assign_geo_identifiers() is broken for every geography: driverName is resolved but never passed

[jinskeep-morpc]

Summary

morpc.assign_geo_identifiers() fails for every geography it supports. The function resolves the correct GDAL driver for the zipped TIGER shapefile and then never passes it to load_spatial_data(), which rejects the .zip extension.

Present on main at v0.5.18. Reproduced against v0.5.17; the relevant lines are identical in both.

Reproduction

import morpc, geopandas as gpd
from shapely.geometry import Point

points = gpd.GeoDataFrame(geometry=[Point(-82.9988, 39.9612)], crs="epsg:4326")
morpc.assign_geo_identifiers(points, ["county"])

morpc.assign_geo_identifiers | INFO | Determining identifiers for geography county
morpc.load_spatial_data | ERROR | File extension is unsupported: .zip.
  It is possible to load zipped spatial data, but you must specify the driver name.
RuntimeError

Cause

morpc/morpc.py. Each geography branch sets driverName — lines 1731, 1736, 1747, 1752 — and the docstring for the block explicitly lists it as one of the four parameters the branches exist to establish:

#   - driverName - What GDAL driver should we use to read the data
...
            driverName = "Census Shapefile" # Custom driver name for load_spatial_data

But the call at line 1776 omits it:

polys = load_spatial_data(sourcePath = filePath, layerName = layerName, verbose=False)

Since every implemented branch (county, tract, zcta, place) points at a .zip, the omission breaks all of them. This is not specific to one geography.

Fix

-        polys = load_spatial_data(sourcePath = filePath, layerName = layerName, verbose=False)
+        polys = load_spatial_data(sourcePath = filePath, layerName = layerName,
+                                  driverName = driverName, verbose=False)

Confirmed sufficient — load_spatial_data() loads the county boundaries correctly once the driver is passed:

morpc.load_spatial_data(
    sourcePath="https://www2.census.gov/geo/tiger/TIGER2020PL/LAYER/COUNTY/2020/tl_2020_39_county20.zip",
    driverName="Census Shapefile", layerName=None, verbose=False)
# -> 88 rows, EPSG:4269, GEOID20 present

Two further problems in the same function

Worth fixing in the same pass, since neither is reachable today:

1. zcta has no identifier field. Line 1748 sets polyIdField = "", so once the driver fix lands, zcta will fail differently: .filter(items=["", "geometry"]) drops the ID column and the subsequent rename is a no-op. It presumably wants ZCTA5CE20 or GEOID20.

2. Stray argument in the rename, line 1782: "id_{}".format(geography, polyIdField) passes two arguments to a single-placeholder format string. Harmless — the extra is ignored — but it reads as though the field name was meant to appear and does not.

Notes

• There is no test coverage for this function under tests/.
• doc/05-morpc-geos-demo.ipynb demonstrates it with morpc.assign_geo_identifiers(libraries, ["county","place"]), so that notebook cannot currently run either.

Impact

Found while building morpc-ohiodbhfacilities-standardize, which needs point-in-polygon county assignment because SAMHSA publishes no county field. That repository does the join inline for now, with a comment pointing back here and instructions to switch to the helper once this is fixed.