Documentation improvements

This release aims only at improving docstrings, and documentation.
There is no changes to the code.

List of changes

• Correction of many typos in the wiki
• Minor corrections in most of the docstrings in the API
• Minor improvements in the output prints of the example scripts

Various improvements

This release adds various improvements to many functions as well improved docstring and documentation.

List of changes

• More docstrings to every functions
• Unified default loss function for both new AE_cls instance an deepAE_load function
• New unified emap function for computing the anomaly score map (NOTE: previous emap_sum and emap_mean functions are kept but deprecated)
• Created a wiki page detailing the AE_cls model specification and configuration (link in README)
• Added an example script showing the basic usage of the main features of the API (requires some external data)

Bug fix

This release provides a few fixes for annoying bugs.

List of change

• Bug fix/improvement in the __init__ method of AE_cls class
• Bug fix in generate_dataset function when source image shuffling is set
• Bug fix in deepAE_load function when the model was not set to the right device

Documentation and functionality imporvements

This release provides mostly some quality improvements of the docstrings and README documentation.

List of changes

• Improved docstrings for most functions and methods (unified formatting, completeness, ...)

• Added possibility to choose device for the deepAE_load function

• Added main features presentation to README

Improved development features for new models

This release include new/improved development feature to make training new models easier.

List of changes:

• Improved function to generate dataset for training/testing new models (defect_detection.generate_dataset)
  - Supports taking images from different sources and combined them into one dataset
  - Improved data augmentation feature

• Improved function to load a batch of images from disk directly to the required device (defect_detection.load_batch)
  - Support option to automatically generate a noisy image batch (useful for training new models)

• Added function to initialize and train a new model on a given dataset (defect_detection.deepAE_train)
  - The structure of the model must be defined in a configuration file
  - The provided training loop supports automatic input noising, learning rate decay and training/validation split

Fix the CD script again

Fix again a problem with pypi-publish GH action.

Fixed the CD script (again)

New attempt at fixing the problem with pypi-publish GH action.

Fix CD script

Fixed problem with pypi-publish action.

Filtering bug fix

Fixed the crash occurring when the anomalous pixel selection is empty.
Also updated CD script.

opencv headless update

Removed dependency to opencv-python package as it may case conflict with apps using PyQT5.
Dependency has been replaced by to opencv-python-headless.
Also, a few bugfix.