def makeDataDict(filenames):
"""
Returns a dictionary having data combined of all the CSV files
ARGS:
filename{list} -- a list containg names of all the files
RETURNS:
-- a dictionary having the data
"""
dictionary = defaultdict(list)
for i, filename in enumerate(filenames):
print('{}. working with {}, please wait...'.format(
i, filename.decode('utf-8')))
# calling csv2dict(), to convert the csv file to dictionary
data = helper_functions.csv2dict('Data/csvfiles/' +
str(filename.decode('utf-8')))
# accessing the key aftershocksyn to check for unique values, similar like (set(list[1, 1, 0, 2, 3])) -> outputs [1, 1, 0, 2, 3]
grid_aftershock_count = np.double(data['aftershocksyn'])
# no use of if
#if len(np.unique(grid_aftershock_count)) < 2:
# continue
temp = grid_aftershock_count.tolist()
# adding a (key, value) to the testingSet
dictionary['aftershocksyn'].extend(temp)
# now adding remaining columns
for column in columns:
dictionary[column].extend(np.double(data[column]))
return dictionary
def singleshot():
fn = str(random.randint(1, 25))
filename = 'rt-data-io/' + fn + '.csv'
weightFile = 'Data/weights.h5'
predFile = 'Data/singleCSV/singlePred.h5'
columns = [
'stresses_full_xx', 'stresses_full_yy', 'stresses_full_xy',
'stresses_full_xz', 'stresses_full_yz', 'stresses_full_zz'
]
testFile = 'single.h5'
dictionary = defaultdict(list)
print('working with {},...'.format(filename.split('/')[-1]))
data = helper_functions.csv2dict(filename)
grid_aftershock_count = np.double(data['aftershocksyn'])
temp = grid_aftershock_count.tolist()
dictionary['aftershocksyn'].extend(temp)
for column in columns:
dictionary[column].extend(np.double(data[column]))
columns.append('aftershocksyn')
helper_functions.dict2HDF('single.h5', columns, dictionary)
features_in = [
'stresses_full_xx', 'stresses_full_yy', 'stresses_full_xy',
'stresses_full_xz', 'stresses_full_yz', 'stresses_full_zz'
]
features_out = 'aftershocksyn'
model = helper_functions.createModel()
model.load_weights(weightFile)
X, y = helper_functions.loadDataFromHDF(testFile, features_in,
features_out)
y_pred = model.predict(X)
helper_functions.writeHDF(predFile, X, y)
auc = sklearn.metrics.roc_auc_score(y, y_pred)
return auc
import csv
##input a file name of csv format from the singleCsv folder in the Data directory.
filename = 'rt-data-io/incoming.csv'
weightFile = 'Data/weights.h5'
predFile = 'Data/singleCSV/singlePred.h5'
columns = ['stresses_full_xx', 'stresses_full_yy',
'stresses_full_xy', 'stresses_full_xz',
'stresses_full_yz','stresses_full_zz']
testFile = 'single.h5'
dictionary = defaultdict(list)
print('working with {},...'.format(filename.split('/')[-1]))
data = helper_functions.csv2dict(filename)
grid_aftershock_count = np.double(data['aftershocksyn'])
temp = grid_aftershock_count.tolist()
dictionary['aftershocksyn'].extend(temp)
for column in columns:
dictionary[column].extend(np.double(data[column]))
columns.append('aftershocksyn')
helper_functions.dict2HDF('single.h5', columns, dictionary)
features_in = ['stresses_full_xx',
'stresses_full_yy',
'stresses_full_xy',
'stresses_full_xz',
'stresses_full_yz',
'stresses_full_zz']
if values2['type'] == "crystal":
space = space + 0.25
if test_atom not in values["connected"]:
dist = hf.distance_checker(values2["coor"],
xyz_list[-1])
if dist < space + 0.25:
break
atom_dict = {**atom_dict, **temp_atom_dict}
break
return atom_dict
if __name__ == "__main__":
# Having a global dict with bonding lengths improves speed a lot
global bond_len_dict
bond_len_dict = hf.csv2dict("bonding_distances.csv")
build = hf.y2true(
input("Create new crystal (y) or use existing file (n)?: "))
if build:
a = float(input("Specify lattice constant (in Ångström): "))
atom_a = input("Element for first element type: ")
atom_b = input("Element for second element type: ")
diameter = float(input("Diameter of quantum dot (in unit cells): "))
atom_dict = crystal_builder(a, atom_a, atom_b, diameter)
else:
crystal_file = input(
"Crystal file to use (don't write the file extension): ") + ".xyz"
atom_dict = crystal_reader(crystal_file)
foldername = input("Save in folder (or main): ")
if foldername == "main":