Separation of two sets of vectors

NOT USED

Outputs up to the first 16 chars of token to fp. */

Output tree in Aroma SPT JSON format. See aroma-schema.json. */

DFS */

Compute and print following matricies: - matrix of numbers of similarity/dissimilarity tests for solutions of pairs of pro

just simplify some non-leaf node, if it has only one child, then it will become the simplified version of its child, recursively done so that only lea

Split array of shaffled samples straight forward way into test and, train and validation Returns: - test dataset

Siamese CNN with merging by computing absolute value of difference

Make test tf.dataset The tf.data.dataset is constructed from_generator Parameters: - batch_size -- batch size

Make test tf.dataset The tf.data.dataset is constructed from_generator Parameters: - batch_size -- batch size

Make test dataset Test datset is always constructed from problems different from ones used for training and validation

Checks whether tokens are equal (same text). */

Checks whether token start with s. */

Tokenization of JavaScript programming language source text. Returns 1 when a next valid token is recognized. Then the token is made avaliable i

Tokenization of Python programming language source text. Returns 1 when a next valid token is recognized. Then the token is made avaliable in th

Tokenize all problems Parameters: - problem_list -- problems to tokenize their solutions

Tokenize all problems Parameters: - problem_list -- problems to tokenize their solutions

Tokenize solution files of given problem and combine results in a single file, collect statistics about tokenized problems

Tokenize solution files of given problem and combine results in a single file, collect statistics about tokenized problems

Print report on tokens and their occurrence in processed files and write down info on tokens to tokenized dataset

Train given DNN on given dataset - dnn -- DNN compiled and optimized - val_ds -- Valdiation dataset as a pair

Train given DNN on given dataset - dnn -- DNN compiled and optimized - train_ds -- Training dataset as a pair

Make training and validation datasets by splitting loaded data set Parameters: - valpart -- Fraction of dataset s

Make symmetric two way convolutional dnn for sequence similarity analysis Parameters: - n_tokens -- number

Update statistics for dissimilarity sample

Update statistics for dissimilarity sample

Update statistics for similarity sample Parameters: s1, s2 -- solutions used in similarity sample

Compute worst confusions of classification Returns list of pairs , each of which is a classified problem and the problem pred

Write large matrix to file Parameters: - m -- matrix to write as numpy array of floats - f -- file to write

Write large matrix to file Parameters: - m -- matrix to write as numpy array of floats - f -- file to write

Write down similarity predictions Parameters: - test -- path to the testset file specifying pairs of s

Print report on processed problems sorted according to their number of solutions Parameters: - problems -- dictionar

Write matrix of accuracy of similarity/dissimilarity detection Parameters: - tested_problems -- list of tested problems

Write distribution of problem solutions