From 21e20cf67865642aa68fba56c175a4bdd5917290 Mon Sep 17 00:00:00 2001
From: Piyush Gupta <piyushg9794@gmail.com>
Date: Thu, 19 Nov 2020 18:05:07 +0530
Subject: [PATCH] Global shared variable in test (#144) (#149)

* using global shared variable in test (lux-org#144)

* modified fixture scope as session, resolved dependency test cases

* run black

Co-authored-by: Doris Lee <dorisjunglinlee@gmail.com>
---
 .gitignore                    |   1 +
 lux/core/frame.py             |   1 +
 tests/conftest.py             |  10 +++
 tests/test_action.py          |  32 +++----
 tests/test_compiler.py        |  67 +++++++--------
 tests/test_config.py          |  20 ++---
 tests/test_dates.py           |   8 +-
 tests/test_display.py         |  16 ++--
 tests/test_error_warning.py   |  19 +++--
 tests/test_executor.py        |  44 +++++-----
 tests/test_interestingness.py |  37 ++++----
 tests/test_maintainence.py    |  22 ++---
 tests/test_nan.py             |   4 +-
 tests/test_pandas.py          |   4 +-
 tests/test_pandas_coverage.py |  54 ++++++------
 tests/test_parser.py          | 153 ++++++++++++++++++----------------
 tests/test_performance.py     |   3 +-
 tests/test_type.py            |   7 +-
 tests/test_vis.py             |  74 ++++++++--------
 19 files changed, 301 insertions(+), 275 deletions(-)
 create mode 100644 tests/conftest.py

diff --git a/.gitignore b/.gitignore
index 4f6eff0f..ecd2d1cc 100644
--- a/.gitignore
+++ b/.gitignore
@@ -10,5 +10,6 @@ static/
 *.egg-info*
 build/
 .DS_Store
+.idea/
 tests/.coverage
 tests/coverage.xml
diff --git a/lux/core/frame.py b/lux/core/frame.py
index fe8b9b97..45e91ed4 100644
--- a/lux/core/frame.py
+++ b/lux/core/frame.py
@@ -246,6 +246,7 @@ def intent(self, intent_input: Union[List[Union[str, Clause]], Vis]):
 
     def clear_intent(self):
         self.intent = []
+        self.expire_recs()
 
     def set_intent(self, intent: List[Union[str, Clause]]):
         """
diff --git a/tests/conftest.py b/tests/conftest.py
new file mode 100644
index 00000000..2b95b1cd
--- /dev/null
+++ b/tests/conftest.py
@@ -0,0 +1,10 @@
+import pytest
+import pandas as pd
+
+
+@pytest.fixture(scope="session")
+def global_var():
+    url = "https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true"
+    pytest.olympic = pd.read_csv(url)
+    pytest.car_df = pd.read_csv("lux/data/car.csv")
+    pytest.college_df = pd.read_csv("lux/data/college.csv")
diff --git a/tests/test_action.py b/tests/test_action.py
index 44337181..8af08658 100644
--- a/tests/test_action.py
+++ b/tests/test_action.py
@@ -19,16 +19,16 @@
 from lux.vis.Vis import Vis
 
 
-def test_vary_filter_val():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vary_filter_val(global_var):
+    df = pytest.olympic
     vis = Vis(["Height", "SportType=Ball"], df)
     df.set_intent_as_vis(vis)
     df._repr_html_()
     assert len(df.recommendation["Filter"]) == len(df["SportType"].unique()) - 1
 
 
-def test_filter_inequality():
-    df = pd.read_csv("lux/data/car.csv")
+def test_filter_inequality(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent(
@@ -48,9 +48,9 @@ def test_filter_inequality():
     assert fltr_clause.value == 10
 
 
-def test_generalize_action():
+def test_generalize_action(global_var):
     # test that generalize action creates all unique visualizations
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(
         df["Year"], format="%Y"
     )  # change pandas dtype for the column "Year" to datetype
@@ -72,9 +72,10 @@ def test_generalize_action():
     assert check1 and check2 and check3
 
 
-def test_row_column_group():
-    url = "https://github.com/lux-org/lux-datasets/blob/master/data/state_timeseries.csv?raw=true"
-    df = pd.read_csv(url)
+def test_row_column_group(global_var):
+    df = pd.read_csv(
+        "https://github.com/lux-org/lux-datasets/blob/master/data/state_timeseries.csv?raw=true"
+    )
     df["Date"] = pd.to_datetime(df["Date"])
     tseries = df.pivot(index="State", columns="Date", values="Value")
     # Interpolating missing values
@@ -85,8 +86,8 @@ def test_row_column_group():
     assert list(tseries.recommendation.keys()) == ["Row Groups", "Column Groups"]
 
 
-def test_groupby():
-    df = pd.read_csv("lux/data/college.csv")
+def test_groupby(global_var):
+    df = pytest.college_df
     groupbyResult = df.groupby("Region").sum()
     groupbyResult._repr_html_()
     assert list(groupbyResult.recommendation.keys()) == ["Column Groups"]
@@ -159,17 +160,17 @@ def test_crosstab():
     assert list(result.recommendation.keys()) == ["Row Groups", "Column Groups"]
 
 
-def test_custom_aggregation():
+def test_custom_aggregation(global_var):
     import numpy as np
 
-    df = pd.read_csv("lux/data/college.csv")
+    df = pytest.college_df
     df.set_intent(["HighestDegree", lux.Clause("AverageCost", aggregation=np.ptp)])
     df._repr_html_()
     assert list(df.recommendation.keys()) == ["Enhance", "Filter", "Generalize"]
 
 
-def test_year_filter_value():
-    df = pd.read_csv("lux/data/car.csv")
+def test_year_filter_value(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df.set_intent(["Acceleration", "Horsepower"])
     df._repr_html_()
@@ -191,3 +192,4 @@ def test_year_filter_value():
     assert (
         "T00:00:00.000000000" not in vis.to_Altair()
     ), "Year filter title contains extraneous string, not displayed as summarized string"
+    df.clear_intent()
diff --git a/tests/test_compiler.py b/tests/test_compiler.py
index 037b7534..27857598 100644
--- a/tests/test_compiler.py
+++ b/tests/test_compiler.py
@@ -19,11 +19,11 @@
 from lux.vis.VisList import VisList
 
 
-def test_underspecified_no_vis(test_recs):
+def test_underspecified_no_vis(global_var, test_recs):
     no_vis_actions = ["Correlation", "Distribution", "Occurrence", "Temporal"]
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     test_recs(df, no_vis_actions)
-    assert len(df.current_vis) == 0
+    assert df.current_vis is None or len(df.current_vis) == 0
 
     # test only one filter context case.
     df.set_intent([lux.Clause(attribute="Origin", filter_op="=", value="USA")])
@@ -31,9 +31,9 @@ def test_underspecified_no_vis(test_recs):
     assert len(df.current_vis) == 0
 
 
-def test_underspecified_single_vis(test_recs):
+def test_underspecified_single_vis(global_var, test_recs):
     one_vis_actions = ["Enhance", "Filter", "Generalize"]
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     df.set_intent([lux.Clause(attribute="MilesPerGal"), lux.Clause(attribute="Weight")])
     test_recs(df, one_vis_actions)
     assert len(df.current_vis) == 1
@@ -42,6 +42,7 @@ def test_underspecified_single_vis(test_recs):
         assert attr.data_model == "measure"
     for attr in df.current_vis[0]._inferred_intent:
         assert attr.data_type == "quantitative"
+    df.clear_intent()
 
 
 # def test_underspecified_vis_collection(test_recs):
@@ -75,8 +76,8 @@ def test_underspecified_single_vis(test_recs):
 # 	df.set_intent([lux.Clause(attribute ="?", data_model="measure"), lux.Clause(attribute ="?", data_model="measure")])
 # 	assert len(df.current_vis) == len([vis.get_attr_by_data_model("measure") for vis in df.current_vis]) #should be 25
 # 	test_recs(df, multiple_vis_actions)
-def test_set_intent_as_vis(test_recs):
-    df = pd.read_csv("lux/data/car.csv")
+def test_set_intent_as_vis(global_var, test_recs):
+    df = pytest.car_df
     df._repr_html_()
     vis = df.recommendation["Correlation"][0]
     df.intent = vis
@@ -95,19 +96,19 @@ def test_recs_function(df, actions):
     return test_recs_function
 
 
-def test_parse():
-    df = pd.read_csv("lux/data/car.csv")
+def test_parse(global_var):
+    df = pytest.car_df
     vlst = VisList([lux.Clause("Origin=?"), lux.Clause(attribute="MilesPerGal")], df)
     assert len(vlst) == 3
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vlst = VisList([lux.Clause("Origin=?"), lux.Clause("MilesPerGal")], df)
     assert len(vlst) == 3
 
 
-def test_underspecified_vis_collection_zval():
+def test_underspecified_vis_collection_zval(global_var):
     # check if the number of charts is correct
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vlst = VisList(
         [
             lux.Clause(attribute="Origin", filter_op="=", value="?"),
@@ -123,11 +124,11 @@ def test_underspecified_vis_collection_zval():
     # assert len(vlst) == 8
 
 
-def test_sort_bar():
+def test_sort_bar(global_var):
     from lux.processor.Compiler import Compiler
     from lux.vis.Vis import Vis
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vis = Vis(
         [
             lux.Clause(attribute="Acceleration", data_model="measure", data_type="quantitative"),
@@ -138,7 +139,7 @@ def test_sort_bar():
     assert vis.mark == "bar"
     assert vis._inferred_intent[1].sort == ""
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vis = Vis(
         [
             lux.Clause(attribute="Acceleration", data_model="measure", data_type="quantitative"),
@@ -150,8 +151,8 @@ def test_sort_bar():
     assert vis._inferred_intent[1].sort == "ascending"
 
 
-def test_specified_vis_collection():
-    df = pd.read_csv("lux/data/car.csv")
+def test_specified_vis_collection(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
@@ -181,8 +182,8 @@ def test_specified_vis_collection():
     assert "Origin = Europe" not in chart_titles
 
 
-def test_specified_channel_enforced_vis_collection():
-    df = pd.read_csv("lux/data/car.csv")
+def test_specified_channel_enforced_vis_collection(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     visList = VisList(
@@ -193,9 +194,9 @@ def test_specified_channel_enforced_vis_collection():
         check_attribute_on_channel(vis, "MilesPerGal", "x")
 
 
-def test_autoencoding_scatter():
+def test_autoencoding_scatter(global_var):
     # No channel specified
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     vis = Vis([lux.Clause(attribute="MilesPerGal"), lux.Clause(attribute="Weight")], df)
@@ -234,9 +235,9 @@ def test_autoencoding_scatter():
         )
 
 
-def test_autoencoding_histogram():
+def test_autoencoding_histogram(global_var):
     # No channel specified
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     vis = Vis([lux.Clause(attribute="MilesPerGal", channel="y")], df)
@@ -247,8 +248,8 @@ def test_autoencoding_histogram():
     assert vis.get_attr_by_channel("y")[0].attribute == "Record"
 
 
-def test_autoencoding_line_chart():
-    df = pd.read_csv("lux/data/car.csv")
+def test_autoencoding_line_chart(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     vis = Vis([lux.Clause(attribute="Year"), lux.Clause(attribute="Acceleration")], df)
@@ -287,8 +288,8 @@ def test_autoencoding_line_chart():
         )
 
 
-def test_autoencoding_color_line_chart():
-    df = pd.read_csv("lux/data/car.csv")
+def test_autoencoding_color_line_chart(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     intent = [
@@ -302,8 +303,8 @@ def test_autoencoding_color_line_chart():
     check_attribute_on_channel(vis, "Origin", "color")
 
 
-def test_autoencoding_color_scatter_chart():
-    df = pd.read_csv("lux/data/car.csv")
+def test_autoencoding_color_scatter_chart(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     vis = Vis(
@@ -327,10 +328,10 @@ def test_autoencoding_color_scatter_chart():
     check_attribute_on_channel(vis, "Acceleration", "color")
 
 
-def test_populate_options():
+def test_populate_options(global_var):
     from lux.processor.Compiler import Compiler
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     df.set_intent([lux.Clause(attribute="?"), lux.Clause(attribute="MilesPerGal")])
     col_set = set()
     for specOptions in Compiler.populate_wildcard_options(df._intent, df)["attributes"]:
@@ -355,8 +356,8 @@ def test_populate_options():
     )
 
 
-def test_remove_all_invalid():
-    df = pd.read_csv("lux/data/car.csv")
+def test_remove_all_invalid(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     # with pytest.warns(UserWarning,match="duplicate attribute specified in the intent"):
     df.set_intent(
diff --git a/tests/test_config.py b/tests/test_config.py
index adfd2655..ba9fe05d 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -48,8 +48,8 @@ def contain_horsepower(df):
     return df
 
 
-def test_default_actions_registered():
-    df = pd.read_csv("lux/data/car.csv")
+def test_default_actions_registered(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert "Distribution" in df.recommendation
     assert len(df.recommendation["Distribution"]) > 0
@@ -91,13 +91,13 @@ def test_no_validator():
     assert "bars" in df.recommendation
 
 
-def test_invalid_function():
+def test_invalid_function(global_var):
     df = pd.read_csv("lux/data/car.csv")
     with pytest.raises(ValueError, match="Value must be a callable"):
         lux.register_action("bars", "not a Callable")
 
 
-def test_invalid_validator():
+def test_invalid_validator(global_var):
     df = pd.read_csv("lux/data/car.csv")
 
     def random_categorical(ldf):
@@ -136,14 +136,14 @@ def test_remove_action():
     )
 
 
-def test_remove_invalid_action():
-    df = pd.read_csv("lux/data/car.csv")
+def test_remove_invalid_action(global_var):
+    df = pytest.car_df
     with pytest.raises(ValueError, match="Option 'bars' has not been registered"):
         lux.remove_action("bars")
 
 
-def test_remove_default_actions():
-    df = pd.read_csv("lux/data/car.csv")
+def test_remove_default_actions(global_var):
+    df = pytest.car_df
     df._repr_html_()
 
     lux.remove_action("Distribution")
@@ -178,8 +178,8 @@ def test_remove_default_actions():
 
 
 # TODO: This test does not pass in pytest but is working in Jupyter notebook.
-# def test_plot_setting():
-# 	df = pd.read_csv("lux/data/car.csv")
+# def test_plot_setting(global_var):
+# 	df = pytest.car_df
 # 	df["Year"] = pd.to_datetime(df["Year"], format='%Y')
 # 	def change_color_add_title(chart):
 # 		chart = chart.configure_mark(color="green") # change mark color to green
diff --git a/tests/test_dates.py b/tests/test_dates.py
index 8a5cc823..b8e902b9 100644
--- a/tests/test_dates.py
+++ b/tests/test_dates.py
@@ -20,7 +20,7 @@
 from lux.executor.PandasExecutor import PandasExecutor
 
 
-def test_dateformatter():
+def test_dateformatter(global_var):
     ldf = pd.read_csv("lux/data/car.csv")
     # change pandas dtype for the column "Year" to datetype
     ldf["Year"] = pd.to_datetime(ldf["Year"], format="%Y")
@@ -37,7 +37,7 @@ def test_dateformatter():
     assert date_utils.date_formatter(timestamp, ldf) == "2019-8-26"
 
 
-def test_period_selection():
+def test_period_selection(global_var):
     ldf = pd.read_csv("lux/data/car.csv")
     ldf["Year"] = pd.to_datetime(ldf["Year"], format="%Y")
 
@@ -56,7 +56,7 @@ def test_period_selection():
     assert all(ldf.current_vis[2].data.columns == ["Year", "Acceleration"])
 
 
-def test_period_filter():
+def test_period_filter(global_var):
     ldf = pd.read_csv("lux/data/car.csv")
     ldf["Year"] = pd.to_datetime(ldf["Year"], format="%Y")
 
@@ -70,7 +70,7 @@ def test_period_filter():
     assert isinstance(ldf.recommendation["Filter"][2]._inferred_intent[2].value, pd.Period)
 
 
-def test_period_to_altair():
+def test_period_to_altair(global_var):
     chart = None
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
diff --git a/tests/test_display.py b/tests/test_display.py
index 54da6ca4..7716867c 100644
--- a/tests/test_display.py
+++ b/tests/test_display.py
@@ -19,23 +19,23 @@
 from lux.vis.VisList import VisList
 
 
-def test_to_pandas():
-    df = pd.read_csv("lux/data/car.csv")
+def test_to_pandas(global_var):
+    df = pytest.car_df
     df.to_pandas()
 
 
-def test_display_LuxDataframe():
-    df = pd.read_csv("lux/data/car.csv")
+def test_display_LuxDataframe(global_var):
+    df = pytest.car_df
     df._repr_html_()
 
 
-def test_display_Vis():
-    df = pd.read_csv("lux/data/car.csv")
+def test_display_Vis(global_var):
+    df = pytest.car_df
     vis = Vis(["Horsepower", "Acceleration"], df)
     vis._repr_html_()
 
 
-def test_display_VisList():
-    df = pd.read_csv("lux/data/car.csv")
+def test_display_VisList(global_var):
+    df = pytest.car_df
     vislist = VisList(["?", "Acceleration"], df)
     vislist._repr_html_()
diff --git a/tests/test_error_warning.py b/tests/test_error_warning.py
index dbbffe7a..0c6bca73 100644
--- a/tests/test_error_warning.py
+++ b/tests/test_error_warning.py
@@ -17,27 +17,28 @@
 import pandas as pd
 from lux.vis.Vis import Vis
 
+
 # Test suite for checking if the expected errors and warnings are showing up correctly
-def test_intent_str_error():
-    df = pd.read_csv("lux/data/college.csv")
+def test_intent_str_error(global_var):
+    df = pytest.college_df
     with pytest.raises(TypeError, match="Input intent must be either a list"):
         df.intent = "bad string input"
 
 
-def test_export_b4_widget_created():
+def test_export_b4_widget_created(global_var):
     df = pd.read_csv("lux/data/college.csv")
     with pytest.warns(UserWarning, match="No widget attached to the dataframe"):
         df.exported
 
 
-def test_bad_filter():
-    df = pd.read_csv("lux/data/college.csv")
+def test_bad_filter(global_var):
+    df = pytest.college_df
     with pytest.warns(UserWarning, match="Lux can not operate on an empty dataframe"):
         df[df["Region"] == "asdfgh"]._repr_html_()
 
 
-def test_multi_vis():
-    df = pd.read_csv("lux/data/college.csv")
+def test_multi_vis(global_var):
+    df = pytest.college_df
     with pytest.raises(
         SyntaxError,
         match="The intent that you specified corresponds to more than one visualization.",
@@ -70,10 +71,10 @@ def test_multi_vis():
 
 
 # Test Properties with Private Variables Readable but not Writable
-def test_vis_private_properties():
+def test_vis_private_properties(global_var):
     from lux.vis.Vis import Vis
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vis = Vis(["Horsepower", "Weight"], df)
     vis._repr_html_()
     assert isinstance(vis.data, lux.core.frame.LuxDataFrame)
diff --git a/tests/test_executor.py b/tests/test_executor.py
index 268243f0..c3d6aa78 100644
--- a/tests/test_executor.py
+++ b/tests/test_executor.py
@@ -20,8 +20,8 @@
 from lux.vis.VisList import VisList
 
 
-def test_lazy_execution():
-    df = pd.read_csv("lux/data/car.csv")
+def test_lazy_execution(global_var):
+    df = pytest.car_df
     intent = [
         lux.Clause(attribute="Horsepower", aggregation="mean"),
         lux.Clause(attribute="Origin"),
@@ -33,8 +33,8 @@ def test_lazy_execution():
     assert type(vis.data) == lux.core.frame.LuxDataFrame
 
 
-def test_selection():
-    df = pd.read_csv("lux/data/car.csv")
+def test_selection(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     intent = [
@@ -46,8 +46,8 @@ def test_selection():
     assert all(vislist[2].data.columns == ["Year", "Acceleration"])
 
 
-def test_aggregation():
-    df = pd.read_csv("lux/data/car.csv")
+def test_aggregation(global_var):
+    df = pytest.car_df
     intent = [
         lux.Clause(attribute="Horsepower", aggregation="mean"),
         lux.Clause(attribute="Origin"),
@@ -73,11 +73,11 @@ def test_aggregation():
     assert int(result_df[result_df["Origin"] == "Europe"]["Horsepower"]) == 133
 
 
-def test_colored_bar_chart():
+def test_colored_bar_chart(global_var):
     from lux.vis.Vis import Vis
     from lux.vis.Vis import Clause
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
 
     x_clause = Clause(attribute="MilesPerGal", channel="x")
     y_clause = Clause(attribute="Origin", channel="y")
@@ -92,11 +92,11 @@ def test_colored_bar_chart():
     assert len(new_vis.data) == 15 > group_by_cardinality < color_cardinality * group_by_cardinality
 
 
-def test_colored_line_chart():
+def test_colored_line_chart(global_var):
     from lux.vis.Vis import Vis
     from lux.vis.Vis import Clause
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
@@ -114,8 +114,8 @@ def test_colored_line_chart():
     assert len(new_vis.data) == 60 > group_by_cardinality < color_cardinality * group_by_cardinality
 
 
-def test_filter():
-    df = pd.read_csv("lux/data/car.csv")
+def test_filter(global_var):
+    df = pytest.car_df
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     intent = [
@@ -129,8 +129,8 @@ def test_filter():
     assert len(vis.data) == len(df[df["Origin"] == "USA"])
 
 
-def test_inequalityfilter():
-    df = pd.read_csv("lux/data/car.csv")
+def test_inequalityfilter(global_var):
+    df = pytest.car_df
     vis = Vis(
         [
             lux.Clause(attribute="Horsepower", filter_op=">", value=50),
@@ -157,21 +157,21 @@ def test_inequalityfilter():
     assert len(vis.data) == Nbins
 
 
-def test_binning():
-    df = pd.read_csv("lux/data/car.csv")
+def test_binning(global_var):
+    df = pytest.car_df
     vis = Vis([lux.Clause(attribute="Horsepower")], df)
     nbins = list(filter(lambda x: x.bin_size != 0, vis._inferred_intent))[0].bin_size
     assert len(vis.data) == nbins
 
 
-def test_record():
-    df = pd.read_csv("lux/data/car.csv")
+def test_record(global_var):
+    df = pytest.car_df
     vis = Vis([lux.Clause(attribute="Cylinders")], df)
     assert len(vis.data) == len(df["Cylinders"].unique())
 
 
-def test_filter_aggregation_fillzero_aligned():
-    df = pd.read_csv("lux/data/car.csv")
+def test_filter_aggregation_fillzero_aligned(global_var):
+    df = pytest.car_df
     intent = [
         lux.Clause(attribute="Cylinders"),
         lux.Clause(attribute="MilesPerGal"),
@@ -187,8 +187,8 @@ def test_filter_aggregation_fillzero_aligned():
     assert result[result["Cylinders"] == 6]["MilesPerGal"].values[0] == externalValidation[6]
 
 
-def test_exclude_attribute():
-    df = pd.read_csv("lux/data/car.csv")
+def test_exclude_attribute(global_var):
+    df = pytest.car_df
     intent = [lux.Clause("?", exclude=["Name", "Year"]), lux.Clause("Horsepower")]
     vislist = VisList(intent, df)
     for vis in vislist:
diff --git a/tests/test_interestingness.py b/tests/test_interestingness.py
index d62b4b40..eceadcde 100644
--- a/tests/test_interestingness.py
+++ b/tests/test_interestingness.py
@@ -18,9 +18,10 @@
 import numpy as np
 from lux.interestingness.interestingness import interestingness
 
+
 # The following test cases are labelled for vis with <Ndim, Nmsr, Nfilter>
-def test_interestingness_1_0_0():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_1_0_0(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent([lux.Clause(attribute="Origin")])
@@ -57,8 +58,8 @@ def test_interestingness_1_0_0():
     assert rank1 < rank2 and rank1 < rank3 and rank2 < rank3
 
 
-def test_interestingness_1_0_1():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_1_0_1(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent(
@@ -71,8 +72,8 @@ def test_interestingness_1_0_1():
     assert df.current_vis[0].score == 0
 
 
-def test_interestingness_0_1_0():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_0_1_0(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent([lux.Clause(attribute="Horsepower")])
@@ -115,8 +116,8 @@ def test_interestingness_0_1_0():
     assert rank1 < rank2 and rank1 < rank3 and rank2 < rank3
 
 
-def test_interestingness_0_1_1():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_0_1_1(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent(
@@ -130,8 +131,8 @@ def test_interestingness_0_1_1():
     assert str(df.recommendation["Current Vis"][0]._inferred_intent[2].value) == "USA"
 
 
-def test_interestingness_1_1_0():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_1_1_0(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent([lux.Clause(attribute="Horsepower"), lux.Clause(attribute="Year")])
@@ -160,8 +161,8 @@ def test_interestingness_1_1_0():
     assert interestingness(df.recommendation["Filter"][0], df) != None
 
 
-def test_interestingness_1_1_1():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_1_1_1(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent(
@@ -198,12 +199,12 @@ def test_interestingness_1_1_1():
     assert interestingness(df.recommendation["Filter"][0], df) != None
 
 
-def test_interestingness_1_2_0():
+def test_interestingness_1_2_0(global_var):
     from lux.vis.Vis import Vis
     from lux.vis.Vis import Clause
     from lux.interestingness.interestingness import interestingness
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     y_clause = Clause(attribute="Name", channel="y")
     color_clause = Clause(attribute="Cylinders", channel="color")
 
@@ -215,8 +216,8 @@ def test_interestingness_1_2_0():
     assert interestingness(new_vis, df) < 0.01
 
 
-def test_interestingness_0_2_0():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_0_2_0(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent([lux.Clause(attribute="Horsepower"), lux.Clause(attribute="Acceleration")])
@@ -262,8 +263,8 @@ def test_interestingness_0_2_0():
     assert interestingness(df.recommendation["Generalize"][0], df) != None
 
 
-def test_interestingness_0_2_1():
-    df = pd.read_csv("lux/data/car.csv")
+def test_interestingness_0_2_1(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     df.set_intent(
diff --git a/tests/test_maintainence.py b/tests/test_maintainence.py
index 1c2137ca..4527c21d 100644
--- a/tests/test_maintainence.py
+++ b/tests/test_maintainence.py
@@ -19,32 +19,32 @@
 from lux.vis.Vis import Vis
 
 
-def test_metadata_subsequent_display():
-    df = pd.read_csv("lux/data/car.csv")
+def test_metadata_subsequent_display(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
     df._repr_html_()
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
 
 
-def test_metadata_subsequent_vis():
-    df = pd.read_csv("lux/data/car.csv")
+def test_metadata_subsequent_vis(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
     vis = Vis(["Acceleration", "Horsepower"], df)
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
 
 
-def test_metadata_inplace_operation():
-    df = pd.read_csv("lux/data/car.csv")
+def test_metadata_inplace_operation(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
     df.dropna(inplace=True)
     assert df._metadata_fresh == False, "Failed to expire metadata after in-place Pandas operation"
 
 
-def test_metadata_new_df_operation():
-    df = pd.read_csv("lux/data/car.csv")
+def test_metadata_new_df_operation(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._metadata_fresh == True, "Failed to maintain metadata after display df"
     df[["MilesPerGal", "Acceleration"]]
@@ -53,7 +53,7 @@ def test_metadata_new_df_operation():
     assert not hasattr(df2, "_metadata_fresh")
 
 
-def test_metadata_column_group_reset_df():
+def test_metadata_column_group_reset_df(global_var):
     df = pd.read_csv("lux/data/car.csv")
     assert not hasattr(df, "_metadata_fresh")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
@@ -70,8 +70,8 @@ def test_metadata_column_group_reset_df():
         assert rec.mark == "bar", "Column Group not displaying bar charts"
 
 
-def test_recs_inplace_operation():
-    df = pd.read_csv("lux/data/car.csv")
+def test_recs_inplace_operation(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._recs_fresh == True, "Failed to maintain recommendation after display df"
     assert len(df.recommendation["Occurrence"]) == 4
diff --git a/tests/test_nan.py b/tests/test_nan.py
index df2b8e9f..f91f7c0c 100644
--- a/tests/test_nan.py
+++ b/tests/test_nan.py
@@ -20,8 +20,8 @@
 from lux.vis.Vis import Vis
 
 
-def test_nan_column():
-    df = pd.read_csv("lux/data/college.csv")
+def test_nan_column(global_var):
+    df = pytest.college_df
     df["Geography"] = np.nan
     df._repr_html_()
     for visList in df.recommendation.keys():
diff --git a/tests/test_pandas.py b/tests/test_pandas.py
index 60f6a91c..c3895779 100644
--- a/tests/test_pandas.py
+++ b/tests/test_pandas.py
@@ -28,8 +28,8 @@
 #     assert series.name == "Weight", "Pandas Series original `name` property not retained."
 
 
-def test_head_tail():
-    df = pd.read_csv("lux/data/car.csv")
+def test_head_tail(global_var):
+    df = pytest.car_df
     df._repr_html_()
     assert df._message.to_html() == ""
     df.head()._repr_html_()
diff --git a/tests/test_pandas_coverage.py b/tests/test_pandas_coverage.py
index d5ebfeb3..e224e404 100644
--- a/tests/test_pandas_coverage.py
+++ b/tests/test_pandas_coverage.py
@@ -21,7 +21,7 @@
 ###################
 
 
-def test_deepcopy():
+def test_deepcopy(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df._repr_html_()
@@ -30,7 +30,7 @@ def test_deepcopy():
     check_metadata_equal(df, saved_df)
 
 
-def test_rename_inplace():
+def test_rename_inplace(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df._repr_html_()
@@ -65,7 +65,7 @@ def test_rename_inplace():
     assert df.pre_aggregated == new_df.pre_aggregated
 
 
-def test_rename():
+def test_rename(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df._repr_html_()
@@ -95,7 +95,7 @@ def test_rename():
     assert df.pre_aggregated == new_df.pre_aggregated
 
 
-def test_rename3():
+def test_rename3(global_var):
 
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
@@ -122,7 +122,7 @@ def test_rename3():
     assert "col2" in list(df.cardinality.keys())
 
 
-def test_concat():
+def test_concat(global_var):
 
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
@@ -136,7 +136,7 @@ def test_concat():
     assert len(new_df.cardinality) == 5
 
 
-def test_groupby_agg():
+def test_groupby_agg(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.groupby("Year").agg(sum)
@@ -145,20 +145,20 @@ def test_groupby_agg():
     assert len(new_df.cardinality) == 7
 
 
-def test_qcut():
+def test_qcut(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df["Weight"] = pd.qcut(df["Weight"], q=3)
     df._repr_html_()
 
 
-def test_cut():
+def test_cut(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Weight"] = pd.cut(df["Weight"], bins=[0, 2500, 7500, 10000], labels=["small", "medium", "large"])
     df._repr_html_()
 
 
-def test_groupby_agg_very_small():
+def test_groupby_agg_very_small(global_var):
 
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
@@ -168,7 +168,7 @@ def test_groupby_agg_very_small():
     assert len(new_df.cardinality) == 7
 
 
-# def test_groupby_multi_index():
+# def test_groupby_multi_index(global_var):
 #     url = 'https://github.com/lux-org/lux-datasets/blob/master/data/cars.csv?raw=true'
 #     df = pd.read_csv(url)
 #     df["Year"] = pd.to_datetime(df["Year"], format='%Y')
@@ -178,7 +178,7 @@ def test_groupby_agg_very_small():
 #     assert len(new_df.cardinality) == 7 # TODO
 
 
-def test_query():
+def test_query(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.query("Weight > 3000")
@@ -192,7 +192,7 @@ def test_query():
     assert len(new_df.cardinality) == 10
 
 
-def test_pop():
+def test_pop(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df.pop("Weight")
@@ -206,7 +206,7 @@ def test_pop():
     assert len(df.cardinality) == 9
 
 
-def test_transform():
+def test_transform(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.iloc[:, 1:].groupby("Origin").transform(sum)
@@ -215,7 +215,7 @@ def test_transform():
     assert len(new_df.cardinality) == 7
 
 
-def test_get_group():
+def test_get_group(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     gbobj = df.groupby("Origin")
@@ -230,7 +230,7 @@ def test_get_group():
     assert len(new_df.cardinality) == 10
 
 
-def test_applymap():
+def test_applymap(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     mapping = {"USA": 0, "Europe": 1, "Japan": 2}
@@ -245,7 +245,7 @@ def test_applymap():
     assert len(df.cardinality) == 10
 
 
-def test_strcat():
+def test_strcat(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df["combined"] = df["Origin"].str.cat(df["Brand"], sep=", ")
@@ -259,7 +259,7 @@ def test_strcat():
     assert len(df.cardinality) == 11
 
 
-def test_named_agg():
+def test_named_agg(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.groupby("Brand").agg(
@@ -272,7 +272,7 @@ def test_named_agg():
     assert len(new_df.cardinality) == 4
 
 
-def test_change_dtype():
+def test_change_dtype(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df["Cylinders"] = pd.Series(df["Cylinders"], dtype="Int64")
@@ -286,7 +286,7 @@ def test_change_dtype():
     assert len(df.data_type_lookup) == 10
 
 
-def test_get_dummies():
+def test_get_dummies(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = pd.get_dummies(df)
@@ -300,7 +300,7 @@ def test_get_dummies():
     assert len(new_df.data_type_lookup) == 339
 
 
-def test_drop():
+def test_drop(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.drop([0, 1, 2], axis="rows")
@@ -315,7 +315,7 @@ def test_drop():
     assert len(new_df2.cardinality) == 7
 
 
-def test_merge():
+def test_merge(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.drop([0, 1, 2], axis="rows")
@@ -330,7 +330,7 @@ def test_merge():
     assert len(new_df2.cardinality) == 11
 
 
-def test_prefix():
+def test_prefix(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.add_prefix("1_")
@@ -345,7 +345,7 @@ def test_prefix():
     assert new_df.cardinality["1_Name"] == 300
 
 
-def test_loc():
+def test_loc(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.loc[:, "Displacement":"Origin"]
@@ -374,7 +374,7 @@ def test_loc():
     assert len(new_df.cardinality) == 3
 
 
-def test_iloc():
+def test_iloc(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     new_df = df.iloc[:, 3:9]
@@ -472,7 +472,7 @@ def compare_vis(vis1, vis2):
 ################
 
 
-def test_df_to_series():
+def test_df_to_series(global_var):
     # Ensure metadata is kept when going from df to series
     df = pd.read_csv("lux/data/car.csv")
     df._repr_html_()  # compute metadata
@@ -503,7 +503,7 @@ def test_df_to_series():
     assert series.name == "Weight", "Pandas Series original `name` property not retained."
 
 
-def test_value_counts():
+def test_value_counts(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df._repr_html_()  # compute metadata
     assert df.cardinality is not None
@@ -533,7 +533,7 @@ def test_value_counts():
     assert series.name == "Weight", "Pandas Series original `name` property not retained."
 
 
-def test_str_replace():
+def test_str_replace(global_var):
     df = pd.read_csv("lux/data/car.csv")
     df._repr_html_()  # compute metadata
     assert df.cardinality is not None
diff --git a/tests/test_parser.py b/tests/test_parser.py
index b37e5db7..a59be1ec 100644
--- a/tests/test_parser.py
+++ b/tests/test_parser.py
@@ -17,88 +17,97 @@
 import pytest
 
 
-def test_case1():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent(["Horsepower"])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == "Horsepower"
-
-
-def test_case2():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent(["Horsepower", lux.Clause("MilesPerGal", channel="x")])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == "Horsepower"
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "MilesPerGal"
-
-
-def test_case3():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent(["Horsepower", "Origin=USA"])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == "Horsepower"
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "Origin"
-    assert ldf._intent[1].value == "USA"
-
-
-def test_case4():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent(["Horsepower", "Origin=USA|Japan"])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == "Horsepower"
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "Origin"
-    assert ldf._intent[1].value == ["USA", "Japan"]
-
-
-def test_case5():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent([["Horsepower", "MilesPerGal", "Weight"], "Origin=USA"])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == ["Horsepower", "MilesPerGal", "Weight"]
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "Origin"
-    assert ldf._intent[1].value == "USA"
-
-    ldf.set_intent(["Horsepower|MilesPerGal|Weight", "Origin=USA"])
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == ["Horsepower", "MilesPerGal", "Weight"]
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "Origin"
-    assert ldf._intent[1].value == "USA"
-
-
-def test_case6():
-    ldf = pd.read_csv("lux/data/car.csv")
-    ldf.set_intent(["Horsepower", "Origin=?"])
-    ldf._repr_html_()
-    assert type(ldf._intent[0]) is lux.Clause
-    assert ldf._intent[0].attribute == "Horsepower"
-    assert type(ldf._intent[1]) is lux.Clause
-    assert ldf._intent[1].attribute == "Origin"
-    assert ldf._intent[1].value == ["USA", "Japan", "Europe"]
-
-
-def test_case7():
-    df = pd.read_csv("lux/data/car.csv")
+def test_case1(global_var):
+    df = pytest.car_df
+    df.set_intent(["Horsepower"])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == "Horsepower"
+    df.clear_intent()
+
+
+def test_case2(global_var):
+    df = pytest.car_df
+    df.set_intent(["Horsepower", lux.Clause("MilesPerGal", channel="x")])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == "Horsepower"
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "MilesPerGal"
+    df.clear_intent()
+
+
+def test_case3(global_var):
+    df = pytest.car_df
+    df.set_intent(["Horsepower", "Origin=USA"])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == "Horsepower"
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "Origin"
+    assert df._intent[1].value == "USA"
+    df.clear_intent()
+
+
+def test_case4(global_var):
+    df = pytest.car_df
+    df.set_intent(["Horsepower", "Origin=USA|Japan"])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == "Horsepower"
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "Origin"
+    assert df._intent[1].value == ["USA", "Japan"]
+    df.clear_intent()
+
+
+def test_case5(global_var):
+    df = pytest.car_df
+    df.set_intent([["Horsepower", "MilesPerGal", "Weight"], "Origin=USA"])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == ["Horsepower", "MilesPerGal", "Weight"]
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "Origin"
+    assert df._intent[1].value == "USA"
+
+    df.set_intent(["Horsepower|MilesPerGal|Weight", "Origin=USA"])
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == ["Horsepower", "MilesPerGal", "Weight"]
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "Origin"
+    assert df._intent[1].value == "USA"
+    df.clear_intent()
+
+
+def test_case6(global_var):
+    df = pytest.car_df
+    df.set_intent(["Horsepower", "Origin=?"])
+    df._repr_html_()
+    assert type(df._intent[0]) is lux.Clause
+    assert df._intent[0].attribute == "Horsepower"
+    assert type(df._intent[1]) is lux.Clause
+    assert df._intent[1].attribute == "Origin"
+    assert df._intent[1].value == ["USA", "Japan", "Europe"]
+    df.clear_intent()
+
+
+def test_case7(global_var):
+    df = pytest.car_df
     df.intent = [["Horsepower", "MilesPerGal", "Acceleration"], "Origin"]
     df._repr_html_()
     assert len(df.current_vis) == 3
+    df.clear_intent()
 
 
-def test_validator_invalid_value():
-    df = pd.read_csv("lux/data/college.csv")
+def test_validator_invalid_value(global_var):
+    df = pytest.college_df
     with pytest.warns(
         UserWarning,
         match="The input value 'bob' does not exist for the attribute 'Region' for the DataFrame.",
     ):
         df.intent = ["Region=bob"]
 
+    df.clear_intent()
+
 
-def test_validator_invalid_filter():
-    df = pd.read_csv("lux/data/college.csv")
+def test_validator_invalid_filter(global_var):
+    df = pytest.college_df
 
     with pytest.raises(KeyError, match="'New England'"):
         with pytest.warns(
@@ -108,8 +117,8 @@ def test_validator_invalid_filter():
             df.intent = ["New England", "Southeast", "Far West"]
 
 
-def test_validator_invalid_attribute():
-    df = pd.read_csv("lux/data/college.csv")
+def test_validator_invalid_attribute(global_var):
+    df = pytest.college_df
     with pytest.raises(KeyError, match="'blah'"):
         with pytest.warns(
             UserWarning, match="The input attribute 'blah' does not exist in the DataFrame."
diff --git a/tests/test_performance.py b/tests/test_performance.py
index 66a9bd6b..4e557075 100644
--- a/tests/test_performance.py
+++ b/tests/test_performance.py
@@ -17,9 +17,10 @@
 import pandas as pd
 import time
 
+
 # To run the script and see the printed result, run:
 # python -m pytest -s tests/test_performance.py
-def test_q1_performance_census():
+def test_q1_performance_census(global_var):
     df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/census.csv?raw=true")
     tic = time.perf_counter()
     df._repr_html_()
diff --git a/tests/test_type.py b/tests/test_type.py
index aa1b3b53..4c53656a 100644
--- a/tests/test_type.py
+++ b/tests/test_type.py
@@ -16,10 +16,10 @@
 import pytest
 import pandas as pd
 
+
 # Suite of test that checks if data_type inferred correctly by Lux
 def test_check_cars():
     df = pd.read_csv("lux/data/car.csv")
-    df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df.maintain_metadata()
     assert df.data_type_lookup["Name"] == "nominal"
     assert df.data_type_lookup["MilesPerGal"] == "quantitative"
@@ -54,10 +54,7 @@ def test_check_str_id():
 
 
 def test_check_hpi():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/hpi.csv?raw=true").head(
-        10
-    )
-
+    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/hpi.csv?raw=true")
     df.maintain_metadata()
 
     assert df.data_type_lookup == {
diff --git a/tests/test_vis.py b/tests/test_vis.py
index 122c1e3c..f8273cdb 100644
--- a/tests/test_vis.py
+++ b/tests/test_vis.py
@@ -19,22 +19,23 @@
 from lux.vis.Vis import Vis
 
 
-def test_vis():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vis(global_var):
+    df = pytest.olympic
     vis = Vis(["Height", "SportType=Ball"], df)
     assert vis.get_attr_by_attr_name("Height")[0].bin_size != 0
     assert vis.get_attr_by_attr_name("Record")[0].aggregation == "count"
 
 
-def test_vis_set_specs():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vis_set_specs(global_var):
+    df = pytest.olympic
     vis = Vis(["Height", "SportType=Ball"], df)
     vis.set_intent(["Height", "SportType=Ice"])
     assert vis.get_attr_by_attr_name("SportType")[0].value == "Ice"
+    df.clear_intent()
 
 
-def test_vis_collection():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vis_collection(global_var):
+    df = pytest.olympic
     vlist = VisList(["Height", "SportType=Ball", "?"], df)
     vis_with_year = list(filter(lambda x: x.get_attr_by_attr_name("Year") != [], vlist))[0]
     assert vis_with_year.get_attr_by_channel("x")[0].attribute == "Year"
@@ -44,22 +45,23 @@ def test_vis_collection():
     assert len(vlist) == len(df.columns) - 1  # remove 1 for vis with for same attribute
 
 
-def test_vis_collection_set_intent():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vis_collection_set_intent(global_var):
+    df = pytest.olympic
     vlist = VisList(["Height", "SportType=Ice", "?"], df)
     vlist.set_intent(["Height", "SportType=Boat", "?"])
     for v in vlist._collection:
         filter_vspec = list(filter(lambda x: x.channel == "", v._inferred_intent))[0]
         assert filter_vspec.value == "Boat"
+    df.clear_intent()
 
 
-def test_custom_plot_setting():
+def test_custom_plot_setting(global_var):
     def change_color_make_transparent_add_title(chart):
         chart = chart.configure_mark(color="green", opacity=0.2)
         chart.title = "Test Title"
         return chart
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     df.plot_config = change_color_make_transparent_add_title
     df._repr_html_()
     config_mark_addition = 'chart = chart.configure_mark(color="green", opacity=0.2)'
@@ -69,37 +71,38 @@ def change_color_make_transparent_add_title(chart):
     assert title_addition in exported_code_str
 
 
-def test_remove():
-    df = pd.read_csv("lux/data/car.csv")
+def test_remove(global_var):
+    df = pytest.car_df
     vis = Vis([lux.Clause("Horsepower"), lux.Clause("Acceleration")], df)
     vis.remove_column_from_spec("Horsepower", remove_first=False)
     assert vis._inferred_intent[0].attribute == "Acceleration"
 
 
-def test_remove_identity():
-    df = pd.read_csv("lux/data/car.csv")
+def test_remove_identity(global_var):
+    df = pytest.car_df
     vis = Vis(["Horsepower", "Horsepower"], df)
     vis.remove_column_from_spec("Horsepower")
     assert vis._inferred_intent == [], "Remove all instances of Horsepower"
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vis = Vis(["Horsepower", "Horsepower"], df)
     vis.remove_column_from_spec("Horsepower", remove_first=True)
     assert len(vis._inferred_intent) == 1, "Remove only 1 instances of Horsepower"
     assert vis._inferred_intent[0].attribute == "Horsepower", "Remove only 1 instances of Horsepower"
 
 
-def test_refresh_collection():
-    df = pd.read_csv("lux/data/car.csv")
+def test_refresh_collection(global_var):
+    df = pytest.car_df
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     df.set_intent([lux.Clause(attribute="Acceleration"), lux.Clause(attribute="Horsepower")])
     df._repr_html_()
     enhanceCollection = df.recommendation["Enhance"]
     enhanceCollection.refresh_source(df[df["Origin"] == "USA"])
+    df.clear_intent()
 
 
-def test_vis_custom_aggregation_as_str():
-    df = pd.read_csv("lux/data/college.csv")
+def test_vis_custom_aggregation_as_str(global_var):
+    df = pytest.college_df
     import numpy as np
 
     vis = Vis(["HighestDegree", lux.Clause("AverageCost", aggregation="max")], df)
@@ -107,8 +110,8 @@ def test_vis_custom_aggregation_as_str():
     assert vis.get_attr_by_data_model("measure")[0]._aggregation_name == "max"
 
 
-def test_vis_custom_aggregation_as_numpy_func():
-    df = pd.read_csv("lux/data/college.csv")
+def test_vis_custom_aggregation_as_numpy_func(global_var):
+    df = pytest.college_df
     from lux.vis.Vis import Vis
     import numpy as np
 
@@ -117,9 +120,8 @@ def test_vis_custom_aggregation_as_numpy_func():
     assert vis.get_attr_by_data_model("measure")[0]._aggregation_name == "ptp"
 
 
-def test_vis_collection_via_list_of_vis():
-    url = "https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true"
-    df = pd.read_csv(url)
+def test_vis_collection_via_list_of_vis(global_var):
+    df = pytest.olympic
     # change pandas dtype for the column "Year" to datetype
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
     from lux.vis.VisList import VisList
@@ -133,15 +135,15 @@ def test_vis_collection_via_list_of_vis():
     assert len(vlist) == 5
 
 
-def test_vis_to_Altair_basic_df():
-    df = pd.read_csv("lux/data/car.csv")
+def test_vis_to_Altair_basic_df(global_var):
+    df = pytest.car_df
     vis = Vis(["Weight", "Horsepower"], df)
     code = vis.to_Altair()
     assert "alt.Chart(df)" in code, "Unable to export to Altair"
 
 
-def test_vis_to_Altair_custom_named_df():
-    df = pd.read_csv("lux/data/car.csv")
+def test_vis_to_Altair_custom_named_df(global_var):
+    df = pytest.car_df
     some_weirdly_named_df = df.dropna()
     vis = Vis(["Weight", "Horsepower"], some_weirdly_named_df)
     code = vis.to_Altair()
@@ -150,8 +152,8 @@ def test_vis_to_Altair_custom_named_df():
     ), "Unable to export to Altair and detect custom df name"
 
 
-def test_vis_to_Altair_standalone():
-    df = pd.read_csv("lux/data/car.csv")
+def test_vis_to_Altair_standalone(global_var):
+    df = pytest.car_df
     vis = Vis(["Weight", "Horsepower"], df)
     code = vis.to_Altair(standalone=True)
     assert (
@@ -161,8 +163,8 @@ def test_vis_to_Altair_standalone():
     )
 
 
-def test_vis_list_custom_title_override():
-    df = pd.read_csv("https://github.com/lux-org/lux-datasets/blob/master/data/olympic.csv?raw=true")
+def test_vis_list_custom_title_override(global_var):
+    df = pytest.olympic
     df["Year"] = pd.to_datetime(df["Year"], format="%Y")
 
     vcLst = []
@@ -177,10 +179,10 @@ def test_vis_list_custom_title_override():
         assert v.title == "overriding dummy title"
 
 
-def test_vis_set_intent():
+def test_vis_set_intent(global_var):
     from lux.vis.Vis import Vis
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vis = Vis(["Weight", "Horsepower"], df)
     vis._repr_html_()
     assert "Horsepower" in str(vis._code)
@@ -189,10 +191,10 @@ def test_vis_set_intent():
     assert "MilesPerGal" in str(vis._code)
 
 
-def test_vis_list_set_intent():
+def test_vis_list_set_intent(global_var):
     from lux.vis.VisList import VisList
 
-    df = pd.read_csv("lux/data/car.csv")
+    df = pytest.car_df
     vislist = VisList(["Horsepower", "?"], df)
     vislist._repr_html_()
     for vis in vislist: