Fix sorting of Fried Parameter in UnifiedResponse

changelog/651.bugfix.rst

@@ -0,0 +1 @@
+Fix sorting of UnifiedResponse table

dkist/net/client.py

@@ -32,6 +32,17 @@
 __all__ = ["DKISTClient", "DKISTQueryResponseTable"]
 
 
+def process_nones(results, key, replacement=np.nan):
+    # We need to replace Nones with nans here for sorting purposes
+    # We also need to recreate the whole row so that it can have a numerical dtype
+    # Without this is doesn't sort properly and any nans up in strange places
+    if key in results.colnames:
+        old_r = results[key]
+        results[key] = [replacement] * len(results)
+        notnone = results[key] != None
+        results[key][notnone] = old_r[notnone]
+
+
 class DKISTQueryResponseTable(QueryResponseTable):
     """
     Results of a DKIST Dataset search.
@@ -98,10 +109,7 @@ def _process_table(results: "DKISTQueryResponseTable") -> "DKISTQueryResponseTab
                 results[colname][none_values] = np.nan
             results[colname] = u.Quantity(results[colname], unit=unit)
 
-        if "Average Fried Parameter" in results.colnames:
-            r_none_values = np.array(results["Average Fried Parameter"] == None)
-            if r_none_values.any():
-                results["Average Fried Parameter"][r_none_values] = np.nan
+        process_nones(results, "Average Fried Parameter")
 
         if results and "Wavelength" not in results.colnames:
             results["Wavelength"] = u.Quantity([results["Wavelength Min"], results["Wavelength Max"]]).T

dkist/net/tests/test_client.py

@@ -1,3 +1,4 @@
+import copy
 import json
 
 import hypothesis.strategies as st
@@ -104,6 +105,20 @@ def example_api_response():
     }
 
 
+@pytest.fixture
+def example_api_response_multiple_r0(example_api_response):
+    """
+    A larger dummy API response with varied Fried parameter values to test sorting
+    """
+    for _ in range(4):
+        example_api_response["searchResults"].append(copy.copy(example_api_response["searchResults"][0]))
+    example_api_response["searchResults"][1]["qualityAverageFriedParameter"] = 5
+    example_api_response["searchResults"][3]["qualityAverageFriedParameter"] = 1
+    example_api_response["searchResults"][4]["qualityAverageFriedParameter"] = 3
+
+    return example_api_response
+
+
 @pytest.fixture
 def expected_table_keys():
     translated_keys = set(INVENTORY_KEY_MAP.values())
@@ -151,6 +166,13 @@ def test_query_response_from_results(empty_query_response, example_api_response,
     assert np.isnan(qr["Average Fried Parameter"][0])
 
 
+def test_sort_fried_parameter(example_api_response_multiple_r0):
+    qr = DKISTQueryResponseTable.from_results([example_api_response_multiple_r0], client=DKISTClient())
+    qr.sort("Average Fried Parameter")
+    assert all(qr["Average Fried Parameter"][:3] == [1.0, 3.0, 5.0])
+    assert all(np.isnan(qr["Average Fried Parameter"][3:]))
+
+
 def test_query_response_from_results_unknown_field(empty_query_response, example_api_response, expected_table_keys):
     """
     This test asserts that if the API starts returning new fields we don't error, they get passed though verbatim.