def main():
if not os.path.isdir(DATASETS):
raise FileNotFoundError('Folder given to save datasets does not exist.')
datasets = sys.argv[1:] # use datasets given by user
for month in set(datasets):
filename = download_data(month)
out_path = os.path.join(DATASETS, filename)
preprocess_data(filename, out_path)
print(f'Successfully downloaded and preproccessed {len(set(datasets))} file(s).')
def generate_clean_data():
train_data, test_data = get_raw_data.classify_data(ORIGINAL_FILE_PATH)
print("The data is being processed right now. Please wait....")
train_data = pre.preprocess_data(train_data)
train_data.to_csv("./data/train.csv",
sep=",",
encoding="utf-8",
index=False)
test_data = pre.preprocess_data(test_data)
test_data.to_csv("./data/test.csv", sep=",", encoding="utf-8", index=False)
use_classifier()
def run_mr(use_word_to_vec=True):
data, labels = data_importer.load_data_and_labels_mr()
word_to_index_mapping, vectors = w2v.load_mr_word_vectors()
train_data, train_labels, vocabulary, test_data, test_labels, embedding_vectors = \
pre.preprocess_data(data, labels, word_to_index_mapping, vectors)
if not use_word_to_vec:
embedding_vectors = None
t.train(train_data, train_labels, vocabulary, test_data, test_labels, embedding_vectors)
def test_compressing_without_output(self):
"""Test wheather compressed data would be the same after decompression if `output` parameter is not
passed"""
new_path = os.path.join(os.path.dirname(path),
'_temp_' + os.path.basename(path))
shutil.copyfile(
path, new_path) # copy file in order not to modify original file
file_path = preprocess_data(new_path)
self._test_compressing_helper(file_path)
def quick_explore(folder='folder_name'):
# Imports
import pandas as pd
import matplotlib.pyplot as plt
import os
from preprocessor import preprocess_data
directory = '../../data/' + folder + '/'
for filename in os.listdir(directory):
filepath = os.path.join(directory, filename)
# Read CSV files
if filename.endswith('.csv'):
raw_df = pd.read_csv(filepath, infer_datetime_format=True)
neg_index = 4
# Read Excel files
elif filename.endswith('.xlsx'):
raw_df = pd.read_excel(filepath, infer_datetime_format=True)
neg_index = 5
# Skip other file types
else:
continue
print()
print('--------------' * 8)
print('--------------' * 8)
print(filename[:-neg_index].upper())
print()
df, redund_dict = preprocess_data(raw_df)
return
def test_compressing_with_output(self):
"""Test wheather compressed data would be the same after decompression if `output` parameter is passed
"""
file_path = preprocess_data(path, '_temp', delete_original=False)
self._test_compressing_helper(file_path)
import os
os.chdir(os.path.dirname(os.path.abspath(__file__)))
#os.chdir('local\\directory')
import scraper
import preprocessor
import loader
data = scraper.scrape_data()
data = preprocessor.preprocess_data(data)
loader.gsheets_upload(data)
### saving local copy of data
#data.to_excel('./data/processed_data_extract.xlsx')
with open('./data/Time Dataset.json') as f:
dataset = json.load(f)
with open('./data/Time Vocabs.json') as f:
human_vocab, machine_vocab = json.load(f)
human_vocab_size = len(human_vocab)
machine_vocab_size = len(machine_vocab)
# number of training examples
m = len(dataset)
# tokenize the data using vocabularies
Tx = 41 # Max x sequence length
Ty = 5 # y sequence length
X, Y, Xoh, Yoh = preprocess_data(dataset, human_vocab, machine_vocab, Tx, Ty)
# Split dat 80-20 between training and test
train_size = int(0.8 * m)
Xoh_train = Xoh[:train_size]
Yoh_train = Yoh[:train_size]
Xoh_test = Xoh[train_size:]
Yoh_test = Yoh[train_size:]
# Check the code works:
# i = 4
# print('Input data point',str(i),'.\n')
# print('The data input is :',str(dataset[i][0]))
# print('The data output is :',str(dataset[i][1]))
# print()
# print('The tokenized input is :',str(X[i]))
