feat: support automatic per-cell execution history filtering and isolated callbacks

[shuoweil]

This is a draft pr for pantheon testing. Please do not review it.

This PR introduces automatic cell-level isolation and execution history filtering for BigQuery DataFrames in notebook environments (Colab/Jupyter):

• Per-Cell Filtering: bigframes.execution_history() now automatically isolates and filters query history to the active notebook cell (controlled via the new filter_by_cell parameter).
• Isolated Callbacks: Added a callback parameter to _read_gbq_colab to support per-query execution event hooks that automatically clean up after execution completes.
• Context Propagation: Preserves the active cell's execution_count inside ExecutionSpec and propagates it thread-safely to background executors and BigQuery event callbacks.

Verified at:
Before: screen/6BJNDbWRKrAQtLk
After: screen/3SaU4eq9ZZfriQt

Fixes #<513337964> 🦕

[gemini-code-assist]

Code Review

This pull request introduces a callback parameter to the _read_gbq_colab function, allowing users to receive query execution events. The changes include updates to the public API, the session implementation, and the addition of a unit test to verify the callback functionality. A review comment suggests refactoring the _read_gbq_colab method to eliminate code duplication in the query execution logic by using a local helper function.