feat(async): stream daily extract downloads through core

[WilliamK112]

Add ofsc.core.get_daily_extract_file_stream(date, filename, chunk_size=8192) to the async client, as requested in #161. It is an async context manager yielding decoded byte chunks from the shared HTTP client's streaming response, so callers can consume large extracts incrementally.

Keep the response open during iteration and close it on completion, early exit, consumer exceptions, and cancellation. Preserve existing typed API errors by reading error responses, and translate connection/download read failures to OFSCNetworkError; errors from the caller's own upload or processing code retain their identity. Reject nonpositive chunk sizes before making a request. Add a README usage example.

Fixes #161.

Validation:

• 19 tests use real httpx response/context lifecycles with MockTransport: lazy consumption of an 80 MiB logical response with one reusable chunk, logging enabled/disabled, custom chunk boundaries, empty responses, HTTP error details, connection/read errors, cancellation, and consumer errors.
• Full mocked suite: 466 passed, 8 skipped on Python 3.13.
• Full Ruff check, Ruff format check (76 files), and diff checks pass.
• Pyright reports only the same two existing get_all_activities date-default diagnostics reproduced in a pristine master@56ded86 worktree; it reports no diagnostics in the new method.

Validation uses synthetic streams and no tenant credentials; a live OFS extract download has not been run.