Adds a method to the corpus class that uses ffmpeg in order to extract
frame images from a given movie file in the corpus.

Merge branch 'feature/create_corpus_cls' of ssh://gitlab.uni.lu:8022/cittena/cittena into feature/create_corpus_cls

Add a method that identifies missing components in the corpus such as
missing data or diagrams and generates a task dictionary that can be
used by other methods in order to create these components.

Tableaus can now not only be created for diagrams accross the corpus
taht refer to a peculiar contrast but also for all diagrams of a movie.

Adds a method to the Corpus class which allows to arrange all diagrams
for a specific contrast and representation of all movies  together in
one tableau of diagrams.

Add documentation for the basedir attribute.

Add a file crawler for *.png files in the `_report` method of the Corpus
class which adds the availalbe diagrams for specific contrasts.

Add a corpus class which holds information about available movie files,
frames, data and visualizations in a corpus of movie files, following
the cittena movie corpus scheme.
Add a method to report the status of the movie corpus.

Add unit test for the Frame class.
Add integration test for seqmean of luminance and saturation.

add a timestamp column before returning it

turn return value into np.timedelta64
calculate timevalue by pd.Timedelta instead of divmod

The threshold value should depent on the number of pixels in a frame
image and should not be absolute.

and change param name dtype to view for better consistency with class
names

time stamps were counted with fps=1, fps=4 is currently used for the extraction
of frames with ffmpeg

I misinterpreted the fps I use for extracting frames with ffmpeg

the {n…n} glob pattern does not work outside of zsh

not all methods and paramters were controlled

after my 1 year break in which I can not overview
the changes I try to set a new start