def addTrainingData(labeled_pairs, data_model, training_data=[]):
"""
Appends training data to the training data collection.
"""
fields = data_model['fields']
examples = [record_pair for example
in labeled_pairs.values()
for record_pair in example]
new_training_data = numpy.empty(len(examples),
dtype=training_data.dtype)
new_training_data['label'] = ([0] * len(labeled_pairs[0])
+ [1] * len(labeled_pairs[1]))
new_training_data['field_distances'], _ = core.buildRecordDistances(examples, fields)
training_data = numpy.append(training_data, new_training_data)
return training_data
def addTrainingData(labeled_pairs, data_model, training_data=[]):
fields = data_model['fields']
n_distinct_pairs, n_dupe_pairs = len(labeled_pairs[0]), len(labeled_pairs[1])
new_training_data = numpy.zeros(n_distinct_pairs + n_dupe_pairs,
dtype=training_data.dtype)
labels = labeled_pairs.keys()
examples = [record_pair for example in labeled_pairs.values() for record_pair in example]
new_training_data['label'] = [labels[0]] * n_distinct_pairs + [labels[1]] * n_dupe_pairs
new_training_data['field_distances'] = core.buildRecordDistances(examples, fields)[0]
training_data = numpy.append(training_data, new_training_data)
return training_data
