The JCAMP-DX parser was a bit flawed what comes to the JCAMP "block" structuring i.e. separating data sections (of different ##DATATYPEs). The blocks should always start at ##TITLE tag (this was not respected), end to ##END tag and they may also be nested (also did not work). This PR fixes the block reading as it is specified.

Some additional notes:

• The main parser function was split to two separate functions as suggested already in comments of Only consider relevant DATATYPEs (NMR SPECTRUM / FID) when reading JCAMP-DX #120
• Since the files may have multiple spectra, and data is built for only the first one by default, the _getdataarray function was renamed to getdataarray as it is needed to be called from outside the parser as well.
• There is a new test, please find the test data for it here:
  dicstructure2.zip