Fix/build manual

[dcollins15]

This PR applies some housekeeping changes to the package's DESCRIPTION file. Critically, the "BuildManual" field is added and set to true.

As far as I have been able to suss out, as of R v4.3.2, R CMD build requires the BuildManual field to be true in order for the package's PDF manual to be generated, even if there are install-time or render-time /Sexpr macros are present in the package's man/*.Rd files. Currently, the following NOTE is being thrown by CRAN checks (the last line is the problematic one):

❯ checking CRAN incoming feasibility ... [6s/19s] NOTE
  Maintainer: ‘Rahul Satija <seurat@nygenome.org>’
  
  Version contains large components (5.0.99.9003)
  
  Suggests or Enhances not in mainstream repositories:
    BPCells
  Availability using Additional_repositories specification:
    BPCells   yes   https://bnprks.r-universe.dev/
  
  Package has help file(s) containing install/render-stage \Sexpr{} expressions but no prebuilt PDF manual.

I first encountered this issue while cutting the release for SeuratObject v5.0.2 but attributed it to a problem with my local environment. Upon digging deeper, I noticed that the successful build used R v4.2.2 and was eventually able to pinpoint the issue to this change in the R source code, which updated the default value for the "BuildManual" field from TRUE to FALSE 🕵️

Curiously, CRAN's documentation on writing R extensions appears to describe the old behavior 🤷

The only other non-cosmetic change introduced in this PR is that the "Type" field was dropped from the package's DESCRIPTION—this field defaults to "Package" and can only take one other value ("Frontend"). The main motivation for dropping it is for consistency with Seurat 🤓

[dcollins15]

I had been fiddling around with the repository settings and accidentally disabled third-party actions—I've adjusted the relevant setting now. Closing and re-opening to trigger CI checks.