fix: minimize anthropic prompt cache change

[911218sky]

Related to #9196

Changes / 改动

• add explicit cache_control to the last system block in the Anthropic provider request payload

Verification / 验证

• python3 -m py_compile astrbot/core/provider/sources/anthropic_source.py
• tested with the running astrbot Docker container after syncing the updated file
• confirmed the container loaded the updated code

[sourcery-ai]

Hey - I've reviewed your changes and they look great!

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[gemini-code-assist]

Code Review

This pull request introduces prompt caching support for the Anthropic provider by adding cache control parameters and sorting tools to ensure consistent payloads. The review feedback suggests refactoring the duplicated payload preparation logic shared between _query and _query_stream into a single helper method to reduce code duplication and improve maintainability.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

To avoid code duplication and adhere to the repository's general rules, the identical payload preparation logic in _query and _query_stream should be refactored into a shared helper method (e.g., _prepare_query_payload). This helper will handle sorting tools, setting tool choices, configuring max tokens, enabling prompt caching, applying cache breakpoints, applying thinking config, and sanitizing assistant messages.

    @classmethod
    def _enable_prompt_caching(cls, payloads: dict) -> None:
        payloads["cache_control"] = dict(cls._PROMPT_CACHE_CONTROL)

    @staticmethod
    def _sort_anthropic_tools(tool_list: list[dict]) -> list[dict]:
        return sorted(tool_list, key=lambda tool: str(tool.get("name", "")))

    @classmethod
    def _apply_explicit_prompt_cache_breakpoints(cls, payloads: dict) -> None:
        tools = payloads.get("tools")
        if isinstance(tools, list) and tools:
            last_tool = tools[-1]
            if isinstance(last_tool, dict) and "cache_control" not in last_tool:
                last_tool["cache_control"] = dict(cls._PROMPT_CACHE_CONTROL)

        system_blocks = payloads.get("system")
        if not isinstance(system_blocks, list) or not system_blocks:
            return

        last_block = system_blocks[-1]
        if isinstance(last_block, dict) and "cache_control" not in last_block:
            last_block["cache_control"] = dict(cls._PROMPT_CACHE_CONTROL)

    def _prepare_query_payload(self, payloads: dict, tools: ToolSet | None) -> None:
        if tools:
            if tool_list := tools.get_func_desc_anthropic_style():
                payloads["tools"] = self._sort_anthropic_tools(tool_list)
                payloads["tool_choice"] = self._normalize_tool_choice(
                    payloads.get("tool_choice", "auto")
                )

        if "max_tokens" not in payloads:
            payloads["max_tokens"] = 65536
        self._enable_prompt_caching(payloads)
        self._apply_explicit_prompt_cache_breakpoints(payloads)
        self._apply_thinking_config(payloads)
        self._sanitize_assistant_messages(payloads)

References

1. When implementing similar functionality for different cases, refactor the logic into a shared helper function to avoid code duplication.

[gemini-code-assist]

Use the newly introduced _prepare_query_payload helper method here as well to simplify the payload preparation in _query_stream.

        self._prepare_query_payload(payloads, tools)

References

1. When implementing similar functionality for different cases, refactor the logic into a shared helper function to avoid code duplication.

[gemini-code-assist]

Remove this duplicated payload preparation block as it is now fully handled by the _prepare_query_payload helper method.

References

1. When implementing similar functionality for different cases, refactor the logic into a shared helper function to avoid code duplication.