Fix issue where LuxDataFrame(df) wasn't working (#416)

* add series equality and value counts test

* black formatting

* fix old value counts test instead

* add pandas tests

* remove str from column group

* save work on groupby bugs so far

* fix merge conflict again

* add new tests and add groupby bug fixes

* remove tests for staging

* update series tests

* add back getitem

* fix merge conflicts for staging

* remove print statements

* revert Makefile

* add test for name column case

* add test to ensure column is not all None

* re-add tests and commit new groupby implementation

* make parent LuxGroupby

* reformat LuxGroupBy and fix SeriesGroupby issue

* remove pandas tests

* run black

* finish seriesgroupby metadata propagation and add test

* run black

* fall back to normal if intent is invalid

* run black

* taking care of multiple attributes in clause case

* run black

* modified tests

* replace logic with validate_intent

* save changes

* add _testing to override Pandas

* cleanup code

* revert ipython display widget

* run black

* add test for bug fixed

* run black

Co-authored-by: Doris Lee <dorisjunglinlee@gmail.com>

lux/core/__init__.py

@@ -57,7 +57,9 @@ def setOption(overridePandas=True):
             pd.io.pytables.DataFrame
         ) = (
             pd.io.spss.DataFrame
-        ) = pd.io.stata.DataFrame = pd.io.api.DataFrame = pd.core.frame.DataFrame = LuxDataFrame
+        ) = (
+            pd.io.stata.DataFrame
+        ) = pd.io.api.DataFrame = pd.core.frame.DataFrame = pd._testing.DataFrame = LuxDataFrame
         pd.Series = pd.core.series.Series = pd.core.groupby.ops.Series = pd._testing.Series = LuxSeries
         pd.core.groupby.generic.DataFrameGroupBy = LuxDataFrameGroupBy
         pd.core.groupby.generic.SeriesGroupBy = LuxSeriesGroupBy

lux/core/series.py

@@ -50,6 +50,7 @@ class LuxSeries(pd.Series):
         "_pandas_only",
         "pre_aggregated",
         "_type_override",
+        "name",
     ]
 
     _default_metadata = {

tests/conftest.py

@@ -30,4 +30,5 @@ def global_var():
         "_pandas_only",
         "pre_aggregated",
         "_type_override",
+        "name",
     ]

tests/test_pandas_coverage.py

@@ -32,6 +32,17 @@ def test_deepcopy(global_var):
     check_metadata_equal(df, saved_df)
 
 
+def test_create_ldf_from_ldf(global_var):
+    import pandas._testing as tm
+
+    df = tm.makeDataFrame()
+    df._ipython_display_()
+    saved_df = lux.LuxDataFrame(df)
+    saved_df._ipython_display_()
+    assert len(df.columns) == len(saved_df.columns)
+    assert all([a == b for a, b in zip(df.columns, saved_df.columns)])
+
+
 def test_rename_inplace(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")

tests/test_series.py

@@ -47,6 +47,7 @@ def test_df_to_series():
         "_pandas_only",
         "pre_aggregated",
         "_type_override",
+        "name",
     ], "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 == "Weight", "Pandas Series original `name` property not retained."