Optimize repo size

[aspeake]

.github is really large making a fresh clone > 10 GB. This is may be due to the large savings shapes .gz files and binary files in the git history.

We may need to address:

• Reducing or dropping the savings_shapes .gz files on the repo
  • Which will then require downloading these for each CI run
  • Which will then probably require a more targeted selection of ECMs for an integration test (separate issue)
• Figure out ways to clean up the .github directory and remove objects no longer needed

Note that this PR adds binary files which could make this problem worse: #440

[jmythms]

Updates:

Running git gc, not changing history, will result in ~0.5% reduction in size of .git directory.
Running bfg:

• without deleting anything: $ java -jar bfg.jar --strip-blobs-bigger-than 100M scout.git: 20% reduction in size of .git directory.
• after deleting saving_shapes: 90% reduction in size: https://github.com/jmythms2/scout

[jmythms]

Details of pack file contents: https://docs.google.com/spreadsheets/d/1uM5g4BOD4-es9xZlKMyVMoTjkecvvY1fbKIPYx6QPKw/edit?usp=sharing