Temporary workaround for weird dplyr bug

[Jason Vander Heiden]

I temporarily disabled evaluation of findNovelAlleles in the vignette because of an inexplicable error on R-devel preventing tigger from passing the CRAN pre-check:

* checking re-building of vignette outputs ... [17s] WARNING
Error in re-building vignettes:
Quitting from lines 111-116 (Tigger-Vignette.Rmd) 
Error: processing vignette 'Tigger-Vignette.Rmd' failed with diagnostics:
task 10 failed - "Value of SET_STRING_ELT() must be a 'CHARSXP' not a 'NULL'"

I cannot reproduce the error using the rocker/r-devel image. I suspect it's an incompatibility between dplyr and the unstable R-devel build. For now, I did this:

{r, eval=FALSE, warning=FALSE}
# Detect novel alleles
novel_df <- findNovelAlleles(sample_db, germline_ighv, nproc=1)

We need to enable evaluation again ASAP, but I can't get it on CRAN as is and I can't reproduce to debug. So, I'm submitting with eval disabled and using the saved novel_df object.

[Jason Vander Heiden]

We should also make sure this isn't a namespace conflict by converting all @import directives to @importFrom directives.

[Jason Vander Heiden]

Converted everything to @importFrom. Seems to work fine on win-builder devel now.