There are a few major changes in output and input that will likely break existing dependencies. Here are the highlights
the main output (lfsr, lfdr, etc) is rearranged into a dataframe, called
so, for example, the lfsr is now
a$result$lfsr instead of
or, better, use the accessor function
get_lfsr(a) to extract the lfsr etc
I added accessor functions
get_psd etc to access the lfsr, lfdr, posterior mean and posterior standard deviation. Using these functions to access results will help ensure your code remains valid if I happen to change the internal structure of the results again (although not anticipated…)
fitted.g is renamed
flash_data to make the whole package convention more consistent. Also
compdens replaced with
comp_dens, again for consistency.
nonzeromode option is replaced with the option
mode to specify mode. Or use
mode="estimate" to estimate the mode.
more flexible control of output. For example, you can say you want only the logLR output by specifying
outputlevel = c("lfsr","logLR"), or only posterior mean by
outputlevel = c("PosteriorMean").