mechanoChemML.workflows.mr_learning.mrnn_utility

Module Contents

Classes

sys_args

System configurations

Functions

parse_sys_args()	Command line input variables
notebook_args(args)	Additional configurations
split_data(datax, datay, split_ratio=[‘0.6’, ‘0.25’, ‘0.15’])	Split data based on different ratios
get_package_version(tf_version)	get the major and minor version of tensor flow
getlist_str(option, sep=’,’, chars=None)	Return a list from a ConfigParser option. By default,
getlist_int(option, sep=’,’, chars=None)	Return a list from a ConfigParser option. By default,
getlist_float(option, sep=’,’, chars=None)	Return a list from a ConfigParser option. By default,
get_now()	Return the now string: yyyy-mm-dd-hh-mm-ss
exe_cmd(cmd)	execute shell cmd
get_dummy_data(num)	get dummy_data for num of fields
csvDf(dat, **kwargs)	Generate dataframe based on csv file
read_config_file(configfile, print_keys=False)	read configuration file and modify the related path
read_one_vtk(filepath, scalar=’’, vector=’’)	Read one VTK file
read_psi_me_from_mechanical_data(file_path)	Read psi_me from mechanical_data: for temporary label of vtk datatype
load_data_from_npy_for_label_shift_frame(config, dataset_frame, normalization_flag=True, verbose=0)	Load data from npy for label shift frame
load_data_from_vtk_for_label_shift_frame(config, dataset_frame, normalization_flag=True, verbose=0)	Load data from vtk for label shift frame
load_all_data_from_vtk_database(config, normalization_flag=True, verbose=0)	Load all data from vtk database
load_all_data_from_npy_database(config, normalization_flag=True, verbose=0)	load all data from npy database
load_data_from_vtk_database(config, normalization_flag=True, verbose=0)	Load data from vtk database with hard coded label mechanical_data.txt
load_all_data(config, args)	load csv, image, url etc data to the main code
read_csv_fields(file_path, fields, sep=’,’)	Read CSV fields information
dataset_pop_list(data_set, pop_list)	Pop a list of index from the dataset
norm(x, train_stats, DataNormOption=0)	Different data normalization scheme
prepare_data_from_csv_file(config, normalization_flag=True, verbose=0)	load the desired fields from the csv file, not full list
load_all_data_from_csv(config, normalization_flag=True, verbose=0)	Load all the data from a csv file
inspect_cnn_features(model, config, test_dataset, savefig=False)	Output intermediate CNN results after each layer
generate_dummy_dataset(old_config)	based on the label list, generate dummy dataset
special_input_case(inputs, input_case=’’)	deal with special case for old inputs:

mechanoChemML.workflows.mr_learning.mrnn_utility.parse_sys_args()

Command line input variables

class mechanoChemML.workflows.mr_learning.mrnn_utility.sys_args

System configurations

mechanoChemML.workflows.mr_learning.mrnn_utility.notebook_args(args)

Additional configurations

mechanoChemML.workflows.mr_learning.mrnn_utility.split_data(datax, datay, split_ratio=['0.6', '0.25', '0.15'])

Split data based on different ratios

mechanoChemML.workflows.mr_learning.mrnn_utility.get_package_version(tf_version)

get the major and minor version of tensor flow

mechanoChemML.workflows.mr_learning.mrnn_utility.getlist_str(option, sep=', ', chars=None)

Return a list from a ConfigParser option. By default,

split on a comma and strip whitespaces.

mechanoChemML.workflows.mr_learning.mrnn_utility.getlist_int(option, sep=', ', chars=None)

Return a list from a ConfigParser option. By default,

split on a comma and strip whitespaces.

mechanoChemML.workflows.mr_learning.mrnn_utility.getlist_float(option, sep=', ', chars=None)

Return a list from a ConfigParser option. By default,

split on a comma and strip whitespaces.

mechanoChemML.workflows.mr_learning.mrnn_utility.get_now()

Return the now string: yyyy-mm-dd-hh-mm-ss

mechanoChemML.workflows.mr_learning.mrnn_utility.exe_cmd(cmd)

execute shell cmd

mechanoChemML.workflows.mr_learning.mrnn_utility.get_dummy_data(num)

get dummy_data for num of fields

mechanoChemML.workflows.mr_learning.mrnn_utility.csvDf(dat, **kwargs)

Generate dataframe based on csv file

mechanoChemML.workflows.mr_learning.mrnn_utility.read_config_file(configfile, print_keys=False)

read configuration file and modify the related path

mechanoChemML.workflows.mr_learning.mrnn_utility.read_one_vtk(filepath, scalar='', vector='')

Read one VTK file

mechanoChemML.workflows.mr_learning.mrnn_utility.read_psi_me_from_mechanical_data(file_path)

Read psi_me from mechanical_data: for temporary label of vtk datatype

This function should be a standalone script to prepare the label and features for vtk files.

mechanoChemML.workflows.mr_learning.mrnn_utility.load_data_from_npy_for_label_shift_frame(config, dataset_frame, normalization_flag=True, verbose=0)

Load data from npy for label shift frame

mechanoChemML.workflows.mr_learning.mrnn_utility.load_data_from_vtk_for_label_shift_frame(config, dataset_frame, normalization_flag=True, verbose=0)

Load data from vtk for label shift frame

mechanoChemML.workflows.mr_learning.mrnn_utility.load_all_data_from_vtk_database(config, normalization_flag=True, verbose=0)

Load all data from vtk database

mechanoChemML.workflows.mr_learning.mrnn_utility.load_all_data_from_npy_database(config, normalization_flag=True, verbose=0)

load all data from npy database

mechanoChemML.workflows.mr_learning.mrnn_utility.load_data_from_vtk_database(config, normalization_flag=True, verbose=0)

Load data from vtk database with hard coded label mechanical_data.txt

mechanoChemML.workflows.mr_learning.mrnn_utility.load_all_data(config, args)

load csv, image, url etc data to the main code

mechanoChemML.workflows.mr_learning.mrnn_utility.read_csv_fields(file_path, fields, sep=', ')

Read CSV fields information

mechanoChemML.workflows.mr_learning.mrnn_utility.dataset_pop_list(data_set, pop_list)

Pop a list of index from the dataset

mechanoChemML.workflows.mr_learning.mrnn_utility.norm(x, train_stats, DataNormOption=0)

Different data normalization scheme

mechanoChemML.workflows.mr_learning.mrnn_utility.prepare_data_from_csv_file(config, normalization_flag=True, verbose=0)

load the desired fields from the csv file, not full list split the data based on the label fields split the data to three different set [train, validation, test]

mechanoChemML.workflows.mr_learning.mrnn_utility.load_all_data_from_csv(config, normalization_flag=True, verbose=0)

Load all the data from a csv file

mechanoChemML.workflows.mr_learning.mrnn_utility.inspect_cnn_features(model, config, test_dataset, savefig=False)

Output intermediate CNN results after each layer

mechanoChemML.workflows.mr_learning.mrnn_utility.generate_dummy_dataset(old_config)

based on the label list, generate dummy dataset

mechanoChemML.workflows.mr_learning.mrnn_utility.special_input_case(inputs, input_case='')

deal with special case for old inputs: for example, another DNN is for the frame 800, which has input only F11, F12, F21, F22, whereas, the current model might have additional microstructure features, thus, the input needs to be sliced to fit for the old DNN.