Skip to content
Snippets Groups Projects
Commit bc6fcc14 authored by niklasfranz's avatar niklasfranz
Browse files

refactor: extensive logging, add param type hints

parent e853bbf7
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 97 additions and 21 deletions
src/Plotter.py
+ 97
− 21
View file @ bc6fcc14
......@@ -20,6 +20,9 @@ class Plotter:
logging.error("dataset parameter is not of type Dataset")
raise ValueError(f"{dataset} is not of type Dataset")
logging.debug("Plotter.__init__(): all params valid, dataset \
initialized")
self.ds = dataset
self.df = dataset.get_dataframe()
......@@ -45,7 +48,8 @@ class Plotter:
fig.set_figheight(parameters["plotter_fig_height"])
fig.set_figwidth(parameters["plotter_fig_width"])
def distribution_plot(self, target, styling_params={}) -> None:
def distribution_plot(self, target: str,
styling_params: dict = {}) -> None:
"""plot a distribution plot.
Args:
......@@ -70,6 +74,11 @@ class Plotter:
logging.error("parameter styling params should be a dict.")
raise ValueError("parameter styling params should be a dict.")
logging.debug("Plotter.distribution_plot(): all params valid,\
plotting plot_categorical_bar_chart(): \n \
target: {target} \n \
styling_params: {styling_params}")
# plotting the plot
fig, ax = plt.subplots()
grouped_data = self.df.groupby(target).size()
......@@ -84,7 +93,7 @@ class Plotter:
self.customize_plot(fig, ax, styling_params)
def plot_categorical_bar_chart(
self, category1, category2, styling_params={}
self, category1: str, category2: str, styling_params: str = {}
) -> None:
"""plot a categorical bar chart.
......@@ -99,33 +108,44 @@ class Plotter:
"""
# implementing sensible logging and error catching
if type(category1) != str:
logging.error("parameter category1 should be a string.")
logging.error("Plotter.plot_categorical_bar_chart(): parameter \
category1 should be a string.")
raise ValueError("parameter category1 should be a string.")
if not (category1 in self.df.columns):
logging.error(
"parameter category1 cannot be found in the dataset."
"Plotter.plot_categorical_bar_chart(): parameter category1 \
cannot be found in the dataset."
)
raise KeyError(
"parameter category1 cannot be found in the dataset."
)
if type(category2) != str:
logging.error("parameter category2 should be a string.")
logging.error("Plotter.plot_categorical_bar_chart(): parameter \
category2 should be a string.")
raise ValueError("parameter category2 should be a string.")
if not (category2 in self.df.columns):
logging.error(
"parameter category2 cannot be found in the dataset."
"Plotter.plot_categorical_bar_chart(): \
parameter category2 cannot be found in the dataset."
)
raise KeyError(
"parameter category2 cannot be found in the dataset."
)
if type(styling_params) != dict:
logging.error("parameter styling params should be a dict.")
logging.error("Plotter.plot_categorical_bar_chart(): \
parameter styling params should be a dict.")
raise ValueError("parameter styling params should be a dict.")
logging.debug("Plotter.plot_categorical_bar_chart(): all params valid,\
plotting plot_categorical_bar_chart(): \n \
category1: {category1} \n \
category2: {category2} \n \
styling_params: {styling_params}")
# plotting the plot
ct = pd.crosstab(self.df[category1], self.df[category2])
# Calculate percentages by row
......@@ -136,7 +156,7 @@ class Plotter:
ct_percent.plot(kind="bar", ax=ax)
def plot_categorical_boxplot(
self, target, category, styling_params={}
self, target: str, category: str, styling_params: dict = {}
) -> None:
"""plot a categorical boxplot.
......@@ -149,6 +169,39 @@ class Plotter:
Returns:
None
"""
# implementing sensible logging and error catching
if type(target) != str:
logging.error("Plotter.plot_categorical_boxplot(): parameter \
target should be a string.")
raise ValueError("parameter target should be a string.")
if not (target in self.df.columns):
logging.error("Plotter.plot_categorical_boxplot(): parameter \
target cannot be found in the dataset.")
raise KeyError("parameter target cannot be found in the dataset.")
if type(category) != str:
logging.error("Plotter.plot_categorical_boxplot(): parameter \
category should be a string.")
raise ValueError("parameter category should be a string.")
if not (category in self.df.columns):
logging.error("Plotter.plot_categorical_boxplot(): parameter \
category cannot be found in the dataset.")
raise KeyError(
"parameter category cannot be found in the dataset."
)
if type(styling_params) != dict:
logging.error("Plotter.plot_categorical_boxplot(): parameter \
styling params should be a dict.")
raise ValueError("parameter styling params should be a dict.")
logging.debug("Plotter.plot_categorical_boxplot(): all params valid, \
plotting plot_categorical_boxplot(): \n \
target: {target} \n \
category: {category} \n \
styling_params: {styling_params}")
fig, ax = plt.subplots()
ax.set_ylabel("Percent")
......@@ -156,8 +209,8 @@ class Plotter:
sns.boxplot(x=category, y=target, data=self.df, palette="rainbow")
def plot_categorical_histplot(
self, target, category, styling_params={}, bins=30
) -> None:
self, target: str, category: str,
styling_params: dict = {}, bins: int = 30) -> None:
"""plot a categorical hisplot.
Args:
......@@ -171,27 +224,38 @@ class Plotter:
"""
# implementing sensible logging and error catching
if type(target) != str:
logging.error("parameter target should be a string.")
logging.error("Plotter.plot_categorical_histplot(): parameter \
target should be a string.")
raise ValueError("parameter target should be a string.")
if not (target in self.df.columns):
logging.error("parameter target cannot be found in the dataset.")
logging.error("Plotter.plot_categorical_histplot(): parameter \
target cannot be found in the dataset.")
raise KeyError("parameter target cannot be found in the dataset.")
if type(category) != str:
logging.error("parameter category should be a string.")
logging.error("Plotter.plot_categorical_histplot(): parameter \
category should be a string.")
raise ValueError("parameter category should be a string.")
if not (category in self.df.columns):
logging.error("parameter category cannot be found in the dataset.")
logging.error("Plotter.plot_categorical_histplot(): parameter \
category cannot be found in the dataset.")
raise KeyError(
"parameter category cannot be found in the dataset."
)
if type(styling_params) != dict:
logging.error("parameter styling params should be a dict.")
logging.error("Plotter.plot_categorical_histplot(): parameter \
styling params should be a dict.")
raise ValueError("parameter styling params should be a dict.")
logging.debug("Plotter.plot_categorical_histplot(): all params valid, \
plotting plot_categorical_histplot(): \n \
target: {target} \n \
category: {category} \n \
styling_params: {styling_params}")
# plotting the plot
uniques = self.ds.get_unique_column_values(category)
......@@ -213,7 +277,8 @@ class Plotter:
label_list.append(val)
ax.legend(labels=label_list)
def plot_scatterplot(self, target1, target2, styling_params={}) -> None:
def plot_scatterplot(self, target1: str, target2: str,
styling_params: dict = {}) -> None:
"""plot a scatterplot.
Args:
......@@ -228,25 +293,36 @@ class Plotter:
# implementing sensible logging and error catching
if type(target1) != str:
logging.error("parameter target1 should be a string.")
logging.error("Plotter.plot_scatterplot(): parameter target1 \
should be a string.")
raise ValueError("parameter target1 should be a string.")
if not (target1 in self.df.columns):
logging.error("parameter target1 cannot be found in the dataset.")
logging.error("Plotter.plot_scatterplot(): parameter target1 \
cannot be found in the dataset.")
raise KeyError("parameter target1 cannot be found in the dataset.")
if type(target2) != str:
logging.error("parameter target2 should be a string.")
logging.error("Plotter.plot_scatterplot(): parameter target2 \
should be a string.")
raise ValueError("parameter target2 should be a string.")
if not (target2 in self.df.columns):
logging.error("parameter target2 cannot be found in the dataset.")
logging.error("Plotter.plot_scatterplot(): parameter target2 \
cannot be found in the dataset.")
raise KeyError("parameter target2 cannot be found in the dataset.")
if type(styling_params) != dict:
logging.error("parameter styling params should be a dict.")
logging.error("Plotter.plot_scatterplot(): parameter styling \
params should be a dict.")
raise ValueError("parameter styling params should be a dict.")
logging.debug("Plotter.plot_scatterplot(): all params valid, plotting\
scatterplot using given parameters: \n \
target1: {target1} \n \
target2: {target2} \n \
styling_params: {styling_params}")
# plotting the plot
fig, ax = plt.subplots()
ax.set_xlabel(target1)
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment