Skip to content

fix: normalize plugin-wrapped conversation content before memory extraction#140

Open
Exploreunive wants to merge 1 commit intoEverMind-AI:mainfrom
Exploreunive:fix/openclaw-plugin-wrapped-content
Open

fix: normalize plugin-wrapped conversation content before memory extraction#140
Exploreunive wants to merge 1 commit intoEverMind-AI:mainfrom
Exploreunive:fix/openclaw-plugin-wrapped-content

Conversation

@Exploreunive
Copy link

What

  • normalize nested plugin-wrapped conversation content into plain text before MemCell extraction
  • keep plain text messages unchanged
  • add regression tests for wrapped content normalization

Why

OpenClaw / Feishu plugin payloads may wrap the real user-visible text in nested content structures. EverMemOS currently passes these nested objects downstream, which pollutes memory extraction and makes the recalled memory harder for the LLM to understand.

Fixes

Fixes #139

Test

  • added test_data_process_normalizes_plugin_wrapped_content
  • added test_data_process_preserves_plain_text_content
  • ran:
  • PYTHONPATH=src pytest -q tests/test_conv_memcell_extractor.py -k "data_process_normalizes_plugin_wrapped_content or data_process_preserves_plain_text_content"
  • result: 2 passed, 5 deselected

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

openclaw新插件bug

1 participant