Fix get_results: preserve user-supplied WQX3.0 dataProfile

dataretrieval/wqp.py

@@ -127,31 +127,23 @@ def get_results(
     kwargs = _check_kwargs(kwargs)
 
     if legacy is True:
-        if (
-            "dataProfile" in kwargs
-            and kwargs["dataProfile"] not in result_profiles_legacy
-        ):
-            raise TypeError(
-                f"dataProfile {kwargs['dataProfile']} is not a legacy profile.",
-                f"Valid options are {result_profiles_legacy}.",
-            )
-
+        valid_profiles = result_profiles_legacy
+        kind = "legacy"
         url = wqp_url("Result")
-
     else:
-        if (
-            "dataProfile" in kwargs
-            and kwargs["dataProfile"] not in result_profiles_wqx3
-        ):
-            raise TypeError(
-                f"dataProfile {kwargs['dataProfile']} is not a valid WQX3.0"
-                f"profile. Valid options are {result_profiles_wqx3}.",
-            )
-        else:
-            kwargs["dataProfile"] = "fullPhysChem"
-
+        valid_profiles = result_profiles_wqx3
+        kind = "WQX3.0"
         url = wqx3_url("Result")
 
+    profile = kwargs.get("dataProfile")
+    if profile is not None and profile not in valid_profiles:
+        raise ValueError(
+            f"dataProfile {profile!r} is not a valid {kind} profile. "
+            f"Valid options are {valid_profiles}."
+        )
+    if legacy is not True and profile is None:
+        kwargs["dataProfile"] = "fullPhysChem"
+
     response = query(url, kwargs, delimiter=";", ssl_check=ssl_check)
 
     df = pd.read_csv(StringIO(response.text), delimiter=",", low_memory=False)

tests/wqp_test.py

@@ -216,3 +216,25 @@ def test_check_kwargs():
     kwargs = {"mimeType": "foo"}
     with pytest.raises(ValueError):
         kwargs = _check_kwargs(kwargs)
+
+
+def test_get_results_wqx3_preserves_user_dataProfile(requests_mock):
+    """A valid user-supplied WQX3.0 profile must not be overwritten.
+
+    Regression: previously the `else` branch of the `dataProfile` validation
+    triggered whenever the value was *not invalid*, including any valid
+    user-supplied profile, silently overwriting it with 'fullPhysChem'.
+    """
+    request_url = (
+        "https://www.waterqualitydata.us/wqx3/Result/search?"
+        "siteid=UTAHDWQ_WQX-4993795&mimeType=csv&dataProfile=narrow"
+    )
+    response_file_path = "tests/data/wqp3_results.txt"
+    mock_request(requests_mock, request_url, response_file_path)
+
+    df, _md = get_results(
+        legacy=False, siteid="UTAHDWQ_WQX-4993795", dataProfile="narrow"
+    )
+    assert isinstance(df, DataFrame)
+    sent = requests_mock.request_history[-1]
+    assert sent.qs.get("dataprofile") == ["narrow"]