fix: flatten assistant content for openai payload

[wangyingxuan383-ai]

修复 openai 兼容提供商在 assistant content 为数组时导致的上下文丢失问题（下游 OpenAI→Gemini 转换会忽略数组内容，导致复读）。

Modifications / 改动点

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

• 在 astrbot/core/provider/sources/openai_source.py 的 _finally_convert_payload 中，
  当 assistant 的 content 为 list 且仅包含 text（剔除 think 后）时，合并为 string；
  若存在非 text 内容，则保留 list（多模态兼容）。

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

复现步骤（修复前）：

1. openai_chat_completion 提供商指向 OpenAI 兼容网关（如 CLIProxyAPI）
2. 连续输入 A/B/C
3. B 回复重复 A，C 回复重复 A+B

修复后：assistant 历史保留，B/C 不再复读。

---

Checklist / 检查清单

• 😊 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。/ If there are new features added in the PR, I have discussed it with the authors through issues/emails, etc.
• 👀 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”。/ My changes have been well-tested, and "Verification Steps" and "Screenshots" have been provided above.
• 🤓 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了 requirements.txt 和 pyproject.toml 文件相应位置。/ 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.
• 😮 我的更改没有引入恶意代码。/ My changes do not introduce malicious code.

Summary by Sourcery

错误修复：

• 修复当助手内容为数组时导致助手上下文丢失的问题：通过将纯文本部分扁平化为单个字符串，同时保留非文本部分，以保持多模态兼容性。

Original summary in English

Summary by Sourcery

Bug Fixes:

• Fix loss of assistant context when assistant content is an array by flattening pure-text parts into a single string while retaining non-text parts for multimodal compatibility.

[sourcery-ai]

Hey - 我在这里给出了一些高层次的反馈：

• 当所有非 think 部分都是文本，但其中一些 text 字段缺失或为 None 时，它们会被静默转换为空字符串；建议显式跳过这些部分或记录日志，以避免内容在不经意间丢失。
• 如果 content 列表只包含 think 部分，message["content"] 会变成空字符串而不是空列表；请仔细检查下游使用方在这种边界情况下是否期望得到字符串，或者为了保持一致性，考虑保留列表类型。

给 AI 代理的提示词

Please address the comments from this code review:

## Overall Comments
- When all non-`think` parts are text but some `text` fields are missing or `None`, they are silently converted to empty strings; consider explicitly skipping such parts or logging to avoid losing content unexpectedly.
- If the content list contains only `think` parts, `message["content"]` becomes an empty string rather than an empty list; double-check that downstream consumers expect a string in this edge case, or preserve the list type for consistency.

---

Sourcery 对开源项目免费——如果你觉得我们的评审有帮助，欢迎分享 ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{帮我变得更有用！请在每条评论上点 👍 或 👎，我会根据你的反馈改进后续的评审。}

Original comment in English

Hey - I've left some high level feedback:

• When all non-think parts are text but some text fields are missing or None, they are silently converted to empty strings; consider explicitly skipping such parts or logging to avoid losing content unexpectedly.
• If the content list contains only think parts, message["content"] becomes an empty string rather than an empty list; double-check that downstream consumers expect a string in this edge case, or preserve the list type for consistency.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- When all non-`think` parts are text but some `text` fields are missing or `None`, they are silently converted to empty strings; consider explicitly skipping such parts or logging to avoid losing content unexpectedly.
- If the content list contains only `think` parts, `message["content"]` becomes an empty string rather than an empty list; double-check that downstream consumers expect a string in this edge case, or preserve the list type for consistency.

---

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.}