Inconsistent handling of marked duplicate Reactions

[speth]

According to the documentation (reactions.rst):

If the ``'duplicate'`` option is specified, then the reaction not only *may*
have a duplicate, it *must*. Any reaction that specifies that it is a duplicate,
but cannot be paired with another reaction in the phase that qualifies as
its duplicate generates an error.

However, no such check is performed when importing reactions (see Kinetics::checkDuplicates)

As a point of reference, Chemkin does enforce this requirement when parsing reactions. Whether this is actually a useful check or not is a different question.