buggy sitemap.xml generation

#9182 and #9240 attempted to fix this, but it seems part of the problem is that the default ingestion date for volumes (when no `ingest-date` attribute in XML) is 1900-01-01. And this is inherited for papers within the volume if not otherwise specified.

https://github.com/acl-org/acl-anthology/blob/7d1e9e40ba8a49b2827af97f1c7f5f7b5f3c4a1a/python/acl_anthology/constants.py#L30-L31

`` entries are filled in with the ingestion date. Google considers dates in 1900 to be invalid. We should probably just omit `` if we don't have a recorded ingestion date.

@mbollmann is there a reason not to use None as the default, for example?

Open python/acl_anthology/constants.py at the linked lines to verify the default ingest date value. Then trace where sitemap.xml lastmod is generated from that value and adjust logic so items without a recorded ingestion date omit lastmod instead of inheriting 1900-01-01. Run the sitemap-generation path or related tests and inspect generated sitemap.xml so valid dates remain and no 1900 dates are emitted.

Written by the indexing model from the issue text.