feat: don't simply toss extensions

[mikix]

• Print a report at the end of the ETL run of all extensions stripped by resource, URL, and count.
• Plus a report on resources skipped due to unrecognized modifier extensions.
• Leave stub extension in place, with just the url field, so that we can inspect stripped extensions in SQL if we want.

This commit allows a bunch of extensions in use today. Scanned some sample data from Cerner and Epic to find common extensions. We'll surely add to the list going forward, to avoid annoying folks with the unknown-extension warning.

Ignore the following extensions:

• http://hl7.org/fhir/StructureDefinition/iso21090-TEL-address
• http://hl7.org/fhir/StructureDefinition/rendered-value
• http://hl7.org/fhir/us/core/StructureDefinition/us-core-direct
• https://fhir-ehr.cerner.com/r4/StructureDefinition/clinical-instruction
• https://fhir-ehr.cerner.com/r4/StructureDefinition/estimated-financial-responsibility-amount
• http://open.epic.com/FHIR/StructureDefinition/extension/birth-location

Drop support for the following extension:

• http://hl7.org/fhir/us/core/StructureDefinition/us-core-sex-for-clinical-use (this was in US Core 6.0 ballot, but didn't make final cut - it seems to be replaced by just 'us-core-sex')

Fixes #346

Example output

────────────────────────────────────────────────────────────────────────────────
Unrecognized extensions dropped from resources:
 Condition                            
 └── http://example.com/blarg (1 time)
 Encounter                                                                      
 ├── http://cerner.com/long-url/extensions/estimated-financial-responsibility-am
 │   ount (1 time)                                                              
 └── http://example.com/blarg (2 times)                                         
────────────────────────────────────────────────────────────────────────────────
🚨 Resources skipped due to unrecognized modifier extensions: 🚨
 Encounter                          
 └── http://example.com/foo (1 time)
────────────────────────────────────────────────────────────────────────────────
⭐ All tasks completed successfully! ⭐

Checklist

• Consider if documentation (like in docs/) needs to be updated
• Consider if tests should be added

[github-actions]

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines	Covered	Coverage	Threshold	Status
3595	3534	98%	98%	🟢

New Files

No new covered files...

Modified Files

File	Coverage	Status
cumulus_etl/deid/scrubber.py	100%	🟢
cumulus_etl/etl/cli.py	100%	🟢
cumulus_etl/etl/tasks/nlp_task.py	100%	🟢
TOTAL	100%	🟢

updated for commit: 820acfc by action🐍

[mikix]

Sorry that there's some refactoring noise - I basically tweaked how the recursive scrubbing worked - basically just reduced the amount of full-object access each scrubber method had.

[mikix]

These test files are for the ms-tool tests, which now no longer strips the extensions - so these changes just add the input extensions to the expected output.