TEST: smoke-test own-data workflow examples

[biru-codeastromer]

Summary

This PR adds smoke coverage for the repository's runnable own-data examples and makes the SQL example degrade cleanly when its optional dependency is not installed.

Closes #398.

Why

The repo documents MCP workflows on user data (pandas, csv, sql), but those example scripts were not covered by tests. In a standard dev environment without sqlalchemy, examples/sql_example.py currently attempts the runnable SQLite workflow anyway and emits noisy executor error logs before reaching its template-only guidance.

Changes

• make examples/sql_example.py check for SQL support up front and print a clean install/skip message when sqlalchemy is unavailable
• add smoke tests for:
  • examples/03_pipeline_instantiation.py
  • examples/pandas_example.py
  • examples/csv_example.py
  • examples/sql_example.py

Verification

• .venv/bin/ruff check examples/sql_example.py tests/test_example_data_workflows.py
• .venv/bin/ruff format --check examples/sql_example.py tests/test_example_data_workflows.py
• .venv/bin/python examples/sql_example.py
• .venv/bin/pytest tests/test_example_data_workflows.py -q
• .venv/bin/pytest -q