def perform_global_dataframe_analysis(df: Optional[pd.DataFrame]) -> dict:
"""
Returns a python dict containing global information about a pandas DataFrame :
Number of features, Number of observations, missing values...
Parameters
----------
df : pd.DataFrame
The dataframe that will be used to compute global information.
Returns
-------
global_d : dict
dictionary that contains an ensemble of global information about the input dataframe.
"""
if df is None:
return dict()
missing_values = df.isna().sum().sum()
global_d = {
'number of features': len(df.columns),
'number of observations': df.shape[0],
'missing values': missing_values,
'% missing values': missing_values / (df.shape[0] * df.shape[1]),
}
for stat in global_d.keys():
if stat == 'number of observations':
global_d[stat] = int(global_d[stat]) # Keeping the exact number
elif isinstance(global_d[stat], float):
global_d[stat] = round_to_k(global_d[stat], 3)
replace_dict_values(global_d, display_value, ',', '.')
return global_d
def perform_univariate_dataframe_analysis(df: Optional[pd.DataFrame],
col_types: dict) -> dict:
"""
Returns a python dict containing information about each column of a pandas DataFrame.
The computed information depends on the type of the column.
Parameters
----------
df : pd.DataFrame
The dataframe on which the analysis will be performed
col_types : dict
Dict of types for each column
Returns
-------
d : dict
A dict containing each column as keys and the corresponding dict of information for each column as values.
"""
if df is None:
return dict()
d = df.describe().to_dict()
for col in df.columns:
if col_types[col] == VarType.TYPE_CAT:
d[col] = {
'distinct values': df[col].nunique(),
'missing values': df[col].isna().sum()
}
for col in d.keys():
for stat in d[col].keys():
if stat in ['count', 'distinct values']:
d[col][stat] = int(
d[col][stat]) # Keeping the exact number here
elif isinstance(d[col][stat], float):
d[col][stat] = round_to_k(d[col][stat],
3) # Rounding to 3 important figures
replace_dict_values(d, display_value, ',', '.')
return d
def display_model_performance(self):
"""
Displays the performance of the model. The metrics are computed using the config dict.
Metrics should be given as a list of dict. Each dict contains they following keys :
'path' (path to the metric function, ex: 'sklearn.metrics.mean_absolute_error'),
'name' (optional, name of the metric as displayed in the report),
and 'use_proba_values' (optional, possible values are False (default) or True
if the metric uses proba values instead of predicted values).
For example :
config['metrics'] = [
{
'path': 'sklearn.metrics.mean_squared_error',
'name': 'Mean absolute error', # Optional : name that will be displayed next to the metric
'y_pred': 'predicted_values' # Optional
},
{
'path': 'Scoring_AP.utils.lift10', # Custom function path
'name': 'Lift10',
'y_pred': 'proba_values' # Use proba values instead of predicted values
}
]
"""
if self.y_test is None:
logging.info(
"No labels given for test set. Skipping model performance part"
)
return
print_md("### Univariate analysis of target variable")
df = pd.concat([
pd.DataFrame({
self.target_name: self.y_pred
}).assign(_dataset="pred"),
pd.DataFrame({
self.target_name: self.y_test
}).assign(_dataset="true") if self.y_test is not None else None
])
self._perform_and_display_analysis_univariate(
df=df,
col_splitter="_dataset",
split_values=["pred", "true"],
names=["Prediction values", "True values"],
group_id='target-distribution')
if 'metrics' not in self.config.keys():
logging.info(
"No 'metrics' key found in report config dict. Skipping model performance part."
)
return
print_md("### Metrics")
for metric in self.config['metrics']:
if 'name' not in metric.keys():
metric['name'] = metric['path']
if metric['path'] in ['confusion_matrix', 'sklearn.metrics.confusion_matrix'] or \
metric['name'] == 'confusion_matrix':
print_md(f"**{metric['name']} :**")
print_html(
convert_fig_to_html(
generate_confusion_matrix_plot(y_true=self.y_test,
y_pred=self.y_pred)))
else:
try:
metric_fn = get_callable(path=metric['path'])
# Look if we should use proba values instead of predicted values
if 'use_proba_values' in metric.keys(
) and metric['use_proba_values'] is True:
y_pred = self.explainer.proba_values
else:
y_pred = self.y_pred
res = metric_fn(self.y_test, y_pred)
except Exception as e:
logging.info(
f"Could not compute following metric : {metric['path']}. \n{e}"
)
continue
if isinstance(res, Number):
res = display_value(round_to_k(res, 3))
print_md(f"**{metric['name']} :** {res}")
elif isinstance(res, (list, tuple, np.ndarray)):
print_md(f"**{metric['name']} :**")
print_html(
pd.DataFrame(res).to_html(classes="greyGridTable"))
elif isinstance(res, str):
print_md(f"**{metric['name']} :**")
print_html(f"<pre>{res}</pre>")
else:
logging.info(
f"Could not compute following metric : {metric['path']}. \n"
f"Result of type {res} cannot be displayed")
print_md('---')
def test_round_to_k_1(self):
x = 123456789
expected_r_x = 123000000
assert round_to_k(x, 3) == expected_r_x
def test_round_to_k_6(self):
x = 0.0000123456789
expected_r_x = 0.0000123
assert round_to_k(x, 3) == expected_r_x
def test_round_to_k_4(self):
x = 123.456789
expected_r_x = 123
assert round_to_k(x, 3) == expected_r_x
def test_round_to_k_3(self):
x = 123456789
expected_r_x = 100000000
assert round_to_k(x, 1) == expected_r_x
def __init__(self, explainer):
"""
Init on class instantiation, everything to be able to run the app on server.
Parameters
----------
explainer : SmartExplainer
SmartExplainer object
"""
# APP
self.server = Flask(__name__)
self.app = dash.Dash(
server=self.server,
external_stylesheets=[dbc.themes.BOOTSTRAP],
)
self.app.title = 'Shapash Monitor'
if explainer.title_story:
self.app.title += ' - ' + explainer.title_story
self.explainer = explainer
# SETTINGS
self.logo = self.app.get_asset_url('shapash-fond-fonce.png')
self.color = '#f4c000'
self.bkg_color = "#343736"
self.settings_ini = {
'rows': 1000,
'points': 1000,
'violin': 10,
'features': 20,
}
self.settings = self.settings_ini.copy()
self.predict_col = ['_predict_']
self.explainer.features_imp = self.explainer.state.compute_features_import(
self.explainer.contributions)
if self.explainer._case == 'classification':
self.label = self.explainer.check_label_name(
len(self.explainer._classes) - 1, 'num')[1]
self.selected_feature = self.explainer.features_imp[-1].idxmax()
self.max_threshold = int(
max([
x.applymap(lambda x: round_to_k(x, k=1)).max().max()
for x in self.explainer.contributions
]))
else:
self.label = None
self.selected_feature = self.explainer.features_imp.idxmax()
self.max_threshold = int(
self.explainer.contributions.applymap(
lambda x: round_to_k(x, k=1)).max().max())
self.list_index = []
self.subset = None
# DATA
self.dataframe = pd.DataFrame()
self.round_dataframe = pd.DataFrame()
self.init_data()
# COMPONENTS
self.components = {
'menu': {},
'table': {},
'graph': {},
'filter': {},
'settings': {}
}
self.init_components()
# LAYOUT
self.skeleton = {'navbar': {}, 'body': {}}
self.make_skeleton()
self.app.layout = html.Div(
[self.skeleton['navbar'], self.skeleton['body']])
# CALLBACK
self.callback_fullscreen_buttons()
self.init_callback_settings()
self.callback_generator()