[kcheng486]

Summary

Why: Release DB-760's automatic reconnect (#80) and DB-773's session key (#79). The MCP server in wherobots/studio-backend#2894 is pinned to an unreleased commit of this repo until this version is on PyPI. It's a minor bump: it adds public API.

What ships since 0.29.0:

• Reconnect after a lost connection (feat: re-attach to a running execution after a lost connection (DB-760) #80, with feat: reconnect and re-attach automatically after a lost connection (DB-760) #82 stacked into it; DB-760).
  • A connection from connect() or connect_direct() now reconnects to the same SQL session on its own and re-attaches its pending queries, so their results still arrive.
  • New keyword reconnect_timeout_seconds: default 60 s, None turns reconnecting off.
  • New errors, all subclasses of OperationalError:
    • ConnectionLostError: reconnecting failed or is off.
    • ExecutionCancelledError: the session cancelled the query while it was detached.
    • ExecutionNotFoundError: the session no longer has the query.
  • New property Cursor.execution_id.
  • Server side, wherobots/sql-session#219 (merged) gives the query back to a re-attaching connection and cancels one nobody re-attaches to. Against a session without #219, the driver polls the query for its outcome instead.
• Per-caller SQL sessions (feat(DB-773): take a SQL session key and report the API's session fallback #79, DB-773). connect(session_key=...), Connection.session_fallback, and Connection.session_id, which is now public.
• Renovate policy (chore: adopt the org Renovate policy #81). The package itself is unchanged.

Behavior change for callers: in 0.29.0 a dropped connection failed pending queries at once, with a plain OperationalError. Now they wait up to reconnect_timeout_seconds for the reconnect, then fail with ConnectionLostError if it fails. Code that catches OperationalError still catches it.

What: Bump the version to 0.30.0 in pyproject.toml and uv.lock. After merge, push tag v0.30.0 on the merge commit to publish to PyPI.

Related Issues

Relates to DB-760 (#80, #82; server side wherobots/sql-session#219) and DB-773 (#79). Consumer: wherobots/studio-backend#2894.

---

Requester Checklist

Complete these before marking Ready for Review

• I have self-reviewed my own code
• I have added/updated tests that prove my fix/feature works (N/A, version bump only)
• I have included visual proof (screenshot, video, or test output) if applicable
• All CI checks are passing
• PR size is S/M, OR I have justified the size and added a walkthrough
• I have updated documentation if needed (N/A)

Visual Proof

On the bump commit:

$ uv run pytest tests/ -q
314 passed
$ uv run pre-commit run --all-files
all hooks Passed
$ uv lock --check && uv version
wherobots-python-dbapi 0.30.0   # what pypi-publish.yaml compares with the tag
$ uv build
wherobots_python_dbapi-0.30.0-py3-none-any.whl, wherobots_python_dbapi-0.30.0.tar.gz

I installed the built wheel in a clean venv:

• connect() takes reconnect_timeout_seconds and session_key.
• Cursor.execution_id and Connection.session_fallback exist.
• Cursor.attach and Connection.reconnect do not: feat: reconnect and re-attach automatically after a lost connection (DB-760) #82 removed them before feat: re-attach to a running execution after a lost connection (DB-760) #80 merged, so they never ship.

Size Justification (if L/XL)

N/A

---

Reviewer Checklist

If these are not met, close the tab — this PR is not ready for review

• Requester checklist above is complete
• All CI checks are passing
• Tests adequately cover the changes

Model: Claude Opus 5.5 (claude-opus-5-5) · Harness: Claude Code

🤖 Generated with Claude Code

[salty-hambot]

Reviewed by Salty Hambot 🤖🧂 at 4c942c1

No issues found.
_{💰 $0.0254 · 11.0k in / 633 out tokens · ⏱️ 7.2s · us.openai.gpt-6-sol high, prompt coverage-v1
💬 To request a re-review, comment @salty-hambot review}