Add index generation system that uses offsets into the WACZ itself.

[anjackson]

This proposed new py-wacz command allows you to generate a CDXJ file where the filenames and offsets refer to the WACZ itself rather than the WARC files within.

The idea is that we should be able to store the WACZ files directly without unpacking them, and use this modified CDXJ data to ingest into OutbackCDX. This should mean we can play back WACZ files as easily as we can WARCs.

Currently DRAFT:

• In case I've missed something and this is a terrible idea?
• Because perhaps the operation should be given a different name than just 'index'? Not sure it's the best name.
• Because it needs a way to override the path prefix, and possibly even the actual WACZ name, depending on the use case.
• And it should probably have some tests!

[edsu]

This is a pretty ingenious idea!

[anjackson]

The CI errors don't seem to be related to my patch?

pydantic.errors.PydanticUserError: A non-annotated attribute was detected: `version = '0.1.0'`. All model fields require a type annotation; if `version` is not meant to be a field, you may be able to resolve this error by annotating it as a `ClassVar` or updating `model_config['ignored_types']`.

Not sure what to do about that.