Expose Tantivy's TermSetQuery

[aecio]

This PR adds the Query.term_set_query method and corresponding tests.

Expected usage:

terms = ["old", "man"]
query = Query.term_set_query(index.schema, "title", terms)
result = index.searcher().search(query, 10)

The new term_set_query method has a signature similar to the existing term_query:

@staticmethod
def term_query(schema: Schema, field_name: str, field_value: Any, index_option: str = "position") -> Query:
    pass

@staticmethod
def term_set_query(schema: Schema, field_name: str, field_values: Sequence[Any]) -> Query:
    pass

See also issue #20.

[cjrh]

Since we know how long field_values is, perhaps we can use with_capacity? Then we won't need multiple allocations in the loop. Just an idea.

[aecio]

Yes, makes total sense. I wasn't aware of this API. I just updated and rebased the PR with the latest master.

[adamreichold]

Sorry for being late to the party. More often than not, it is preferable to use the built-in iterator adaptors with automatically do things like pre-allocate collections if possible, i.e. in this case collect should achieve the same effect but avoid adding the loop and allocation noise.

[adamreichold]

Specifically, something like

let terms = field_values
  .into_iter()
  .map(|field_value| {
    make_term(&schema.inner, field_name, &field_value).map_err(|err| {
      exceptions::PyTypeError::new_err(format!(
          "Cannot create a term from the value: {}",
          field_value
      ))
    })
  })
  .collect::<Result<Vec<_>, _>>()?;

should work and be just as efficient.

[aecio]

No worries, @adamreichold. I'm very new to Rust and I had tried a similar approach using iterator + map earlier but wasn't able to make the compiler happy. But with your snippet, I could make it work and was able to further simplify the code. It turns out that the original exception from make_term() is expressive enough so I just removed the map_errwhich was ignoring the argument |err| and causing a warning. Does it look better now?

[adamreichold]

Much better. Thanks for going the extra mile!