added PlainExtractor based on numpy and scipy functions

restructured the stats containers

Remove StarVarianceExtractor since is functionality is featured in the existing Extractors

allow overlapping extraction sequences

renaming to chunk(s) and chunk_size and _shift

added test for chunk_shift and boundary case

dl1 table is a Qtable from astropy and not a ndarray

Length of table should be greater OR EQUAL than the size of chunk.

The mask is needed in the lstchain flatfield class to calculate the relative gain, which we will do outside the StatisticsExtractor.

Check for outliers depends if one process flatfield (signal) or pedestal (noise) events.

Change time to time_mono in order to read the telescope trigger and not the subarray.

Fix tests to have the correct types.

The for loop needs only run till length of the table minus the size of a chunk. Before the last chunks were created multiple times at the end.

When the chunks are shifted (in the second pass) we want to drop overflow chunks, while for the first pass we want to keep the last overflow chunk and extract the stats for the last possible chunk.

move outlier_check to traitlet; and remove redundant test on the input verification

init extractor classes without fixture inside the test itself

before two chunk iterations were used for looping over two columns of the same table

remove pylint command since not used in CI system of ctapipe

polish doc string of __call__ function with explanation of overlapping and non-overlapping chunks as well as the overflow of the bounds of the dl1 table

moved extraction start and stop time outside stats container and only provide the extraction start, i.e. validity start. The returning of the caller is a dict with key the validity start and items the StatisticsContainer.

added check for exceptions

before we returned a list, but now a dict

also update description of the returning dict

added end of the chunk as well

added also event IDs for the range

remove second space