[CORRUPTED] Synthetic Benchmark PR #27749 - refactor(models): Refine MessageAgentThought SQLAlchemy typing

[ofir-frd]

Benchmark PR langgenius#27749

Type: Corrupted (contains bugs)

Original PR Title: refactor(models): Refine MessageAgentThought SQLAlchemy typing
Original PR Description: ## Summary

• Fixes [Chore/Refactor] Tighten MessageAgentThought type annotations langgenius/dify#27748.
• Add explicit Mapped[...] annotations to MessageAgentThought text, numeric, and temporal columns for SQLAlchemy 2.0 typing.
• Ensure pricing fields rely on Decimal typing so static analysis can treat message cost calculations consistently.

Screenshots

Before	After
n/a	n/a

Checklist

• This change requires a documentation update, included: Dify Document
• I understand that this PR may be closed in case there was no previous discussion or issues. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.
• I ran dev/reformat(backend) and cd web && npx lint-staged(frontend) to appease the lint gods

Original PR URL: langgenius#27749

Compliance Violation

• Rule: Python Functions Must Include Type Hints and Relevant Special Methods
• Language: Python
• File: api/core/agent/base_agent_runner.py