Fix issue with read_json and add tests for different file types

lux/core/__init__.py

@@ -22,7 +22,9 @@
 
 def setOption(overridePandas=True):
     if overridePandas:
-        pd.DataFrame = pd.io.parsers.DataFrame = pd.core.frame.DataFrame = LuxDataFrame
+        pd.DataFrame = (
+            pd.io.json._json.DataFrame
+        ) = pd.io.parsers.DataFrame = pd.core.frame.DataFrame = LuxDataFrame
     else:
         pd.DataFrame = pd.io.parsers.DataFrame = pd.core.frame.DataFrame = originalDF
 

tests/test_pandas_coverage.py

@@ -560,3 +560,29 @@ def test_str_replace(global_var):
     ], "Metadata is lost when going from Dataframe to Series."
     assert df.cardinality is not None, "Metadata is lost when going from Dataframe to Series."
     assert series.name == "Brand", "Pandas Series original `name` property not retained."
+
+
+################
+# Read Tests #
+################
+
+
+def test_read_json(global_var):
+    url = "https://raw.githubusercontent.com/lux-org/lux-datasets/master/data/car.json"
+    df = pd.read_json(url)
+    df._repr_html_()
+    assert list(df.recommendation.keys()) == [
+        "Correlation",
+        "Distribution",
+        "Occurrence",
+        "Temporal",
+    ]
+    assert len(df.data_type_lookup) == 10
+
+
+def test_read_sas(global_var):
+    url = "https://github.com/lux-org/lux-datasets/blob/master/data/airline.sas7bdat?raw=true"
+    df = pd.read_sas(url, format="sas7bdat")
+    df._repr_html_()
+    assert list(df.recommendation.keys()) == ["Correlation", "Distribution", "Temporal"]
+    assert len(df.data_type_lookup) == 6