issues related to topic correlations

[cschwem2er]

Hi Brandon! :)

stminsights got archived due to failing unit tests of a function built upon stm::topicCorr(). I looked into that and am not sure what is going on.

For instance, the following code (pretty much the one taken from the documentation of the function)

cormat <- topicCorr(gadarianFit, cutoff = 0.01, method  = "simple")
> plot(cormat)

results in 0 edges and I am fairly certain that it used to show more of them in earlier stm versions. cor(gadarianFit$theta) produces mostly negative values. I wonder whether this can in any way be related to changes implemented in version 3.7 (e.g. RcppArmadillo fix)? In any case, I'd appreciate it if you could briefly check whether topicCorr() still works as intended.