Fix support for small floats

[TendTo]

This solves the issue with RealVal where some smaller values (in absolute terms) would cause an issue when converted to string, since they use the scientific notation.
Alternatively, the mkReal should be updated to support the e notation.

Some examples that would be fixed by this change:

cvc5.pythonic.RealVal(1e-5) # => "0.0000100000"
"1/100000"
cvc5.pythonic.RealVal(-1e-5) # => "-0.0000100000"
"-1/100000"
cvc5.pythonic.RealVal(1e-50)
"8424983333484575/842498333348457493583344221469363458551160763204392890034487820288"
cvc5.pythonic.RealVal(-1e-50)
"-8424983333484575/842498333348457493583344221469363458551160763204392890034487820288"