def test_binning():
df = pd.read_csv("lux/data/car.csv")
df.set_context([lux.Spec(attribute="Horsepower")])
PandasExecutor.execute(df.current_view, df)
nbins = list(filter(lambda x: x.bin_size != 0,
df.current_view[0].spec_lst))[0].bin_size
assert len(df.current_view[0].data) == nbins
def custom(ldf):
"""
Generates user-defined vis based on the intent.
Parameters
----------
ldf : lux.core.frame
LuxDataFrame with underspecified intent.
Returns
-------
recommendations : Dict[str,obj]
object with a collection of visualizations that result from the Distribution action.
"""
recommendation = {
"action": "Current Vis",
"description": "Shows the list of visualizations generated based on user specified intent",
}
recommendation["collection"] = ldf.current_vis
vlist = ldf.current_vis
PandasExecutor.execute(vlist, ldf)
for vis in vlist:
vis.score = interestingness(vis, ldf)
# ldf.clear_intent()
vlist.sort(remove_invalid=True)
return recommendation
def test_binning():
df = pd.read_csv("lux/data/car.csv")
df.setContext([lux.Spec(attribute="Horsepower")])
PandasExecutor.execute(df.viewCollection, df)
Nbins = list(filter(lambda x: x.binSize != 0,
df.viewCollection[0].specLst))[0].binSize
assert len(df.viewCollection[0].data) == Nbins
def user_defined(ldf):
'''
Generates user-defined views based on the context.
Parameters
----------
ldf : lux.luxDataFrame.LuxDataFrame
LuxDataFrame with underspecified context.
Returns
-------
recommendations : Dict[str,obj]
object with a collection of visualizations that result from the Distribution action.
'''
recommendation = {
"action": "Current Views",
"description": "Shows a view collection defined by the context"
}
recommendation["collection"] = ldf.current_view
vc = ldf.current_view
PandasExecutor.execute(vc, ldf)
for view in vc:
view.score = interestingness(view, ldf)
# ldf.clear_context()
vc.sort(remove_invalid=True)
return recommendation
def load(self, ldf) -> View:
"""
Loading the data into the view by instantiating the specification and populating the view based on the data, effectively "materializing" the view.
Parameters
----------
ldf : LuxDataframe
Input Dataframe to be attached to the view
Returns
-------
View
Complete View with fully-specified fields
See Also
--------
lux.view.ViewCollection.load
"""
from lux.compiler.Parser import Parser
from lux.compiler.Validator import Validator
from lux.compiler.Compiler import Compiler
from lux.executor.PandasExecutor import PandasExecutor #TODO: temporary (generalize to executor)
#TODO: handle case when user input vanilla Pandas dataframe
self.specLst = Parser.parse(self.specLst)
Validator.validateSpec(self.specLst, ldf)
vc = Compiler.compile(ldf,
ldf.context, [self],
enumerateCollection=False)
PandasExecutor.execute(vc, ldf)
return vc[0]
def __init__(self, *args, **kw):
from lux.executor.PandasExecutor import PandasExecutor
self._history = History()
self._intent = []
self._recommendation = {}
self._saved_export = None
self._current_vis = []
self._prev = None
super(LuxDataFrame, self).__init__(*args, **kw)
self.executor_type = "Pandas"
self.executor = PandasExecutor()
self.SQLconnection = ""
self.table_name = ""
self._sampled = None
self._toggle_pandas_display = True
self._message = Message()
self._pandas_only = False
# Metadata
self.data_type_lookup = None
self.data_type = None
self.data_model_lookup = None
self.data_model = None
self.unique_values = None
self.cardinality = None
self._min_max = None
self.pre_aggregated = None
warnings.formatwarning = lux.warning_format
def test_binning():
df = pd.read_csv("lux/data/car.csv")
df.set_intent([lux.Clause(attribute="Horsepower")])
PandasExecutor.execute(df.current_vis, df)
nbins = list(
filter(lambda x: x.bin_size != 0,
df.current_vis[0]._inferred_intent))[0].bin_size
assert len(df.current_vis[0].data) == nbins
def test_lazyExecution():
df = pd.read_csv("lux/data/car.csv")
df.setContext([
lux.Spec(attribute="Horsepower", aggregation="mean"),
lux.Spec(attribute="Origin")
])
# Check data field in view is empty before calling executor
assert df.viewCollection[0].data == None
PandasExecutor.execute(df.viewCollection, df)
assert type(df.viewCollection[0].data
) == lux.luxDataFrame.LuxDataframe.LuxDataFrame
def test_histogram_code_export(global_var):
df = pytest.car_df
vis = Vis([lux.Clause("Horsepower")], df)
PandasExecutor.execute([vis], df)
code = vis.to_code("python")
try:
exec(code, globals())
create_chart_data(df, vis)
except:
assert False
def test_color_barchart_code_export(global_var):
df = pytest.car_df
vis = Vis([lux.Clause("Origin"), lux.Clause("Cylinders")], df)
PandasExecutor.execute([vis], df)
code = vis.to_code("python")
try:
exec(code, globals())
create_chart_data(df, vis)
except:
assert False
def test_lazy_execution():
df = pd.read_csv("lux/data/car.csv")
df.set_intent([
lux.Clause(attribute="Horsepower", aggregation="mean"),
lux.Clause(attribute="Origin")
])
# Check data field in vis is empty before calling executor
assert df.current_vis[0].data == None
PandasExecutor.execute(df.current_vis, df)
assert type(
df.current_vis[0].data) == lux.luxDataFrame.LuxDataframe.LuxDataFrame
def test_lazy_execution():
df = pd.read_csv("lux/data/car.csv")
intent = [
lux.Clause(attribute="Horsepower", aggregation="mean"),
lux.Clause(attribute="Origin"),
]
vis = Vis(intent)
# Check data field in vis is empty before calling executor
assert vis.data is None
PandasExecutor.execute([vis], df)
assert type(vis.data) == lux.core.frame.LuxDataFrame
def test_period_filter():
ldf = pd.read_csv("lux/data/car.csv")
ldf["Year"] = pd.to_datetime(ldf["Year"], format="%Y")
ldf["Year"] = pd.DatetimeIndex(ldf["Year"]).to_period(freq="A")
ldf.set_intent([lux.Clause(attribute="Acceleration"), lux.Clause(attribute="Horsepower")])
PandasExecutor.execute(ldf.current_vis, ldf)
ldf._repr_html_()
assert isinstance(ldf.recommendation["Filter"][2]._inferred_intent[2].value, pd.Period)
def test_exclude_attribute():
df = pd.read_csv("lux/data/car.csv")
df.setContext(
[lux.Spec("?", exclude=["Name", "Year"]),
lux.Spec("Horsepower")])
view = df.viewCollection[0]
view.data = df
PandasExecutor.executeFilter(view)
for vc in df.viewCollection:
assert (vc.getAttrByChannel("x")[0].attribute != "Year")
assert (vc.getAttrByChannel("x")[0].attribute != "Name")
assert (vc.getAttrByChannel("y")[0].attribute != "Year")
assert (vc.getAttrByChannel("y")[0].attribute != "Year")
def test_exclude_attribute():
df = pd.read_csv("lux/data/car.csv")
df.set_intent(
[lux.Clause("?", exclude=["Name", "Year"]),
lux.Clause("Horsepower")])
vis = df.current_vis[0]
vis.data = df
PandasExecutor.execute_filter(vis)
for vc in df.current_vis:
assert (vc.get_attr_by_channel("x")[0].attribute != "Year")
assert (vc.get_attr_by_channel("x")[0].attribute != "Name")
assert (vc.get_attr_by_channel("y")[0].attribute != "Year")
assert (vc.get_attr_by_channel("y")[0].attribute != "Year")
def test_filter():
df = pd.read_csv("lux/data/car.csv")
# change pandas dtype for the column "Year" to datetype
df["Year"] = pd.to_datetime(df["Year"], format="%Y")
intent = [
lux.Clause(attribute="Horsepower"),
lux.Clause(attribute="Year"),
lux.Clause(attribute="Origin", filter_op="=", value="USA"),
]
vis = Vis(intent, df)
vis._vis_data = df
PandasExecutor.execute_filter(vis)
assert len(vis.data) == len(df[df["Origin"] == "USA"])
def set_executor_type(self, exe):
if (exe =="SQL"):
import pkgutil
if (pkgutil.find_loader("psycopg2") is None):
raise ImportError("psycopg2 is not installed. Run `pip install psycopg2' to install psycopg2 to enable the Postgres connection.")
else:
import psycopg2
from lux.executor.SQLExecutor import SQLExecutor
self.executor = SQLExecutor
else:
from lux.executor.PandasExecutor import PandasExecutor
self.executor = PandasExecutor()
self.executor_type = exe
def test_filter():
df = pd.read_csv("lux/data/car.csv")
df["Year"] = pd.to_datetime(
df["Year"],
format='%Y') # change pandas dtype for the column "Year" to datetype
df.set_intent([
lux.Clause(attribute="Horsepower"),
lux.Clause(attribute="Year"),
lux.Clause(attribute="Origin", filter_op="=", value="USA")
])
vis = df.current_vis[0]
vis.data = df
PandasExecutor.execute_filter(vis)
assert len(vis.data) == len(df[df["Origin"] == "USA"])
def test_filter():
df = pd.read_csv("lux/data/car.csv")
df["Year"] = pd.to_datetime(
df["Year"],
format='%Y') # change pandas dtype for the column "Year" to datetype
df.setContext([
lux.Spec(attribute="Horsepower"),
lux.Spec(attribute="Year"),
lux.Spec(attribute="Origin", filterOp="=", value="USA")
])
view = df.viewCollection[0]
view.data = df
PandasExecutor.executeFilter(view)
assert len(view.data) == len(df[df["Origin"] == "USA"])
def test_period_to_altair():
chart = None
df = pd.read_csv("lux/data/car.csv")
df["Year"] = pd.to_datetime(df["Year"], format="%Y")
df["Year"] = pd.DatetimeIndex(df["Year"]).to_period(freq="A")
df.set_intent([lux.Clause(attribute="Acceleration"), lux.Clause(attribute="Horsepower")])
PandasExecutor.execute(df.current_vis, df)
df._repr_html_()
exported_code = df.recommendation["Filter"][2].to_Altair()
assert "Year = 1972" in exported_code
def test_period_filter():
ldf = pd.read_csv("lux/data/car.csv")
ldf["Year"] = pd.to_datetime(ldf["Year"], format='%Y')
ldf["Year"] = pd.DatetimeIndex(ldf["Year"]).to_period(freq='A')
ldf.set_context(
[lux.Spec(attribute="Acceleration"),
lux.Spec(attribute="Horsepower")])
PandasExecutor.execute(ldf.current_view, ldf)
ldf.show_more()
assert isinstance(ldf.recommendation['Filter'][2].spec_lst[2].value,
pd.Period)
def test_color_scatter_code_export(global_var):
df = pytest.car_df
vis = Vis([
lux.Clause("Horsepower"),
lux.Clause("Acceleration"),
lux.Clause("Origin")
], df)
PandasExecutor.execute([vis], df)
code = vis.to_code("python")
try:
exec(code, globals())
create_chart_data(df, vis)
except:
assert False
def __init__(self, *args, **kw):
self._history = History()
self._intent = []
self._inferred_intent = []
self._recommendation = {}
self._saved_export = None
self._current_vis = []
self._prev = None
self._widget = None
super(LuxDataFrame, self).__init__(*args, **kw)
self.table_name = ""
if lux.config.SQLconnection == "":
from lux.executor.PandasExecutor import PandasExecutor
lux.config.executor = PandasExecutor()
else:
from lux.executor.SQLExecutor import SQLExecutor
lux.config.executor = SQLExecutor()
self._sampled = None
self._toggle_pandas_display = True
self._message = Message()
self._pandas_only = False
# Metadata
self._data_type = {}
self.unique_values = None
self.cardinality = None
self._min_max = None
self.pre_aggregated = None
self._type_override = {}
warnings.formatwarning = lux.warning_format
def test_period_selection():
ldf = pd.read_csv("lux/data/car.csv")
ldf["Year"] = pd.to_datetime(ldf["Year"], format="%Y")
ldf["Year"] = pd.DatetimeIndex(ldf["Year"]).to_period(freq="A")
ldf.set_intent(
[
lux.Clause(attribute=["Horsepower", "Weight", "Acceleration"]),
lux.Clause(attribute="Year"),
]
)
PandasExecutor.execute(ldf.current_vis, ldf)
assert all([type(vlist.data) == lux.core.frame.LuxDataFrame for vlist in ldf.current_vis])
assert all(ldf.current_vis[2].data.columns == ["Year", "Acceleration"])
def test_period_to_altair():
chart = None
df = pd.read_csv("lux/data/car.csv")
df["Year"] = pd.to_datetime(df["Year"], format='%Y')
df["Year"] = pd.DatetimeIndex(df["Year"]).to_period(freq='A')
df.set_context(
[lux.Spec(attribute="Acceleration"),
lux.Spec(attribute="Horsepower")])
PandasExecutor.execute(df.current_view, df)
df.show_more()
exported_code = df.recommendation['Filter'][2].to_Altair()
assert 'Year = 1971' in exported_code
def test_heatmap_code_export(global_var):
df = pd.read_csv(
"https://raw.githubusercontent.com/lux-org/lux-datasets/master/data/airbnb_nyc.csv"
)
lux.config._heatmap_start = 100
vis = Vis(["price", "longitude"], df)
PandasExecutor.execute([vis], df)
code = vis.to_code("python")
try:
exec(code, globals())
create_chart_data(df, vis)
except:
assert False
lux.config._heatmap_start = 5000
def test_selection():
df = pd.read_csv("lux/data/car.csv")
df["Year"] = pd.to_datetime(
df["Year"],
format='%Y') # change pandas dtype for the column "Year" to datetype
df.setContext([
lux.Spec(attribute=["Horsepower", "Weight", "Acceleration"]),
lux.Spec(attribute="Year")
])
PandasExecutor.execute(df.viewCollection, df)
assert all([
type(vc.data) == lux.luxDataFrame.LuxDataframe.LuxDataFrame
for vc in df.viewCollection
])
assert all(df.viewCollection[2].data.columns == ["Year", 'Acceleration'])
def test_period_selection():
ldf = pd.read_csv("lux/data/car.csv")
ldf["Year"] = pd.to_datetime(ldf["Year"], format='%Y')
ldf["Year"] = pd.DatetimeIndex(ldf["Year"]).to_period(freq='A')
ldf.set_context([
lux.Spec(attribute=["Horsepower", "Weight", "Acceleration"]),
lux.Spec(attribute="Year")
])
PandasExecutor.execute(ldf.current_view, ldf)
assert all([
type(vc.data) == lux.luxDataFrame.LuxDataframe.LuxDataFrame
for vc in ldf.current_view
])
assert all(ldf.current_view[2].data.columns == ["Year", 'Acceleration'])
def test_filter_aggregation_fillzero_aligned():
df = pd.read_csv("lux/data/car.csv")
df.setContext([
lux.Spec(attribute="Cylinders"),
lux.Spec(attribute="MilesPerGal"),
lux.Spec("Origin=Japan")
])
PandasExecutor.execute(df.viewCollection, df)
result = df.viewCollection[0].data
externalValidation = df[df["Origin"] == "Japan"].groupby(
"Cylinders").mean()["MilesPerGal"]
assert result[result["Cylinders"] == 5]["MilesPerGal"].values[0] == 0
assert result[result["Cylinders"] == 8]["MilesPerGal"].values[0] == 0
assert result[result["Cylinders"] ==
3]["MilesPerGal"].values[0] == externalValidation[3]
assert result[result["Cylinders"] ==
4]["MilesPerGal"].values[0] == externalValidation[4]
assert result[result["Cylinders"] ==
6]["MilesPerGal"].values[0] == externalValidation[6]
def set_executor_type(self, exe):
if exe == "SQL":
from lux.executor.SQLExecutor import SQLExecutor
self.executor = SQLExecutor()
elif exe == "Pandas":
from lux.executor.PandasExecutor import PandasExecutor
self.SQLconnection = ""
self.executor = PandasExecutor()
else:
raise ValueError("Executor type must be either 'Pandas' or 'SQL'")
