[FIX] Augment tempfile.SpooledTemporaryFile() for expected behavior

[tony-sappe]

Description:
This PR creates a new class to give tempfile.SpooledTemporaryFile an uplift. The current implementation in the standard library isn't satisfying the expected API. Augmenting the class methods and importing io.IOBase should get the class close to where it needs to be. Until python/cpython#3249 is merge, released, and used by this project we will need to keep this augmented class around.

Technical details:
While this issue has existed since Python 3.0, it was only recently felt in this project with the upgrade from pandas v0.45.x to pandas v1.0.2+ (pandas v1.0.0 and v1.0.1 introduced another bug which they fixed in v1.0.2, relying on API missing in the SpooledTemporaryFile implementation). See https://bugs.python.org/issue26175 and python/cpython#3249

Locally loadcfda, load_rosetta, load_city_county_state_code, and downloads were tested to ensure the changes did not break other areas. FABS and FPDS ETL scripts also use the class for obtaining delete records and should be tested. Several other ETL scripts (load_tas, load_dabs_submission_window_schedule) contain a path for loading a file which isn't used by the nightly pipeline which connects directly to broker for refreshing data.

Requirements for PR merge:

1. Unit & integration tests updated
2. API documentation updated (N/A)
3. Necessary PR reviewers:
   • Backend
   • Operations
4. Matview impact assessment completed (N/A)
5. Frontend impact assessment completed (N/A)
6. Data validation completed
7. Appropriate Operations ticket(s) created (N/A)
8. Jira Ticket (N/A)

Area for explaining above N/A when needed:

PR is only a change to class used for retrieving temporary files

[sethstoudenmier]

Stepped through comments on PR related to this change in cpython and everything seems to be in line with what is planned for that. LGTM.