Improve performance of exporting variable-width strings to numpy StringDType

The only way to create a string-typed numpy array from Python is from a list of Python string objects. This is very slow because you first need to copy your data into Python string objects.

(There are some low-level `unsafe` numpy C APIs, but I don't want to be mucking around manually in the numpy C API.)

We currently export variable-width string data _first_ to a list of `PyString` and _then_ to the numpy array constructor.

Hopefully the Rust-numpy crate will add safe APIs for passing a vector of `String` or `&str` to numpy from Rust.

No file or test is named; start by tracing the current variable-width string export and reviewing the safe Rust-numpy APIs available for constructing StringDType arrays. Done means exporting the data without first copying it into a list of Python string objects, while preserving the existing result.

Written by the indexing model from the issue text.