config_template = 'session_config_template.json'

configs_location = 'session_configs'

config_name = 'conf'

corpus_importer_key = 'corpus_importer'

corpus_identifier_key = 'corpus_identifier'

categories_identifier_key = 'categories_identifier'

preprocessor_key = 'preprocessor'

noise_remover_key = 'noise_remover'

stopword_remover_key = 'stopword_remover'

stopwords_identifier_key = 'stopwords_identifier'

additional_stopwords_key = 'additional_stopwords'

lemmatizer_key = 'lemmatizer'

train_test_splitter_key = 'train_test_splitter'

train_test_split_ratio_key = 'train_test_split_ratio'

train_test_split_random_state_key = 'train_test_split_random_state'

vectorizer_key = 'vectorizer'

vec_model_id_key = 'vec_model_id'

word2vec_size_key = 'word2vec_size'

word2vec_window_key = 'word2vec_window'

word2vec_min_count_key = 'word2vec_min_count'

word2vec_workers_key = 'word2vec_workers'

vectorizer_gensim_word2vec = 'gensim-word2vec'

word_vec_to_doc_vec_key = 'word-vec_to_doc-vec'

feature_vec_dim_key = 'feature_vec_dim'

classifier_key = 'classifier'

classifier_keras_nn = 'keras_nn'

classification_interpreter_key = 'classification_interpreter'

keras_nn_default_output_layer_size_key = 'keras_nn_default_output_layer_size'

keras_nn_model_id_key = 'keras_nn_model_id'

classification_interpreter_custom1 = 'custom1'

classification_interpreter_custom2 = 'custom2'

classification_interpreter_output_threshold_key = 'classification_interpreter_output_threshold'

similarity_function_key = 'similarity_function'

keras_nn_loss_key = 'keras_nn_loss'

keras_nn_optimizer_key = 'keras_nn_optimizer'

keras_nn_metrics_key = 'keras_nn_metrics'

keras_nn_epochs_key = 'keras_nn_epochs'

keras_nn_batch_size_key = 'keras_nn_batch_size'

keras_nn_default_activation_key = 'keras_nn_default_activation'

keras_nn_default_output_activation_key = 'keras_nn_default_output_activation'

keras_nn_default_layer_size_key = 'keras_nn_default_layer_size'

keras_nn_default_layer_type_key = 'keras_nn_default_layer_type'

keras_nn_default_layer_count_key = 'keras_nn_default_layer_count'

keras_nn_layers_key = 'keras_nn_layers'

# expects a config json list

# adds keras layers info to configs

# returns a list with new configs

@staticmethod

def add_keras_layers_info(configs):

new_configs = list()

for conf in configs:

new_conf = copy.deepcopy(conf)

classifier_type = new_conf[SessionConfigBuilderCustom1.classifier_key][0]

conf_has_output_dim = SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key in new_conf

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

if SessionConfigBuilderCustom1.keras_nn_layers_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_layers_key)

if SessionConfigBuilderCustom1.keras_nn_default_activation_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_activation_key)

if SessionConfigBuilderCustom1.keras_nn_default_output_activation_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_activation_key)

if SessionConfigBuilderCustom1.keras_nn_default_layer_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_layer_size_key)

if SessionConfigBuilderCustom1.keras_nn_default_layer_count_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_layer_count_key)

if conf_has_output_dim:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key)

new_configs.append(new_conf)

else:

conf_layers = list()

if SessionConfigBuilderCustom1.keras_nn_layers_key in new_conf:

conf_layers = new_conf[SessionConfigBuilderCustom1.keras_nn_layers_key]

output_dim = new_conf[SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key][0]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key)

# individual layers

idx = 0

n_custom_layers = len(conf_layers)

while idx < n_custom_layers:

new_c_conf = copy.deepcopy(new_conf)

layers_list = copy.deepcopy(conf_layers)

layers = layers_list.pop(idx)

last_layer = layers[len(layers)-1]

last_layer_out_dim = last_layer[1]

if last_layer_out_dim == output_dim:

new_c_conf[SessionConfigBuilderCustom1.keras_nn_layers_key] = [layers]

new_configs.append(new_c_conf)

idx = idx + 1

# constructed layers

if (

SessionConfigBuilderCustom1.keras_nn_layers_key in new_conf and

SessionConfigBuilderCustom1.keras_nn_default_output_activation_key in new_conf and

SessionConfigBuilderCustom1.keras_nn_default_layer_size_key in new_conf and

SessionConfigBuilderCustom1.keras_nn_default_layer_type_key in new_conf and

SessionConfigBuilderCustom1.keras_nn_default_layer_count_key in new_conf and

conf_has_output_dim

):

def_activations = new_conf[SessionConfigBuilderCustom1.keras_nn_default_activation_key]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_activation_key)

def_out_activations = new_conf[SessionConfigBuilderCustom1.keras_nn_default_output_activation_key]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_activation_key)

def_layer_sizes = new_conf[SessionConfigBuilderCustom1.keras_nn_default_layer_size_key]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_layer_size_key)

layer_types = new_conf[SessionConfigBuilderCustom1.keras_nn_default_layer_type_key]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_layer_type_key)

def_layer_counts = new_conf[SessionConfigBuilderCustom1.keras_nn_default_layer_count_key]

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_layer_count_key)

for activation in def_activations:

for out_activation in def_out_activations:

for layer_size in def_layer_sizes:

for layer_type in layer_types:

for layer_count in def_layer_counts:

new_c_conf = copy.deepcopy(new_conf)

layers = []

idx = 0

while idx < layer_count:

layer = []

layer.append([layer_type])

if idx == layer_count-1:

layer.append([output_dim])

layer.append([out_activation])

else:

layer.append([layer_size])

layer.append([activation])

layers.append(layer)

idx = idx + 1

new_c_conf[SessionConfigBuilderCustom1.keras_nn_layers_key] = [layers]

new_configs.append(new_c_conf)

else:

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds batch size info to configs

# returns a list with new configs

@staticmethod

def add_keras_batchsize_info(configs):

new_configs = list()

for conf in configs:

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_batch_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_batch_size_key)

new_configs.append(new_conf)

else:

conf_batchsizes = conf[SessionConfigBuilderCustom1.keras_nn_batch_size_key]

idx = 0

while idx < len(conf_batchsizes):

new_conf = copy.deepcopy(conf)

batchsizes = copy.deepcopy(conf_batchsizes)

batchsize = batchsizes.pop(idx)

new_conf[SessionConfigBuilderCustom1.keras_nn_batch_size_key] = [batchsize]

new_configs.append(new_conf)

idx = idx + 1

return new_configs

# expects a config json list

# adds epochs info to configs

# returns a list with new configs

@staticmethod

def add_keras_epochs_info(configs):

new_configs = list()

for conf in configs:

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_epochs_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_epochs_key)

new_configs.append(new_conf)

else:

conf_epochs_list = conf[SessionConfigBuilderCustom1.keras_nn_epochs_key]

idx = 0

while idx < len(conf_epochs_list):

new_conf = copy.deepcopy(conf)

epochs_list = copy.deepcopy(conf_epochs_list)

epochs = epochs_list.pop(idx)

new_conf[SessionConfigBuilderCustom1.keras_nn_epochs_key] = [epochs]

new_configs.append(new_conf)

idx = idx + 1

return new_configs

# expects a config json list

# adds nn metrics info to configs

# returns a list with new configs

@staticmethod

def add_keras_metrics_info(configs):

new_configs = list()

for conf in configs:

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_metrics_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_metrics_key)

new_configs.append(new_conf)

else:

conf_metrics_list = conf[SessionConfigBuilderCustom1.keras_nn_metrics_key]

idx = 0

while idx < len(conf_metrics_list):

new_conf = copy.deepcopy(conf)

metrics_list = copy.deepcopy(conf_metrics_list)

metrics = metrics_list.pop(idx)

new_conf[SessionConfigBuilderCustom1.keras_nn_metrics_key] = [metrics]

new_configs.append(new_conf)

idx = idx + 1

return new_configs

# expects a config json list

# adds optimizer info to configs

# returns a list with new configs

@staticmethod

def add_keras_optimizer_info(configs):

new_configs = list()

for conf in configs:

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_optimizer_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_optimizer_key)

new_configs.append(new_conf)

else:

conf_optimizers = conf[SessionConfigBuilderCustom1.keras_nn_optimizer_key]

idx = 0

while idx < len(conf_optimizers):

new_conf = copy.deepcopy(conf)

optimizers = copy.deepcopy(conf_optimizers)

optimizer = optimizers.pop(idx)

new_conf[SessionConfigBuilderCustom1.keras_nn_optimizer_key] = [optimizer]

new_configs.append(new_conf)

idx = idx + 1

return new_configs

# expects a config json list

# adds loss function info to configs

# returns a list with new configs

@staticmethod

def add_keras_loss_function_info(configs):

new_configs = list()

for conf in configs:

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

if classifier_type != SessionConfigBuilderCustom1.classifier_keras_nn:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_loss_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_loss_key)

new_configs.append(new_conf)

else:

conf_loss_functions = conf[SessionConfigBuilderCustom1.keras_nn_loss_key]

idx = 0

while idx < len(conf_loss_functions):

new_conf = copy.deepcopy(conf)

loss_functions = copy.deepcopy(conf_loss_functions)

loss_function = loss_functions.pop(idx)

new_conf[SessionConfigBuilderCustom1.keras_nn_loss_key] = [loss_function]

new_configs.append(new_conf)

idx = idx + 1

return new_configs

# expects a config json list

# adds similarity function info to configs

# returns a list with new configs

@staticmethod

def add_similarity_function_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.classification_interpreter_key in conf:

classification_interpreter_type = conf[SessionConfigBuilderCustom1.classification_interpreter_key][0]

if classification_interpreter_type != SessionConfigBuilderCustom1.classification_interpreter_custom1:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.similarity_function_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.similarity_function_key)

new_configs.append(new_conf)

else:

conf_similarity_functions = conf[SessionConfigBuilderCustom1.similarity_function_key]

idx = 0

while idx < len(conf_similarity_functions):

new_conf = copy.deepcopy(conf)

similarity_functions = copy.deepcopy(conf_similarity_functions)

similarity_function = similarity_functions.pop(idx)

new_conf[SessionConfigBuilderCustom1.similarity_function_key] = [similarity_function]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.similarity_function_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.similarity_function_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds evaluation threshold for classification output info to configs

# returns a list with new configs

@staticmethod

def add_evaluation_threshold_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.classification_interpreter_key in conf:

classification_interpreter_type = conf[SessionConfigBuilderCustom1.classification_interpreter_key][0]

if classification_interpreter_type != SessionConfigBuilderCustom1.classification_interpreter_custom2:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key in conf:

new_conf.pop(SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key)

new_configs.append(new_conf)

else:

if SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key in conf:

conf_interpreter_thresholds = conf[SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key]

idx = 0

while idx < len(conf_interpreter_thresholds):

new_conf = copy.deepcopy(conf)

interpreter_thresholds = copy.deepcopy(conf_interpreter_thresholds)

interpreter_threshold = interpreter_thresholds.pop(idx)

new_conf[SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key] = [interpreter_threshold]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

new_configs.append(new_conf)

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key in conf:

new_conf.pop(SessionConfigBuilderCustom1.classification_interpreter_output_threshold_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds classification output layer info to configs

# returns a list with new configs

@staticmethod

def add_classification_output_layer_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.classification_interpreter_key in conf:

conf_classification_interpreters = conf[SessionConfigBuilderCustom1.classification_interpreter_key]

classifier_type = conf[SessionConfigBuilderCustom1.classifier_key][0]

idx = 0

while idx < len(conf_classification_interpreters):

new_conf = copy.deepcopy(conf)

classification_interpreters = copy.deepcopy(conf_classification_interpreters)

classification_interpreter = classification_interpreters.pop(idx)

new_conf[SessionConfigBuilderCustom1.classification_interpreter_key] = [classification_interpreter]

if classifier_type == SessionConfigBuilderCustom1.classifier_keras_nn:

if SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key in new_conf:

keras_nn_default_output_layer_sizes = copy.deepcopy(conf[SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key])

if len(keras_nn_default_output_layer_sizes) > idx:

keras_nn_default_output_layer_size = keras_nn_default_output_layer_sizes.pop(idx)

else:

keras_nn_default_output_layer_size = keras_nn_default_output_layer_sizes.pop(0)

new_conf[SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key] = [keras_nn_default_output_layer_size]

else:

if SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key)

if classifier_type == SessionConfigBuilderCustom1.classifier_keras_nn:

keras_nn_model_ids = copy.deepcopy(conf[SessionConfigBuilderCustom1.keras_nn_model_id_key])

if len(keras_nn_model_ids) > idx:

keras_nn_model_id = keras_nn_model_ids.pop(idx)

else:

keras_nn_model_id = keras_nn_model_ids.pop(0)

new_conf[SessionConfigBuilderCustom1.keras_nn_model_id_key] = [keras_nn_model_id]

else:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_model_id_key)

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_default_output_layer_size_key)

if SessionConfigBuilderCustom1.keras_nn_model_id_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.keras_nn_model_id_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds word vector to document vector info to configs

# returns a list with new configs

@staticmethod

def add_wv2dv_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.word_vec_to_doc_vec_key in conf:

conf_wv2dvs = conf[SessionConfigBuilderCustom1.word_vec_to_doc_vec_key]

idx = 0

while idx < len(conf_wv2dvs):

new_conf = copy.deepcopy(conf)

wv2dvs = copy.deepcopy(conf_wv2dvs)

wv2dv = wv2dvs.pop(idx)

new_conf[SessionConfigBuilderCustom1.word_vec_to_doc_vec_key] = [wv2dv]

feature_vec_dims = copy.deepcopy(conf[SessionConfigBuilderCustom1.feature_vec_dim_key])

if len(feature_vec_dims) > idx:

feature_vec_dim = feature_vec_dims.pop(idx)

else:

feature_vec_dim = feature_vec_dims.pop(0)

new_conf[SessionConfigBuilderCustom1.feature_vec_dim_key] = [feature_vec_dim]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds word2vec model workers info to configs

# returns a list with new configs

@staticmethod

def add_w2v_workers_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.vectorizer_key in conf:

vectorizer_type = conf[SessionConfigBuilderCustom1.vectorizer_key][0]

if vectorizer_type != SessionConfigBuilderCustom1.vectorizer_gensim_word2vec:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_workers_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_workers_key)

new_configs.append(new_conf)

else:

conf_w2v_model_worker_counts = conf[SessionConfigBuilderCustom1.word2vec_workers_key]

idx = 0

while idx < len(conf_w2v_model_worker_counts):

new_conf = copy.deepcopy(conf)

w2v_model_worker_counts = copy.deepcopy(conf_w2v_model_worker_counts)

w2v_model_worker_count = w2v_model_worker_counts.pop(idx)

new_conf[SessionConfigBuilderCustom1.word2vec_workers_key] = [w2v_model_worker_count]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_workers_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_workers_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds word2vec model min count info to configs

# returns a list with new configs

@staticmethod

def add_w2v_min_count_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.vectorizer_key in conf:

vectorizer_type = conf[SessionConfigBuilderCustom1.vectorizer_key][0]

if vectorizer_type != SessionConfigBuilderCustom1.vectorizer_gensim_word2vec:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_min_count_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_min_count_key)

new_configs.append(new_conf)

else:

conf_w2v_model_min_counts = conf[SessionConfigBuilderCustom1.word2vec_min_count_key]

idx = 0

while idx < len(conf_w2v_model_min_counts):

new_conf = copy.deepcopy(conf)

w2v_model_mincounts = copy.deepcopy(conf_w2v_model_min_counts)

w2v_model_mincount = w2v_model_mincounts.pop(idx)

new_conf[SessionConfigBuilderCustom1.word2vec_min_count_key] = [w2v_model_mincount]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_min_count_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_min_count_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds word2vec model window size info to configs

# returns a list with new configs

@staticmethod

def add_w2v_window_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.vectorizer_key in conf:

vectorizer_type = conf[SessionConfigBuilderCustom1.vectorizer_key][0]

if vectorizer_type != SessionConfigBuilderCustom1.vectorizer_gensim_word2vec:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_window_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_window_key)

new_configs.append(new_conf)

else:

conf_w2v_model_window_sizes = conf[SessionConfigBuilderCustom1.word2vec_window_key]

idx = 0

while idx < len(conf_w2v_model_window_sizes):

new_conf = copy.deepcopy(conf)

w2v_model_window_sizes = copy.deepcopy(conf_w2v_model_window_sizes)

w2v_model_window_size = w2v_model_window_sizes.pop(idx)

new_conf[SessionConfigBuilderCustom1.word2vec_window_key] = [w2v_model_window_size]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_window_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_window_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds word2vec model size info to configs

# returns a list with new configs

@staticmethod

def add_w2v_size_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.vectorizer_key in conf:

vectorizer_type = conf[SessionConfigBuilderCustom1.vectorizer_key][0]

if vectorizer_type != SessionConfigBuilderCustom1.vectorizer_gensim_word2vec:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_size_key)

new_configs.append(new_conf)

else:

conf_w2v_model_sizes = conf[SessionConfigBuilderCustom1.word2vec_size_key]

idx = 0

while idx < len(conf_w2v_model_sizes):

new_conf = copy.deepcopy(conf)

w2v_model_sizes = copy.deepcopy(conf_w2v_model_sizes)

w2v_model_size = w2v_model_sizes.pop(idx)

new_conf[SessionConfigBuilderCustom1.word2vec_size_key] = [w2v_model_size]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.word2vec_size_key in new_conf:

new_conf.pop(SessionConfigBuilderCustom1.word2vec_size_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds vectorizer info to configs

# returns a list with new configs

@staticmethod

def add_vectorizer_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.vectorizer_key in conf:

conf_vectorizers = conf[SessionConfigBuilderCustom1.vectorizer_key]

idx = 0

while idx < len(conf_vectorizers):

new_conf = copy.deepcopy(conf)

vectorizers = copy.deepcopy(conf_vectorizers)

vectorizer = vectorizers.pop(idx)

new_conf[SessionConfigBuilderCustom1.vectorizer_key] = [vectorizer]

if SessionConfigBuilderCustom1.vec_model_id_key in conf:

vectorizer_model_ids = copy.deepcopy(conf[SessionConfigBuilderCustom1.vec_model_id_key])

if len(vectorizer_model_ids) > idx:

vectorizer_model_id = vectorizer_model_ids.pop(idx)

else:

vectorizer_model_id = vectorizer_model_ids.pop(0)

new_conf[SessionConfigBuilderCustom1.vec_model_id_key] = [vectorizer_model_id]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.vec_model_id_key in new_conf:

new_conf[SessionConfigBuilderCustom1.vec_model_id_key] = [new_conf[SessionConfigBuilderCustom1.vec_model_id_key][0]]

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds stopword removal info to configs

# returns a list with new configs

@staticmethod

def add_stopword_removal_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.stopword_remover_key in conf:

conf_stopword_removers = conf[SessionConfigBuilderCustom1.stopword_remover_key]

idx = 0

while idx < len(conf_stopword_removers):

new_conf = copy.deepcopy(conf)

stopword_removers = copy.deepcopy(conf_stopword_removers)

stopword_remover = stopword_removers.pop(idx)

new_conf[SessionConfigBuilderCustom1.stopword_remover_key] = [stopword_remover]

if SessionConfigBuilderCustom1.stopwords_identifier_key in conf:

stopwords_ids = copy.deepcopy(conf[SessionConfigBuilderCustom1.stopwords_identifier_key])

if len(stopwords_ids) > idx:

stopwords_id = stopwords_ids.pop(idx)

else:

stopwords_id = stopwords_ids.pop(0)

new_conf[SessionConfigBuilderCustom1.stopwords_identifier_key] = [stopwords_id]

if SessionConfigBuilderCustom1.additional_stopwords_key in conf:

additional_stopwords_list = copy.deepcopy(conf[SessionConfigBuilderCustom1.additional_stopwords_key])

if len(additional_stopwords_list) > idx:

additional_stopwords = additional_stopwords_list.pop(idx)

else:

additional_stopwords = additional_stopwords_list.pop(0)

new_conf[SessionConfigBuilderCustom1.additional_stopwords_key] = [additional_stopwords]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

if SessionConfigBuilderCustom1.stopwords_identifier_key in conf:

new_conf.pop(SessionConfigBuilderCustom1.stopwords_identifier_key)

if SessionConfigBuilderCustom1.additional_stopwords_key in conf:

new_conf.pop(SessionConfigBuilderCustom1.additional_stopwords_key)

new_configs.append(new_conf)

return new_configs

# expects a config json list

# adds corpus info to configs

# returns a list with new configs

@staticmethod

def add_corpus_info(configs):

new_configs = list()

for conf in configs:

if SessionConfigBuilderCustom1.corpus_importer_key in conf:

conf_corpus_importers = conf[SessionConfigBuilderCustom1.corpus_importer_key]

idx = 0

while idx < len(conf_corpus_importers):

new_conf = copy.deepcopy(conf)

corpus_importers = copy.deepcopy(conf_corpus_importers)

importer = corpus_importers.pop(idx)

new_conf[SessionConfigBuilderCustom1.corpus_importer_key] = [importer]

corpus_ids = copy.deepcopy(conf[SessionConfigBuilderCustom1.corpus_identifier_key])

if len(corpus_ids) > idx:

corpus_id = corpus_ids.pop(idx)

else:

corpus_id = corpus_ids.pop(0)

new_conf[SessionConfigBuilderCustom1.corpus_identifier_key] = [corpus_id]

categories_ids = copy.deepcopy(conf[SessionConfigBuilderCustom1.categories_identifier_key])

if len(categories_ids) > idx:

categories_id = categories_ids.pop(idx)

else:

categories_id = categories_ids.pop(0)

new_conf[SessionConfigBuilderCustom1.categories_identifier_key] = [categories_id]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

new_configs.append(new_conf)

return new_configs

# expects a config json list and a parameter key

# adds simple config info to configs

# returns a list with new configs

@staticmethod

def add_simple_config_info(configs, key):

new_configs = list()

for conf in configs:

if key in conf:

conf_values = conf[key]

idx = 0

while idx < len(conf_values):

new_conf = copy.deepcopy(conf)

values = copy.deepcopy(conf_values)

value = values.pop(idx)

new_conf[key] = [value]

new_configs.append(new_conf)

idx = idx + 1

else:

new_conf = copy.deepcopy(conf)

new_configs.append(new_conf)

return new_configs

# expects a list of configs

# adds the complex config information to all configs

# returns a list of configs

@staticmethod

def add_all_config_info(configs):

configs = SessionConfigBuilderCustom1.add_corpus_info(configs)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.preprocessor_key)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.noise_remover_key)

configs = SessionConfigBuilderCustom1.add_stopword_removal_info(configs)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.lemmatizer_key)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.train_test_splitter_key)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.train_test_split_ratio_key)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.train_test_split_random_state_key)

configs = SessionConfigBuilderCustom1.add_vectorizer_info(configs)

configs = SessionConfigBuilderCustom1.add_w2v_size_info(configs)

configs = SessionConfigBuilderCustom1.add_w2v_window_info(configs)

configs = SessionConfigBuilderCustom1.add_w2v_min_count_info(configs)

configs = SessionConfigBuilderCustom1.add_w2v_workers_info(configs)

configs = SessionConfigBuilderCustom1.add_wv2dv_info(configs)

configs = SessionConfigBuilderCustom1.add_simple_config_info(configs, SessionConfigBuilderCustom1.lemmatizer_key)

configs = SessionConfigBuilderCustom1.add_classification_output_layer_info(configs)

configs = SessionConfigBuilderCustom1.add_evaluation_threshold_info(configs)

configs = SessionConfigBuilderCustom1.add_similarity_function_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_loss_function_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_optimizer_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_metrics_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_epochs_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_batchsize_info(configs)

configs = SessionConfigBuilderCustom1.add_keras_layers_info(configs)

return configs

# constructs a number of session configs, stores them

# returns list with config identifiers

@staticmethod

def create_session_configs(configs_location=None, delete_old_configs=1):

if configs_location is None:

configs_location = ConfigReader.get_configs_location()

if delete_old_configs:

Storage.delete_location(configs_location)

configs = [SessionConfigReader.get_config_template()]

configs = SessionConfigBuilderCustom1.add_all_config_info(configs)

n_configs = len(configs)

SessionLogger.log('Constructed ' + str(n_configs) + ' new session configs from template: \'' + ConfigReader.get_config_template_id() + '\'.')

config_ids = list()

idx = 0

for conf in configs:

config_id = configs_location + '/' + SessionConfigBuilderCustom1.config_name + str(idx+1)

SessionConfigReader.set_config(conf, config_id)

config_ids.append(config_id)

idx = idx + 1

SessionLogger.log('Stored ' + str(n_configs) + ' session configs in \'' + configs_location + '\'.')

return config_ids

# returns the configs location

@staticmethod

def get_configs_location():

return ConfigReader.get_configs_location()

# returns the general config name

@staticmethod

def get_configs_name():