fix(waterdata): attach EPSG:4326 CRS to OGC GeoDataFrames (#343)

The waterdata getters build their GeoDataFrame via
gpd.GeoDataFrame.from_features(...) without a crs= argument, so the
returned object has .crs is None even though the coordinates are
published in EPSG:4326 (WGS84) and the get_monitoring_locations
docstring says so. That makes CRS-aware ops like to_crs and spatial
joins fail on an otherwise valid result, and it's inconsistent with
the nldi (EPSG:4326) and nwis (EPSG:4269) modules, which already tag
their frames. Pass the documented CRS at both build sites
(ogc/shaping.py and waterdata/stats.py) via a shared _CRS constant.

Fixes #342

Signed-off-by: Arpit Jain <arpitjain099@gmail.com>

* test(waterdata): reuse _resp_ok helper in CRS regression test

Replace the hand-built mock.MagicMock response in
test_get_resp_data_attaches_wgs84_crs with the file's existing
_resp_ok() helper, which produces the same 200-OK feature payload.
Drops duplicated mock scaffolding and matches the convention used by
the other tests in this module.

Author: arpitjain099

dataretrieval/ogc/shaping.py

@@ -31,6 +31,12 @@
 
 logger = logging.getLogger(__name__)
 
+# Water Data OGC coordinates are published in WGS84, so attach that CRS where the
+# GeoDataFrame is built (otherwise the result is CRS-naive and ``to_crs`` /
+# spatial joins fail). Mirrors the constants in ``nldi`` (EPSG:4326) and ``nwis``
+# (EPSG:4269).
+_CRS = "EPSG:4326"
+
 # Whether geopandas is present is a static, environment-level fact, so warn
 # once here at import time rather than per query/chunk.
 if not GEOPANDAS:
@@ -141,7 +147,8 @@ def _get_resp_data(
     # a plain DataFrame. Features that already carry geometry (the common
     # sites case) are passed through without a per-feature dict copy.
     df = gpd.GeoDataFrame.from_features(
-        [f if "geometry" in f else {**f, "geometry": None} for f in features]
+        [f if "geometry" in f else {**f, "geometry": None} for f in features],
+        crs=_CRS,
     )
     # Mirror the non-geopandas branch's defensive ``f.get("id")`` so a feature
     # missing a top-level ``id`` yields None rather than a KeyError.

dataretrieval/waterdata/stats.py

@@ -22,6 +22,7 @@
     _run_sync,
 )
 from dataretrieval.ogc.shaping import (
+    _CRS,
     GEOPANDAS,
     _attach_coordinates,
     _empty_feature_frame,
@@ -111,7 +112,8 @@ def _handle_nesting(
         # can't ``KeyError`` on a stats feature that omits geometry — mirrors
         # the guard in :func:`engine._get_resp_data`.
         df = gpd.GeoDataFrame.from_features(
-            [f if "geometry" in f else {**f, "geometry": None} for f in features]
+            [f if "geometry" in f else {**f, "geometry": None} for f in features],
+            crs=_CRS,
         ).drop(columns=["data"], errors="ignore")
 
     # Unnest json features, properties, data, and values while retaining necessary

tests/waterdata_utils_test.py

@@ -596,6 +596,58 @@ class _Sentinel:
     assert isinstance(result, _Sentinel)
 
 
+def test_get_resp_data_attaches_wgs84_crs():
+    """A geometry-bearing Water Data page should come back tagged as
+    EPSG:4326 (the CRS the coordinates are published in), so callers can
+    run ``to_crs`` / spatial joins without first patching in a CRS by
+    hand. Regression for the ``.crs is None`` reported in issue #342."""
+    geopandas = pytest.importorskip("geopandas")
+
+    resp = _resp_ok(
+        [
+            {
+                "type": "Feature",
+                "id": "USGS-01",
+                "geometry": {"type": "Point", "coordinates": [-76.5, 39.2]},
+                "properties": {"monitoring_location_id": "USGS-01"},
+            }
+        ]
+    )
+    df = _get_resp_data(resp, geopd=True)
+    assert isinstance(df, geopandas.GeoDataFrame)
+    assert df.crs == "EPSG:4326"
+
+
+def test_handle_nesting_attaches_wgs84_crs():
+    """The stats path builds its GeoDataFrame the same way, so it should
+    carry EPSG:4326 too (issue #342)."""
+    geopandas = pytest.importorskip("geopandas")
+
+    body = {
+        "next": None,
+        "features": [
+            {
+                "type": "Feature",
+                "geometry": {"type": "Point", "coordinates": [-76.5, 39.2]},
+                "properties": {
+                    "monitoring_location_id": "USGS-01",
+                    "data": [
+                        {
+                            "parameter_code": "00060",
+                            "unit_of_measure": "ft^3/s",
+                            "parent_time_series_id": "ts-1",
+                            "values": [{"statistic_id": "mean", "value": 10.0}],
+                        }
+                    ],
+                },
+            }
+        ],
+    }
+    df = _handle_nesting(body, geopd=True)
+    assert isinstance(df, geopandas.GeoDataFrame)
+    assert df.crs == "EPSG:4326"
+
+
 def test_handle_nesting_tolerates_missing_features_key():
     """A 200 response with a body that doesn't carry ``features`` at
     all (rare but seen in error envelopes) must also short-circuit