feat(python): support batched frame iteration over read_database queries

[alexander-beedie]

Adding another component to the database connectivity feature set; this PR enables support for generator return of DataFrame batches from read_database, using a new iter_batches mode/option. Works generically across SQLAlchemy, ODBC, and more (provided that the underlying backend/driver actually has such support).

Note that you can set batch_size without also setting iter_batches (in which case the batches are assembled into a single DataFrame for you), but you cannot set iter_batches without also providing a batch_size.

Example

import polars as pl

for frame in pl.read_database(
    query = "SELECT * FROM some_table",
    connection = user_connection,
    iter_batches = True,  # << new mode
    batch_size = 1000,
):
    do_something(frame)

---

Also: closes #10697 as a drive-by (adds basic mitigation to prevent credential leakage from connectorx errors).

[ritchie46]

Cool stuff!