fix: merge anthropic parallel tool results

[daniel5u]

修复了 #7871 中提到的问题，将 Anthropic API 下的 payload 构建方式进行调整，以匹配 Anthropic SDK 对于并行调用工具的要求（将连续的 tool_result 合并在一条进行返回）

Modifications / 改动点

• This is NOT a breaking change. / 这不是一个破坏性变更。

Screenshots or Test Results / 运行截图或测试结果

---

Checklist / 检查清单

• 😊 If there are new features added in the PR, I have discussed it with the authors through issues/emails, etc.
  / 如果 PR 中有新加入的功能，已经通过 Issue / 邮件等方式和作者讨论过。

• 👀 My changes have been well-tested, and "Verification Steps" and "Screenshots" have been provided above.
  / 我的更改经过了良好的测试，并已在上方提供了“验证步骤”和“运行截图”。

• 🤓 I have ensured that no new dependencies are introduced, OR if new dependencies are introduced, they have been added to the appropriate locations in requirements.txt and pyproject.toml.
  / 我确保没有引入新依赖库，或者引入了新依赖库的同时将其添加到 requirements.txt 和 pyproject.toml 文件相应位置。

• 😮 My changes do not introduce malicious code.
  / 我的更改没有引入恶意代码。

Summary by Sourcery

Adjust Anthropic provider payload construction to correctly group tool call results for parallel tool usage.

Bug Fixes:

• Merge consecutive Anthropic tool results into a single user message to align with the SDK’s parallel tool-calling expectations.

Tests:

• Add unit tests covering merging of consecutive tool results, handling of single tool results, and ensuring non-consecutive results are not merged.

[sourcery-ai]

Hey - I've found 1 issue, and left some high level feedback:

• The can_append_to_previous_tool_results condition in _prepare_payload is fairly dense; consider extracting this check into a small helper (e.g., _is_tool_result_user_message(last_message)) to improve readability and make the intent clearer.
• In the tool-result merge logic, you rely on new_messages[-1] when new_messages is non-empty; adding an early return or guard when the last message's content is not a list (instead of computing last_content and all(...)) could simplify the control flow and make it easier to follow.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- The `can_append_to_previous_tool_results` condition in `_prepare_payload` is fairly dense; consider extracting this check into a small helper (e.g., `_is_tool_result_user_message(last_message)`) to improve readability and make the intent clearer.
- In the tool-result merge logic, you rely on `new_messages[-1]` when `new_messages` is non-empty; adding an early return or guard when the last message's content is not a list (instead of computing `last_content` and `all(...)`) could simplify the control flow and make it easier to follow.

## Individual Comments

### Comment 1
<location path="tests/test_anthropic_kimi_code_provider.py" line_range="98-102" />
<code_context>
         )
+
+
+def _make_anthropic_provider_for_payload_tests() -> anthropic_source.ProviderAnthropic:
+    return anthropic_source.ProviderAnthropic(
+        provider_config={"model": "claude-test"},
+        provider_settings={},
+        use_api_key=False,
+    )
+
</code_context>
<issue_to_address>
**suggestion (testing):** Consider one test where the sequence starts with a tool message to guard the empty-`new_messages` path

The new `last_message = new_messages[-1] if new_messages else None` path isn’t covered by a case where the first input message is `role: "tool"`. Please add a test where `messages` begins with a tool message (no prior user/assistant), to confirm `_prepare_payload` correctly handles initially empty `new_messages` and still produces a single user message with one `tool_result`.
</issue_to_address>

---

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.}

[sourcery-ai]

suggestion (testing): Consider one test where the sequence starts with a tool message to guard the empty-new_messages path

The new last_message = new_messages[-1] if new_messages else None path isn’t covered by a case where the first input message is role: "tool". Please add a test where messages begins with a tool message (no prior user/assistant), to confirm _prepare_payload correctly handles initially empty new_messages and still produces a single user message with one tool_result.

[gemini-code-assist]

Code Review

This pull request modifies the Anthropic provider's payload preparation to merge consecutive tool results into a single user message, ensuring compatibility with Anthropic's API requirements for tool usage. New unit tests have been added to validate this merging behavior and ensure non-consecutive results are handled correctly. I have no feedback to provide.