For integration tests, we were using an S3 bucket to read from and store
previously parsed documents. This PR replaces that with local output. This
means we don't have to have the cloud.gov-based S3 bucket. It does require
that PR submitters account fix any discrepancies in their PR, rather than a
merge fixing it.
H/T to @anthonygarvan for describing this option many moons ago.
For integration tests, we were using an S3 bucket to read from and store previously parsed documents. This PR replaces that with local output. This means we don't have to have the cloud.gov-based S3 bucket. It does require that PR submitters account fix any discrepancies in their PR, rather than a merge fixing it.
H/T to @anthonygarvan for describing this option many moons ago.