 """Extract canonical state and county FIPS tables from the addfips library."""
+
+import shutil
+import tempfile
 from functools import lru_cache
 from importlib.resources import files
+from pathlib import Path
 from typing import Dict
 
 import addfips
@@ -14,6 +18,22 @@
 TRIBAL_LANDS_URI = "gs://dgm-archive/census/tl_2021_us_aiannh.zip"
 
 
+def extract_zipped_shapefile(path: Path) -> gpd.GeoDataFrame:
+    """
+    Create a temporary file from a zipped shapefile and return a GeoDataFrame.
+
+    vsizip doesn't like the '#' in the path so my workaround is to copy the file to a temporary file.
+
+    Args:
+        path: path to zipped shapefile
+    Returns:
+        GeoDataFrame
+    """
+    with tempfile.NamedTemporaryFile(delete=True, suffix=".zip") as temp_file:
+        shutil.copyfile(path, temp_file.name)
+        return gpd.read_file(temp_file.name)
+
+
 @lru_cache(maxsize=1)  # county boundaries are also used in some transform modules
 def _extract_census_counties(census_uri: str) -> pd.DataFrame:
     """Extract canonical county FIPS tables from census data.
@@ -22,8 +42,7 @@ def _extract_census_counties(census_uri: str) -> pd.DataFrame:
         census_uri: path to zipped shapefiles.
     """
     path = dbcp.extract.helpers.cache_gcs_archive_file_locally(census_uri)
-    counties = gpd.read_file(path)
-    return counties
+    return extract_zipped_shapefile(path)
 
 
 def extract_census_tribal_land(archive_uri: str) -> pd.DataFrame:
@@ -36,8 +55,7 @@ def extract_census_tribal_land(archive_uri: str) -> pd.DataFrame:
         output dataframes of county-level info.
     """
     path = dbcp.extract.helpers.cache_gcs_archive_file_locally(archive_uri)
-    counties = gpd.read_file(path)
-    return counties
+    return extract_zipped_shapefile(path)
 
 
 def _extract_state_fips() -> pd.DataFrame: