fix(media): filter first before getting git blob hash

[dcshzj]

Problem

Sites with a lot of images cannot load their images on the CMS.

Solution

Breaking Changes

• Yes - this PR contains breaking changes
• No - this PR is backwards compatible with ALL of the following feature flags in this doc

Improvements:

• Perform a filter first before getting the Git blob hash, which was found to be quite expensive for a large number of files.

Bug Fixes:

• The addedTime attribute was accidentally reverted earlier, fixed that so that the added times are now correct on EFS.
• Skip checking the .git folder, so we should have less error messages in the logs.

Tests

Smoke test:

• Navigate to moe-ast-moehc with the repo on GGS (i.e. email-login)
• Attempt to load the images page
• The page should load fairly quickly

Deploy Notes

None

[seaerchin]

lgtm. was wondering why we still force-fit the pagination stuff into listDirectoryContents given that they have differing goals (and resulted in listDirectoryContent having a huge API but default arguments help).

if we foresee the pagination stuff to always occur with listing the contents then ok, otherwise more in favour of breaking this up

[dcshzj]

if we foresee the pagination stuff to always occur with listing the contents then ok, otherwise more in favour of breaking this up

I don't have a good answer to this, so I have split the function up. Reading a media directory is the only time that needs pagination right now I believe, and it is also the only time that requires the Git blob hash.