fix: preserve group image context order

[PeiPei233]

Fixes #9634.

Built-in group chat context formatting waits for image captioning before appending a message. Because the event bus processes incoming messages concurrently, a following text message could append first and trigger an LLM request without the preceding image context.

Modifications / 改动点

• Hold the existing per-session group context lock while formatting and appending each message, preventing later messages in the same group session from overtaking an image caption request.

• Keep an [Image] placeholder when caption generation fails instead of storing an empty sender/timestamp record.

• Add deterministic async regression coverage for image/text ordering and caption failure fallback.

• 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

Verification steps:

$ uv run pytest -q tests/unit/test_group_chat_context_wiring.py tests/unit/test_event_bus.py
.......................                                                  [100%]
23 passed, 1 warning in 3.62s

$ uv run ruff check .
All checks passed!

$ uv run ruff format --check .
499 files already formatted

The concurrency regression test starts an image event, pauses its caption request, and then starts a text event for the same group session. It verifies that the text event cannot complete before the image caption and that the text-triggered LLM request receives the image record.

---

Checklist / 检查清单

• 😊 No new feature is introduced; this is a bug fix tracked by [Bug] Group image caption can be overtaken by a following text message #9634.
• 👀 My changes have been well-tested, and "Verification Steps" and test results have been provided above.
• 🤓 No new dependencies are introduced.
• 😮 My changes do not introduce malicious code.

Summary by Sourcery

Ensure group chat image messages are formatted and recorded atomically so subsequent text messages cannot overtake them in concurrent processing.

Bug Fixes:

• Prevent text messages in a group session from triggering LLM requests before preceding image caption context is recorded.
• Preserve an explicit image placeholder when caption generation fails instead of storing an empty record.

Enhancements:

• Add deterministic handling of image caption failures to keep group context consistent.

Tests:

• Add async regression tests verifying image/text ordering under concurrent events and the fallback image placeholder when captioning fails.

[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.}

[Copilot]

Pull request overview

This PR fixes a concurrency ordering bug in the built-in GroupChatContext so that image captioning cannot be overtaken by later messages in the same group session, ensuring LLM requests see the correct preceding context.

Changes:

• Hold the per-session group context lock while formatting and appending records to prevent message reordering during image captioning.
• Preserve an explicit [Image] placeholder when caption generation fails, instead of storing an empty sender/timestamp-only record.
• Add async regression tests for image/text ordering and caption-failure placeholder behavior.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
astrbot/builtin_stars/astrbot/group_chat_context.py	Makes message formatting+append atomic under the session lock; ensures caption failures still record an [Image] placeholder.
tests/unit/test_group_chat_context_wiring.py	Adds deterministic async tests covering ordering under concurrent events and placeholder behavior when captioning fails.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.