Repurpose

==============
- Img2Ts now logs only errors and warnings
- Handled an edge case where Img2Ts failed when a whole set of images was missing

============
- PynetCF time series type (OrthoMultiTs or IndexedRaggedTs) are now stored in
the global `timeSeries_format` attribute.
- `time_coverage_end` global attribute added to each time series file, this is
updated when appending data to an existing file.
- Some utility function were added that are often required in our conversion
packages

============
- Updates to repurpose (img2ts) for performant conversion (for non-orthogonal data)
- img2ts logging was improved
- The method for parallelization was updated to allow different backends
- A check was implemented to repeatedly try to append to a (temporarily) unavailable file
- Input grid for img2ts is now used from the input dataset if not specified by the user. This allows e.g. conversion of swath data.

============
- Use joblib for parallel processing framework, improved logging
- Added option to parallelize Img2Ts process
- Fix bug where a wrong grid attribute was used.

============
- Ts2Img module was rebuilt. Allows conversion of time series with NN lookup.
- Added example notebook for converting ASCAT time series into regularly gridded images.
- Added a simple parallelization framework, with logging and error handling.
- Added the option to pass custom pre- and post-processing functions to ts2img.

===========
- Update for new pyscaffold
- Fixed bug where resampling failed when a BasicGrid was passed instead of a CellGrid