Cache Image.format for better throughput

[TomeHirata]

Closes #8834

Cache Image.format for better inference time throughput.

[Copilot]

Pull Request Overview

This PR optimizes adapter methods by adding LRU cache decorators to improve inference time throughput. The changes cache the results of formatting methods that depend only on signature types rather than user inputs.

• Adds @lru_cache(maxsize=64) decorator to formatting methods across adapter classes
• Targets methods that format field structures, descriptions, and task information
• Focuses on methods that are signature-dependent but input-independent for safe caching

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
dspy/adapters/json_adapter.py	Adds caching to field structure and output requirements formatting methods
dspy/adapters/chat_adapter.py	Adds caching to field description, structure, and task description formatting methods
dspy/adapters/base.py	Adds caching to the demo formatting method in the base adapter class

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}