def generate_report(self, name=None, level=None, notes=None):
self.reporting = ClassificationReporting(self.clf.best_estimator_,
self.X_train, self.X_test,
self.y_train, self.y_test,
self.n_classes)
if name:
self.reporting.set_name(name)
if level:
self.reporting.set_level(level)
if notes:
self.reporting.set_notes(notes)
self.report = self.reporting.create_report(output=True, show_roc=False)
# incase it's a function...
self.clf.best_params_["features__text__tfidf__tokenizer"] = \
pickle.dumps(self.clf.best_params_["features__text__tfidf__tokenizer"])
self.report["params"] = self.clf.best_params_
return self
class CustomGridSearch:
def __init__(self, pipeline, param_grid, n_classes, random, **kwargs):
self.clf = (RandomizedSearchCV if random else GridSearchCV)(pipeline,
param_grid,
**kwargs)
self.n_classes = n_classes
def set_data(self, X, y):
self.X_train, self.X_test, self.y_train, self.y_test = \
train_test_split(X, y, test_size=0.15, random_state=42)
return self
def fit(self):
self.clf.fit(self.X_train, self.y_train)
print("[GRIDSCORES]")
print(self.clf.grid_scores_)
print("[GRIDSCORES]")
return self
def generate_report(self, name=None, level=None, notes=None):
self.reporting = ClassificationReporting(self.clf.best_estimator_,
self.X_train, self.X_test,
self.y_train, self.y_test,
self.n_classes)
if name:
self.reporting.set_name(name)
if level:
self.reporting.set_level(level)
if notes:
self.reporting.set_notes(notes)
self.report = self.reporting.create_report(output=True, show_roc=False)
# incase it's a function...
self.clf.best_params_["features__text__tfidf__tokenizer"] = \
pickle.dumps(self.clf.best_params_["features__text__tfidf__tokenizer"])
self.report["params"] = self.clf.best_params_
return self
def write_to_mongo(self):
ClassifierAccess.write_report(self.report)
return self
def __repr__(self):
return self.clf.__repr__()
def __str__(self):
return self.clf.__str__()
class CustomGridSearch:
def __init__(self, pipeline, param_grid, n_classes, random, **kwargs):
self.clf = (RandomizedSearchCV if random else GridSearchCV)(pipeline, param_grid, **kwargs)
self.n_classes = n_classes
def set_data(self, X, y):
self.X_train, self.X_test, self.y_train, self.y_test = \
train_test_split(X, y, test_size=0.15, random_state=42)
return self
def fit(self):
self.clf.fit(self.X_train, self.y_train)
print("[GRIDSCORES]")
print(self.clf.grid_scores_)
print("[GRIDSCORES]")
return self
def generate_report(self, name=None, level=None, notes=None):
self.reporting = ClassificationReporting(
self.clf.best_estimator_, self.X_train, self.X_test, self.y_train, self.y_test, self.n_classes
)
if name:
self.reporting.set_name(name)
if level:
self.reporting.set_level(level)
if notes:
self.reporting.set_notes(notes)
self.report = self.reporting.create_report(output=True, show_roc=False)
# incase it's a function...
self.clf.best_params_["features__text__tfidf__tokenizer"] = \
pickle.dumps(self.clf.best_params_["features__text__tfidf__tokenizer"])
self.report["params"] = self.clf.best_params_
return self
def write_to_mongo(self):
ClassifierAccess.write_report(self.report)
return self
def __repr__(self):
return self.clf.__repr__()
def __str__(self):
return self.clf.__str__()
def generate_report(self, name=None, level=None, notes=None):
self.reporting = ClassificationReporting(
self.clf.best_estimator_, self.X_train, self.X_test, self.y_train, self.y_test, self.n_classes
)
if name:
self.reporting.set_name(name)
if level:
self.reporting.set_level(level)
if notes:
self.reporting.set_notes(notes)
self.report = self.reporting.create_report(output=True, show_roc=False)
# incase it's a function...
self.clf.best_params_["features__text__tfidf__tokenizer"] = \
pickle.dumps(self.clf.best_params_["features__text__tfidf__tokenizer"])
self.report["params"] = self.clf.best_params_
return self
