Add multi component parsing to support XY-datasets

[jgoedeke]

Closes #30

This pull request adds the two components to the parsing of a channel. This way XY-datasets can be parsed and xdata and ydata is properly set.
It utilizes the unused imc_object.hpp to cast every metadata key into a proper object. This may create a slight overhead because all parameters are parsed, but improves the readability IMO. It also implements the 6byte unsigned long datatype. Finally the buffer-size is exposed to the python interface to enable calculation of the file timespan without parsing the buffer.

There is a naming inconsistency in xoffset because this variable name was used for the x0 parameter in normal datasets before but is applicable to the offset value of the x component as well. I renamed the x0 parameter variabel to xstart but kept the name xoffset in the python interface (get_info())