chore(query): improve python udf script

[sundy-li]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

• Check error when window expr inside window expr
• Python UDF bootstrap now runs shared prelude that adds stdlib/platlib paths, injects a _contextvars fallback, appends the restored dependency directory, and automatically adds any
  staged zip/whl/egg artifacts to sys.path. That lets scripts import packages shipped via IMPORTS=@stage/foo.zip without touching pipelines logic.
• Cached venv entries still store only archives, but hits now rehydrate to a temp directory and the stage artifacts stay unpacked; we never import directly from the archive, so
  NumPy/Pandas work as expected.
• After each Python runtime call, update_datablock inspects the returned error column (case‑insensitive). If any row carries a non‑null message (UTF‑8 string or bytes), we raise
  UDFRuntimeError with that text; non‑UTF8 payloads produce a fallback message. This ensures Python exceptions bubble up rather than silently producing partial results.

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is 

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".