Peuqse

Update by Chris B on cov matrix scaling, so it remains symmetrical for greater than 2 x 2

The Geweke Diagnostic function has been updated to use spectral variance and normalized to the length of the windows. This accounts for the variances being sample variances and also for the fact that they are not independent samples. The Geweke Diagnostic now takes longer to converge, and is consistent with the pymc method for Geweke Diagnostic.

Several features have been upgraded to being more intuitive.
In particular, the logs_and_csvs now includes the data of the simulated responses in files like ExportedResponses
Additionally, when uniform distributions are used, and no initial guess is provided, the provided prior is now used as the initial guess rather than using the middle of the uniform distribution.

Improved warnings in likelihood functions

made getLogLikelihood_byResponses (which takes variances for uncertainties)
turned off convergence diagnostics during doe
turned off convergence diagnostics during multi-start (except for best run)
Fixed Example 12 and Example 13 to work with newest PEUQSE and dependencies

New combined plots for series
Extended doSinglePoint compatibility to all situations for reducedParameter space