Dataset-generator: allow search_engine_type=vespa and default vespa_schema to collection_name

[radu-gheorghe]

This should make it a bit easier to use dataset-generator with Vespa.

search_engine_type=vespa didn't seem to work from the command line (uv run...) before this.

Vespa schema is the equivalent of a collection in Solr, so collection_name should be a good default. I don't know why one would override it, but it's still possible now.

Didn't want to be intrusive, it's my first contribution 🙈

Let me know if you need me to change anything.

[radu-gheorghe]

Added another follow-up fix for num_queries_needed.

While testing the Vespa query_template flow, I noticed that lowering num_queries_needed did not limit the evaluation work the generator performed if resources/tmp/datastore.json already contained more queries from a previous run.

This patch should make the behavior more predictable.