Description

Fixes #1448 — /product/search (fine mode) returning empty results for memories added through the fast pipeline. Root cause: tree_text_memory._graph_recall filtered Neo4j candidates by n.key IN parsed_goal.keys and tag overlap ≥ 2, but fast-mode add stores the raw chat-formatted text as key and only ["mode:fast"] as tags, while fine-mode search asks the LLM for high-level semantic keys/tags, so the strict candidate set comes back empty and the structured branch silently dropped the memory.

The fix extracts a lenient _node_matches_parsed_goal post-filter that accepts substring key matches and a single tag overlap, and adds _fallback_candidates_by_substring which uses the existing graph_store.get_all_memory_items API (bounded to 50 candidates per call) to substring-match parsed_goal.keys against stored key + memory text whenever the strict candidate set is empty. The fallback is scoped to WorkingMemory / LongTermMemory / UserMemory / OuterMemory so PreferenceMemory and other specialized scopes keep their dedicated retrievers. Wrapped in try/except so unexpected backends silently degrade to the previous behavior. No API, schema, or migration impact.

Tests: 4 new regression tests in tests/memories/textual/test_tree_retriever.py cover the reporter's exact scenario, substring post-filter, single-tag overlap, and graceful empty return. All 62 tests under tests/memories/textual/ pass after the fix; ruff check and ruff format both clean. The 6 unrelated failures under tests/mem_reader/ and tests/api/ are pre-existing on the dev-20260624-v2.0.22 baseline (verified by stashing the fix and re-running).

Related Issue (Required): Fixes #1448

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Refactor (does not change functionality, e.g. code style improvements, linting)
• Documentation update

How Has This Been Tested?

Automated tests are pending.

• Unit Test
• Test Script Or Test Steps (please provide)
• Pipeline Automated API Test (please provide)

Checklist

• I have performed a self-review of my own code
• I have commented my code in hard-to-understand areas
• I have added tests that prove my fix is effective or that my feature works
• I have created related documentation issue/PR in MemOS-Docs (if applicable)
• I have linked the issue to this PR (if applicable)
• I have mentioned the person who will review this PR

@MatthewZhuang, @CarltonXiang, @syzsunshine219, @World-controller please review this PR.

Reviewer Checklist

• closes fix: 添加包含偏好信息的记忆后，使用搜索接口POST /product/search无法检索到任何记忆 #1448
• Made sure Checks passed
• Tests have been provided