depth-first building tree @param tree @param parent @param ruleNames @param vocabulary @return

Check source dataset directory with respect to loaded test and training datasets Parameters: - source -- directory with source dataset

Check that all problems from source dataset directory are in one of datasets Parameters: - source -- directory with source dataset

Check pairs of samples for similarity. All n(n-1)/2 pairs are considered in principle. Avoid a quadratic number of tests by flagging samples

Return true for a tree, false for anything else. */

Compute lists of accuracy of similarity and dissimilarity detection Returns: - list of tested problems i-th element

Compare training and testing datasets Parameters: - test -- dictionary defining test dataset - train -- dictionary defining traini

Siamese CNN with merging by concatenation

Perform confusion analysis and report its results Parameters: - dnn -- DNN to make classification - ds -- Dataset to cla

Compute cosine similarity of 2 multisets. */

Siamese CNN with merging by cosine

Count how many nodes (including leaves and non-leaves) in the tree @param root @return

Writes absolute and relative numbers of correct and incorrect classifications of problems to csv file Parameters:

Write down classification dataset in csv form Parameters: - fn -- name of file to write to

Check tokenized file Parameters: - source_fn -- file name of source code - tok_fn -- file name of tokenizer outp

Make a stack of dense layers Stack may have no layers It is expected that input shape is known Parameters: -

@param root @return number of nodes

Distance between two vectors

Siamese CNN with merging by dot product

Classificating CNN with both max and average global pooling operations Parameters: - convolutions -- specification

Dump classification dataset Parameters: - fn -- name of file to write to

Checks whether token is a false or true literal. */

Phase 2. Removes line and block comments and condenses white-space up to a newline. Signals presence of any leading white-space by setting ws to

@param srcFileName @param dstFileName @return True, when successfully processed an SPT @throws IOException

Returns length of matched token (ws inclusive) or 0 when in error.

Get list of misclassifications for given classes Parameters: - classes -- indices of classes to report Returns list

Abstract method for getting classification predictions Returns: classification predictions

Get number of accepted solutions of a required problem Parameters: - metadir -- path to directory with metadata - pid -- proble

Get list of correct classifications of given classes Parameters: - classes -- indices of classes to report Returns l

Load tokenized samples of all solutions of programming problems, transform each sample into bag or sequence of tokens, creat

Get list of errors in similarity detection of problem solutions Parameters: - problems -- indices of problems to re

Get name of the token set and number of tokens used for tokenization of problem solutions Parameters: - ds - directory wi

Returns length of matched token (ws inclusive) or 0 when in error.

Build Text vectorization layer Args: texts (list): List of string, i.e., input texts vocab_size (int): vocab size max_seq (int)

Output graph in GraphViz dot format. */

Compute Jaccard similarity of 2 multisets, both with ignoring an element's multiplicity and with considering it. The first case of course equals

* Creates a new parser based over a given buffer with an array of tokens * available. */

* Parse JSON string and fill tokens. */