def main(date, delete, keywords=[], byLeaf=True, saveProportion=0.5):
"""
Generates ML training and testing data from extracted CSV files
:param date: (string) Data collection date YYYY_MMDD
:param delete: (boolean) Determines whether or not to delete the existing
training/testing data files
:param keywords: (list of strings) Data filename keywords
:param byLeaf: (boolean) Should we separate the train/test data
by leaf, or should we randomly separate
the data according to a set proportion?
:param saveProportion: (float) Amount of each CSV file to save as training
and testing data.
:return: (None)
"""
# Get the data files we will be looking at
dataPath = DATA_DIRECTORIES[date]
dataFilenames = FileIO.getDatafileNames(dataPath, keywords)
# If desired, remove the old training data and start fresh
if delete:
mlDataPath = DATA_DIRECTORIES[date+"_ML"]
trainingDataPath = os.path.join(mlDataPath, TRAINING_DATA_PATH)
testingDataPath = os.path.join(mlDataPath, TESTING_DATA_PATH)
sampleCountsPath = os.path.join(mlDataPath, SAMPLE_COUNTS_PATH)
if os.path.exists(trainingDataPath):
os.remove(trainingDataPath)
if os.path.exists(testingDataPath):
os.remove(testingDataPath)
if os.path.exists(sampleCountsPath):
os.remove(sampleCountsPath)
# Consolidate the CSV files into training and testing data
(train_X, train_y, test_X, test_y) = DataManipulation.separateTrainTest(dataPath,
dataFilenames,
byLeaf=byLeaf,
saveProportion=saveProportion)
# Save the training and testing data in the proper spot
FileIO.saveTrainingData(date, train_X, train_y)
FileIO.saveTestingData(date, test_X, test_y)
