chore: bump fastmcp + google-cloud-bigquery, pin packaging<26 in build (

#12)

Three small changes in one PR:

- Bump fastmcp 2.13.0 → 3.3.1. Major version; our surface area
  (FastMCP, @mcp.tool, mcp.run, Annotated/Field) is unchanged. No
  code changes required.
- Bump google-cloud-bigquery 3.31.0 → 3.41.0. Minor.
- Pin packaging<26 in [build-system].requires. packaging==26.2 ships
  a broken wheel (missing requirements.py and other submodules), so
  hatchling crashes in PEP 517 isolated builds. Without this pin,
  'uv sync', 'uv build', and any 'uv run ...' subprocess fail —
  including all 9 integration tests that spawn the server.

Full suite (52/52) passes including real-BigQuery integration tests.
Lint/format/mypy clean.

docs: add BIGQUERY_MAX_BYTES_BILLED to example configs (#8)

Surface the per-query spend cap (introduced in #7) in the copy-paste
examples so users adopt a safe default. 20 GiB (~USD 0.10/query at
$5/TiB) — conservative starting point; the in-code default remains
100 GiB.

Co-authored-by: pvoorschot <pvo@domain.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Add vector search support to README (#6)

Remove max_results parameter from run_query tool (#4)

- Remove max_results parameter from run_query tool completely
- Update tool description to recommend using LIMIT clause in SQL queries
- Remove all CLI arguments and environment variables related to query max_results
- Update tests to remove max_results usage and use LIMIT in SQL instead
- Improve list tools documentation to clarify max_results is optional integer

The run_query tool now relies on SQL LIMIT clause for result control,
making it more explicit and giving users full control over result size.

Co-authored-by: pvoorschot <pvo@domain.com>

Improve arguments handling and documentation for PyPI (#3)

* improve arguments  handling and doc for pypi



---------

Co-authored-by: pvoorschot <pvo@domain.com>

Merge pull request #2 from pvoo/prepare-pypi-release

Prepare package for PyPI release

Release v0.1.0 - BigQuery MCP Server

Initial release with:
- Dataset/table exploration tools
- Safe SQL query execution
- Docker support
- Comprehensive authentication
- BigQuery table analyst agent