def dict_to_json_file(dict_obj,
directory_path,
filename,
remove_file_extension=True):
"""
Writes a dict to a json file.
Args:
dict_obj: dict
Dictionary object.
directory_path: string
Given path that already exists.
filename: string
Json file's name.
"""
directory_path = correct_directory_path(directory_path)
check_if_directory_exists(directory_path)
filename = convert_to_filename(filename,
remove_file_extension=remove_file_extension)
with open(f'{directory_path}{filename}.json', 'w',
encoding='utf-8') as outfile:
json.dump(dict_obj, outfile, ensure_ascii=False, indent=2)
def pickle_object_to_file(obj,
directory_path,
filename,
remove_file_extension=True):
"""
Writes the object to a pickle file.
Args:
obj: any object
Any python object that can be pickled.
directory_path: string
Given path that already exists.
filename: string
Pickle file's name.
"""
try:
directory_path = correct_directory_path(directory_path)
check_if_directory_exists(directory_path)
# Ensures no file extensions in filename
filename = convert_to_filename(
filename, remove_file_extension=remove_file_extension)
file_dir = f'{directory_path}{filename}.pkl'
list_pickle = open(file_dir, 'wb')
pickle.dump(obj, list_pickle)
finally:
list_pickle.close()
return file_dir
def write_object_text_to_file(obj,
directory_path,
filename,
remove_file_extension=True):
"""
Writes the object's string representation to a text file.
Args:
obj: any
Any object that has a string 'repr'.
directory_path: string
Given path that already exists.
filename: string
Text file's name.
"""
directory_path = correct_directory_path(directory_path)
check_if_directory_exists(directory_path)
filename = convert_to_filename(filename,
remove_file_extension=remove_file_extension)
file_dir = f'{directory_path}{filename}.txt'
f = open(file_dir, 'w')
f.write('obj = ' + repr(obj) + '\n')
f.close()
def save_plot(self, filename="Unknown filename", sub_dir=""):
"""
Checks the passed data to see if a plot can be saved; raises
an error if it can't.
Args:
filename: string
If set to 'None' will default to a pre-defined string;
unless it is set to an actual filename.
sub_dir: string
Specify the sub directory to append to the pre-defined folder path.
"""
if not isinstance(sub_dir, str):
raise TypeError(
f"Expected param 'sub_dir' to be type string! Was found to be {type(sub_dir)}"
)
# Create the png to save
create_plt_png(self.folder_path, sub_dir,
convert_to_filename(filename))
def save_plot(self,
sub_dir="",
filename="Unknown filename",
suppress_runtime_errors=True):
"""
Checks the passed data to see if a plot can be saved; raises
an error if it can't.
Args:
df: pd.Dataframe
Pandas DataFrame object
df_features: DataFrameTypes from eflow
DataFrameTypes object.
filename: string
If set to 'None' will default to a pre-defined string;
unless it is set to an actual filename.
sub_dir: string
Specify the sub directory to append to the pre-defined folder path.
suppress_runtime_errors: bool
If set to true; when generating any graphs will suppress any runtime
errors so the program can keep running.
"""
try:
if not isinstance(sub_dir, str):
raise TypeError(
f"Expected param 'sub_dir' to be type string! Was found to be {type(sub_dir)}"
)
# Create the png to save
create_plt_png(self.folder_path, sub_dir,
convert_to_filename(filename))
# Always raise snapshot error
except SnapshotMismatchError as e:
raise e
except Exception as e:
plt.close("all")
if suppress_runtime_errors:
warnings.warn(
f"An error occured when trying to save the plot:\n{str(e)}",
RuntimeWarning)
else:
raise e
def save_table_as_plot(self,
table,
filename="Unknown filename",
sub_dir="",
show_index=False):
"""
Checks the passed data to see if a table can be saved as a plot;
raises an error if it can't.
Args:
table: pd.Dataframe
Dataframe object to convert to plot.
filename: string
If set to 'None' will default to a pre-defined string;
unless it is set to an actual filename.
sub_dir: string
Specify the sub directory to append to the pre-defined folder path.
show_index: bool
Show index of the saved dataframe.
"""
if not isinstance(sub_dir, str):
raise TypeError(
f"Expected param 'sub_dir' to be type string! Was found to be {type(sub_dir)}"
)
# Convert dataframe to plot
df_to_image(table,
self.folder_path,
sub_dir,
convert_to_filename(filename),
show_index=show_index,
format_float_pos=2)
def save_plot(self,
df,
df_features,
filename="Unknown filename",
sub_dir="",
dataframe_snapshot=True,
suppress_runtime_errors=True,
compare_shape=True,
compare_feature_names=True,
compare_random_values=True,
meta_data=True):
"""
Checks the passed data to see if a plot can be saved; raises
an error if it can't.
Args:
df: pd.Dataframe
Pandas DataFrame object
df_features: DataFrameTypes from eflow
DataFrameTypes object.
filename: string
If set to 'None' will default to a pre-defined string;
unless it is set to an actual filename.
sub_dir: string
Specify the sub directory to append to the pre-defined folder path.
dataframe_snapshot: bool
Boolean value to determine whether or not generate and compare a
snapshot of the dataframe in the dataset's directory structure.
Helps ensure that data generated in that directory is correctly
associated to a dataframe.
compare_shape: bool
When comparing and creating the dataframe snapshot of the data's
shape.
compare_feature_names: bool
When comparing and creating the dataframe snapshot of the data's
column names.
compare_random_values: bool
When comparing and creating the dataframe snapshot of the data
sudo random values.
meta_data:bool
If set to true then it will generate meta data on the dataframe.
suppress_runtime_errors: bool
If set to true; when generating any graphs will suppress any runtime
errors so the program can keep running.
"""
try:
if not isinstance(sub_dir, str):
raise TypeError(
f"Expected param 'sub_dir' to be type string! Was found to be {type(sub_dir)}"
)
# Check if dataframe matches saved snapshot; Creates file if needed
if dataframe_snapshot:
df_snapshot = DataFrameSnapshot(
compare_shape=compare_shape,
compare_feature_names=compare_feature_names,
compare_random_values=compare_random_values)
df_snapshot.check_create_snapshot(
df,
df_features,
directory_path=self.folder_path,
sub_dir=f"{sub_dir}/_Extras")
# Create the png to save
create_plt_png(self.folder_path, sub_dir,
convert_to_filename(filename))
if meta_data:
generate_meta_data(df, self.folder_path,
f"{sub_dir}" + "/_Extras")
# Always raise snapshot error
except SnapshotMismatchError as e:
raise e
except Exception as e:
plt.close()
if suppress_runtime_errors:
warnings.warn(
f"An error occured when trying to save the plot:\n{str(e)}",
RuntimeWarning)
else:
raise e
def save_table_as_plot(self,
df,
sub_dir="",
filename="Unknown filename",
suppress_runtime_errors=True,
show_index=False,
format_float_pos=2):
"""
Checks the passed data to see if a table can be saved as a plot;
raises an error if it can't.
Args:
df: pd.Dataframe
Pandas DataFrame object of all data.
sub_dir:
Directory to generate on top of the already made directory (self.folder_path).
overwrite_full_path:
A pre defined directory that already exists.
df_features: DataFrameTypes from eflow
DataFrameTypes object.
table: pd.Dataframe
Dataframe object to convert to plot.
filename: string
If set to 'None' will default to a pre-defined string;
unless it is set to an actual filename.
sub_dir: string
Specify the sub directory to append to the pre-defined folder path.
suppress_runtime_errors: bool
If set to true; when generating any graphs will suppress any runtime
errors so the program can keep running.
meta_data:bool
If set to true then it will generate meta data on the dataframe.
show_index: bool
Show index of the saved dataframe.
"""
try:
if not isinstance(sub_dir, str):
raise TypeError(
f"Expected param 'sub_dir' to be type string! Was found to be {type(sub_dir)}"
)
# Convert dataframe to plot
df_to_image(df,
self.folder_path,
sub_dir,
convert_to_filename(filename),
show_index=show_index,
format_float_pos=format_float_pos)
# Always raise snapshot error
except SnapshotMismatchError as e:
raise e
except Exception as e:
plt.close("all")
if suppress_runtime_errors:
warnings.warn(
f"An error occured when trying to save the table as a plot:\n{str(e)}",
RuntimeWarning)
else:
raise e