def visit(self, node: Node, path=None):
if path is None:
path = []
if isinstance(node, Statement):
node.scope = self.scopes[-1]
self.parents.append(node)
if 'visit_' + class_name(node) in dir(self):
getattr(self, 'visit_' + class_name(node))(node, path)
elif isinstance(node, (list, tuple)):
for i, v in enumerate(node):
v.path = Path(node, i)
self.visit(v, path + [v.path])
else:
getattr(self, 'enter_' + class_name(node))(node, path)
for f in node.FIELDS:
v = getattr(node, f)
if v:
v.path = Path(node, f)
self.visit(v, path + [v.path])
getattr(self, 'leave_' + class_name(node))(node, path)
if isinstance(node, Statement):
self.parents.pop()
def visit(self, node: Node):
if node == self.terminal:
raise Complete()
if isinstance(node, (list, tuple)):
for i, v in enumerate(node):
self.visit(v)
else:
getattr(self, 'enter_' + class_name(node))(node)
for f in node.FIELDS:
v = getattr(node, f)
if v:
self.visit(v)
getattr(self, 'leave_' + class_name(node))(node)
def __init__(self, classifier, datamanager, category):
self.now = str(datetime.now())
self.log_path = os.path.join(datamanager.PATHS["LOGS"], "grid_search",
class_name(classifier), category)
log_file = os.path.join(self.log_path, self.now + ".log")
super(GridSearch, self).__init__(classifier,
datamanager,
log_file=log_file)
self.category = category
self.grid_search_obj = None
def __init__(self, classifier, datamanager, category):
self.now = str(datetime.now())
self.log_path = os.path.join(datamanager.PATHS["LOGS"],
"grid_search",
class_name(classifier),
category)
log_file = os.path.join(self.log_path, self.now + ".log")
super(GridSearch, self).__init__(classifier, datamanager, log_file = log_file)
self.category = category
self.grid_search_obj = None
def visit(self, node: Node):
if isinstance(node, Statement):
self.parents.append(node)
if 'visit_' + class_name(node) in dir(self):
getattr(self, 'visit_' + class_name(node))(node)
elif isinstance(node, (list, tuple)):
for i, v in enumerate(node):
self.visit(v)
else:
getattr(self, 'enter_' + class_name(node))(node)
for f in node.FIELDS:
v = getattr(node, f)
if v:
self.visit(v)
getattr(self, 'leave_' + class_name(node))(node)
if isinstance(node, Statement):
self.parents.pop()
def evaluate(category, clf, datamanager, data=(None, None)):
"""Run evaluation of a classifier, for one category.
If data isn't set explicitly, the test set is
used by default.
"""
log_file = os.path.join(datamanager.PATHS["LOGS"], "evaluation", class_name(clf), category)
log_file = os.path.join(log_file, str(datetime.now()) + ".log")
vcd = VisualConceptDetection(None, datamanager, log_file=log_file)
clf = vcd.load_object("Classifier", category, clf)
vcd.classifier = clf
if (data[0] is None) or (data[1] is None):
return vcd.evaluate_test_set(category)
else:
return vcd.evaluate(X_test=data[0], y_test=data[1])
def evaluate(category, clf, datamanager, data=(None, None)):
"""Run evaluation of a classifier, for one category.
If data isn't set explicitly, the test set is
used by default.
"""
log_file = os.path.join(datamanager.PATHS["LOGS"], "evaluation",
class_name(clf), category)
log_file = os.path.join(log_file, str(datetime.now()) + ".log")
vcd = VisualConceptDetection(None, datamanager, log_file=log_file)
clf = vcd.load_object("Classifier", category, clf)
vcd.classifier = clf
if (data[0] is None) or (data[1] is None):
return vcd.evaluate_test_set(category)
else:
return vcd.evaluate(X_test=data[0], y_test=data[1])
def grid_search(self, params, X=None, y=None, weight_samples=False):
"""Perform a nested grid search.
Args:
params: Parameter hash, specifiying the parameter configurations
for the grid search (see http://scikit-learn.org/stable/modules/grid_search.html).
X: Matrix of sample data.
y: Vector of class labels.
weight_samples: Whether to use grid search with weighted
samples or not.
"""
self.logger.info("Grid search for %s at %s" %
(class_name(self.classifier), datetime.now()))
self.logger.info("Category: %s" % self.category)
self.logger.info("Parameters: %s" % params)
self.logger.info("Initial sample weighting: %s" % weight_samples)
result = self.search(params, X, y, weight_samples)
joblib.dump(result, os.path.join(self.log_path, self.now + ".grid"), compress=3)
return result
def grid_search(self, params, X=None, y=None, weight_samples=False):
"""Perform a nested grid search.
Args:
params: Parameter hash, specifiying the parameter configurations
for the grid search (see http://scikit-learn.org/stable/modules/grid_search.html).
X: Matrix of sample data.
y: Vector of class labels.
weight_samples: Whether to use grid search with weighted
samples or not.
"""
self.logger.info("Grid search for %s at %s" %
(class_name(self.classifier), datetime.now()))
self.logger.info("Category: %s" % self.category)
self.logger.info("Parameters: %s" % params)
self.logger.info("Initial sample weighting: %s" % weight_samples)
result = self.search(params, X, y, weight_samples)
joblib.dump(result,
os.path.join(self.log_path, self.now + ".grid"),
compress=3)
return result
def expr(self, expr: Expr) -> ir.Value:
return getattr(self, class_name(expr))(expr)
def statement(self, statement: Statement):
getattr(self, class_name(statement))(statement)