Conversation
sairin1202
commented
Nov 18, 2025
sairin1202
commented
- Add basic usage example
- Add advanced scenario demonstrations
- Include code snippets for common patterns
There was a problem hiding this comment.
Pull Request Overview
This PR adds comprehensive usage examples and documentation for the MemU agentic memory framework. It streamlines the documentation by removing extensive self-hosting guides and localized README files, while significantly enhancing the main README with practical code examples demonstrating both RAG-based and LLM-based retrieval methods.
Key changes include:
- Addition of a new query rewriter prompt template for conversation-aware retrieval
- Comprehensive basic usage example showing memorization and dual retrieval approaches
- Detailed explanation of RAG vs LLM-based retrieval methods with use cases
- Removal of self-hosting documentation and multiple localized README files
Reviewed Changes
Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| src/memu/prompts/retrieve/query_rewriter.py | New prompt template for rewriting queries using conversation history to resolve references and ambiguities |
| README.md | Major update with installation instructions, comprehensive usage examples, and detailed retrieval method comparison |
| README.self_host.md | Removed self-hosting documentation (454 lines) |
| README.ja.md | Removed Japanese localized README (161 lines) |
| README.es.md | Removed Spanish localized README (197 lines) |
| README.de.md | Removed German localized README (183 lines) |
| README.ar.md | Removed Arabic localized README (190 lines) |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| </div> | ||
|
|
||
| MemU is a future-oriented Agentic Memory System that addresses the core issue of **inaccurate retrieval in traditional RAG-based memory frameworks**. | ||
| MemU is an agentic memory framework for LLM and AI agent backends. It receive multi-modal inputs, extracts them into memory items, and then organizes and summarizes these items into structured memory files. |
There was a problem hiding this comment.
Corrected grammar: 'receive' should be 'receives' to match the singular subject 'It'.
| MemU is an agentic memory framework for LLM and AI agent backends. It receive multi-modal inputs, extracts them into memory items, and then organizes and summarizes these items into structured memory files. | |
| MemU is an agentic memory framework for LLM and AI agent backends. It receives multi-modal inputs, extracts them into memory items, and then organizes and summarizes these items into structured memory files. |
|
|
||
| # Memorize a conversation | ||
| memory = await service.memorize( | ||
| resource_url="tests/data/example_conversation.json", |
There was a problem hiding this comment.
The code example references tests/data/example_conversation.json which does not exist in the repository. This will cause the example to fail when users try to run it. Consider either creating this example file, using a more accessible path, or providing instructions on how to create the required JSON file.
| @@ -0,0 +1,32 @@ | |||
| PROMPT = """Your task is to rewrite a user query by resolving references and ambiguities using the conversation history. | |||
There was a problem hiding this comment.
The new query_rewriter.py file defines a PROMPT variable but is not imported or used anywhere in the codebase. The existing code uses query_rewriter_judger.py instead. This creates confusion about which prompt template is actually being used for query rewriting. Either integrate this new prompt into the retrieval flow or document its intended future use.
