consider *not* re-installing PACTA packages over those installed in the transitionmonitor image

[cjyetman]

workflow.prepare.pacta.indices/Dockerfile

Lines 26 to 31 in dd1e27d

	# install PACTA R packages
	RUN Rscript -e "pak::pkg_install(pkg = 'RMI-PACTA/pacta.data.scraping')"
	RUN Rscript -e "pak::pkg_install(pkg = 'RMI-PACTA/pacta.portfolio.import')"
	RUN Rscript -e "pak::pkg_install(pkg = 'RMI-PACTA/pacta.portfolio.allocate')"
	RUN Rscript -e "pak::pkg_install(pkg = 'RMI-PACTA/pacta.portfolio.audit')"
	RUN Rscript -e "pak::pkg_install(pkg = 'RMI-PACTA/pacta.portfolio.utils')"

The Dockerfile uses a transitionmonitordockerregistry.azurecr.io/rmi_pacta base image, which should already have all the relevant PACTA R packages installed, because it should be a fully functional PACTA image. Currently, the Dockerfile re-installs all the PACTA R packages again (from GitHub, and therefore the most current versions), which can lead to differing versions of the PACTA R packages being used then those that are included in the base image, which can lead to errors if breaking changes occurred between those versions.

[AlexAxthelm]

Given that an important part of preparing indices is ensuring parity between the code running the analysis for a portfolio and the code running the analysis on the index benchmarks, I strongly support this.

[cjyetman]

for some context, I tried running this for 2021Q4 and it didn't work because of a function signature change that happened since 2021Q4 benchmarks were last run

[cjyetman]

superseded by #67