Plots tiles in a grid (performs a normalization!). Useful for plotting the weights of a neural network or data instances. :param tiles_i

Save a histogram of values titled with title in file filename.

Loads data from file (bz2 compressed) :param fn: file name of dumped data :return: loaded data

Saves data to file (bz2 compressed) :param data: data to save :param fn: file name of dumped data

Gets the correct bpm based on the wav_file name. If the wav_file is not contained in the JKU dataset, raises error. Parameters ----------

Loads one or more configuration files and validates it against a specification. Parameters ---------- files : string or

Returns the offset (pick up measure), if any, from the score.

Applies a median filter of size L to the matrix of row observations X.

Formats the patterns into the csv format. Parameters ---------- pattersn : list List of patterns with its occurrences. midi_s

Percentage of variance explained by each component Returns ------- expvar: numpy array Explained varian

Project the data in the M dimensional space Returns ------- transform: numpy array Transformed data

Plot a given dataset and utilizes labels. Plots a given 2d dataset and uses different colors for datapoints belonging to different labels

Reads the wav file and downsamples to 11025 Hz.

Performs contrast normalization (zero mean, unit variance) along the given axis. :param x: array to normalize :param axis: normalize

If `cache_fn` exists, return the unpickled contents of that file (the cache file is treated as a bzipped pickle file). If this fails, com

Computes the CQT filters.

Computes the score for a given segment. The score is computed by taking a look at the trace of the block of the ssm.

Computes a spectrogram.

Computes the self similarity matrix from an audio file. Parameters ---------- wav_file: str Path to the wav file to be read.

Makes sure that the directory dir exists. Parameters ---------- dir: str Path to the directory to be created if needed.

Obtains the set of files containing the patterns and its occcurrences given the annotator directories. Parameters ---------- annotato

Given a set of patterns corresponding to a single musical piece and the output directory, get the output file path. Parameters ----------

Gets the smaller duration of a csv formatted score.

Computes the total duration of a csv formatted score.

Main function.

Main function.

Main function to run the audio polyphonic version of patterns finding.

Main function to parse the ground truth.

This function could do a simple hstack, if it weren't for the separation lines that we want between viewpoints. Furthermore, we need to r

Given a midi pitch (e.g. 60 == C), returns its corresponding chroma class value. A == 0, A# == 1, ..., G# == 11

Given a set of segments, find its occurrences and thus obtain the possible patterns. Parameters ---------- segments : list Li

Given an occurrence, return the csv formatted one into a list (onset,midi). Parameters ---------- start : float Start ind

Parses the set of patterns and saves the results into the output file. Parameters ---------- patterns: list of list of strings (files)

Principal components Returns ------- components: numpy array M Principal components

Main process to find the patterns in a polyphonic audio file. Parameters ---------- wav_file : str Path to the wav file to be ana

Main process to parse the ground truth csv files. Parameters ---------- jku_dir: string Directory where the JKU Dataset is locate

Save a histogram of values titled with title in file filename.

Scales all values in the ndarray ndar to be between 0 and 1

Sets the norms of all convolutional kernels of the C-GAE to a specific value. :param val: norms of kernels are set to this v

Trains the C-GAE for one epoch :param epoch: the current training epoch

Transposes axis 0 (zero-based) of x by [shift] steps. Missing information is padded with zeros. :param x: the array to transp

Mean vector of the observed data Returns ------- x_bar: numpy_array The mean vector of the observed dat

Returns random ngrams from data, can shift and scale data in two dimensions :param data_x: data (2d) :param length_n

Constructor for 1D signal dataset :param filelist: list of audio file names (str) :param sr: desired sam