Rmd refactor (part 5)

[pdiakumis]

Better handling of package conflicts. First up: most of the changes are simple pkgA::fun1 -> fun1 changes, so this isn't as big as it looks ;-)

• Added the {conflicted} R dependency which forces you to choose which package to pull the function from when there are conflicts (see #87).
• Explicitly listing functions used from each package at the top of the Rmd (Python-style)
  • due to this we can remove most of the pkgA::fun1 explicit function calls and just use fun1.
  • I'm keeping the RNAsum-related explicit calls but can remove if we want
• Tidied up the data_transformation_plot chunk
  • simplified data.df construction via tibble + dplyr
  • using dplyr's slice to grab every 25th row
  • fixed plotly warning about markers is not in the mode
  • removing the detach("package:plotly") calls, I don't think they're needed but we can test and see later on.
  • I'm a bit curious why there seem to be duplicated plotting code chunks with minor diffs here. I think in the first part those are explicitly saved to variables, whereas next they get written to a PNG file?
  • renaming legend var to plot_legend1 to distinguish from legend function.

[pdiakumis]

@skanwal thanks for your patience with this, I resolved some conflicts with your branch so I think it's ready for review/merge now. I've stopped at the Manta-fusion prep stage since I need to have a better understanding of the SV requirements there, plus the Manta parsing has changed a bit since we rely on gpgr now. You've been following along with my commits so don't think I need to get into details. In summary I refactored:

• some of the plotting code
• the exprTable
• the copynumber section
• the known fusions section
• the fusion prep sections
• removed the allow_thumbnails + css stuff since I don't think they're getting used any more

[pdiakumis]

@skanwal ready for review/merge whenever you get some time. I've reached the end of the data prep stage (hooray!), leaving it in your hands now to see if you can also reach that stage without errors. It renders for me up to and including the goi_summary_table chunk. Next we'll need to tackle the prose in the rendered report and start testing/comparing with real data.