fix: parse in sim-struct chat-data; unify image analysis prompt with context injection & improve LLM robustness

[CaralHsi]

Description

Summary: This PR improves the multimodal image memory parsing pipeline by aligning prompt design, context injection, and error-handling behavior across image_parser and simple_struct.

Fix: #938

Docs Issue/PR: (docs-issue-or-pr-link)

Reviewer: @(reviewer)

Checklist:

• I have performed a self-review of my own code | 我已自行检查了自己的代码
• I have commented my code in hard-to-understand areas | 我已在难以理解的地方对代码进行了注释
• I have added tests that prove my fix is effective or that my feature works | 我已添加测试以证明我的修复有效或功能正常
• I have created related documentation issue/PR in MemOS-Docs (if applicable) | 我已在 MemOS-Docs 中创建了相关的文档 issue/PR（如果适用）
• I have linked the issue to this PR (if applicable) | 我已将 issue 链接到此 PR（如果适用）
• I have mentioned the person who will review this PR | 我已提及将审查此 PR 的人