Boldigger

- integrated the API verification module from BOLDigger into the command-line version
- rewrote all identification engine scripts to match the speed of original BOLDigger
- added the new JAMP sorting method also to the command line version
- this update will break backward compatibility with old assignments, so to reanalyze a dataset the identification engine has to be rerun

- bugfix on the API verification option

- added a failsave in case the connection to BOLD is closed by BOLD before results come in. In this case the request is going to be resent until it gets a clean response

- fixed the hdf to excel converter so split the data in multiple parts to be able to save large datasets in multiple excel files. If the dataset exceeds 50.000 OTUs / sequences, the results will be saved in chunks of 50.000.

- found a bug where a NoMatch table would be saved when BOLD sends a malformed response and fixed it
- rewrote the complete code for fetching and saving the data
- Intermediate BOLDResults are now saved in .hdf format. This allows to append new data really quickly and is much faster than writing excel files in general. For large datasets, several hours of saving operations are therefore saved.
- When the download is complete the hdf will be translated into an excel file once.
- The hdf file will be retained and can be edited with e.g. Python or R

- increased the request speed by more than 100% by using asynchronous coding. Depending on the internet connection the download speed of the results can be up to 4x as fast. Initial tests shows less than a minute for 50 sequences compared to the usual 4 minutes per 50 sequences. speeds up to 1000 sequences per hour are now possible.
- cleaned up a lot of old code snippets