Include Ebsco in reindex status counts

[paul-butcher]

We need to include the individual EBSCO records in the count reported by get_reindex_status, otherwise it will erroneously report that we have created more Works than we started with.

Originally posted by @paul-butcher in #2556 (comment)

[paul-butcher]

My initial plan WRT transforming EBSCO content was to split the collection into records in the transformer (assuming the documents are not huge). This means that the count would also have to somehow split the XML to find the expected record count.

It may be appropriate, instead, to crack the document in an adapter.