Wavecracker

---------------------
Improved error handling. If libraries are missing for non-core functions, this is handled more gracefully.

new argument --downsampling (useful for large datasets)
Note, this is part of LOADING, not of preprocessing.
Note, this works only with values <= 50%.

Improved docs.

---------------------
Cosmetic changes in logs and in plots.

Bug corrected (CSV with ONE LINE of data was mishandled)

---------------------
Config file structure has changed.

Now audio files can be processed: mp3, mp4, m4a, wav, wmv (mono)

FFMPEG installation required for this. Ref. setup document.

Internal changes now will make extensions to other file formats easier (data loaders abstraction - ref. config file).
Calculations completely isolated in util/ft_calculator.py

Minor bugs fixed

Improved docs.

---------------------
Config file structure has changed.

!!!! Program now seamlessly manages directory or files !!!!
--dataset: suppressed. Now, we use --input-path, which can be one file or one dir.
If directories, the argument --input-datafile-extensions allows extension filtering,
Eg with '--input-datafile-extensions csv txt' we're processing CSV and TXT

!!!! Aggiunto filtraggio di frequenza!!!!
With argument --frequency-filter you can select frequency bands, and observe variations in the original signal.

Plots added (with configurability in YAML file):
--include-wl-transform (wavelet transform)
--include-histogram (--plot-hist-val-window --plot-hist-freq-window to select interval in both axes)
--include-psd (power spectrum density, with window selection in both axes --plot-psd-freq-window, --plot-psd-pow-window)
--include-phase (fourier transform phase, with window selection in both axes --plot-ftphase-freq-window --plot-ftphase-val-window)
--include-envelope

Added configurability for output CSV (encoding, separators, index column).

Output files now have timestamp in the name.

More configurability in all plots (linewidth, etc.)

Auto-detection of the encoding for CSV files.

Minor bugs resolved.

Improved docs.

---------------------
ENCODING is detected, and then used.

---------------------
Config file structure has changed.

New features:

Advanced separators handling:
- suppressed arguments: --decimal-sep, --alt-decimal-sep (until 1.3.0, they would permit specifying different separators)
- now it is all about the configuration file. There are DEFAULTS and you can configure an arbitrary number of attempts,
each of them with specific separators etc.

Now PREPROCESSING step exists. Happens after loading, but BEFORE the FT is calculated. So far, these are the preprocessing options:
- DC Offset removal
Ref. new command line argument, available in example script.

As a part of the loading process, SORTING can be requested (ref new argument --sort, see example script).
(useful in case the x-axis dimension is not sorted). Note, sorting is NOT preprocessing. It's part of the LOADING.

Now you can focus on an interval, in the x-axis, for the original signal.
Ref. argument --plot-signal-time-window, see details in example script.

Now you can focus on an interval for the values of the FT amplitude, ref. new argument --plot-ft-magn-window, see details in example script.

Header search improved (more flexible, not strict sequence of the headers).

Tool now shows, in logs, the time window under exam. Eg.:
.... [INFO] Signal sample time window: [-1.00000000000000002082E-02, 9.97855999999999916661E-03]

Tool now shows, in logs, the avg/min/max of the signal. Eg.:
.... [INFO] Signal sample values AVG (MIN/MAX): 2.63136613329176372034E-01 (-2.21608001738999998986E-02/5.56733608246000044595E-01)

Log formats for decimal numbers are configurable.

Log formats for sampling frequency are configurable.

Log now less verbose.

Bug fixing
- bug when no rows are present in a cvs, before header
- bug when FT output generation fails.

Documental changes
Setup instructions improved.
Known issues section eniched.