split_continuous_references: only split actual reference tags

[simpleqt]

Fixes #2446

Summary

split_continuous_references in src/memos/mem_os/utils/reference_utils.py converts any single [...] block containing a comma into split tags, even when the content is not a reference list:

>>> split_continuous_references("The set is [apple, banana] here")
'The set is [apple][banana] here'    # plain prose corrupted

Fix

Require every item to match the reference-tag shape int:id before splitting:

The set is [apple, banana] here     -> unchanged
[1:92ff35fb, 4:bfe6f044]            -> [1:92ff35fb][4:bfe6f044]   (unchanged behavior)

This tightens the shape guard added in #2418 (which required a comma-separated list but did not check the item shapes).

Regression test

Added tests/mem_os/test_reference_split_shape.py:

• red/green verified: the prose cases fail against the old implementation and pass with the fix
• covers split/no-space reference lists (unchanged behavior) plus plain lists, number lists, and bracketless text
• ruff check (repo-pinned 0.11.8) passes on both touched files

AI Disclosure

• Tool(s): Claude (ZCode CLI)
• Used for: debugging assistance, code suggestions, and drafting this PR description. All changes human-reviewed.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[Memtensor-AI]

🤖 Open Code Review

Target: PR #2447
Task: 5669ea117d4e38f7
Base: main
Head: sq/ref-split-shape

🔍 OpenCodeReview found 1 issue(s) in this PR.

---

1. src/memos/mem_os/utils/reference_utils.py (L46-L50)

The import re statement is placed mid-function, after the early-return guards. While Python caches module imports so this is not a runtime correctness issue, placing imports inside a function body (especially after executable statements) violates PEP 8 and makes the dependency harder to discover. Since re is already used in the sibling function process_streaming_references_complete via a local import, the cleanest fix is to hoist import re to the top of the file alongside the existing import.

💡 Suggested Change

Before:

    import re

    if not re.fullmatch(
        r"\s*\d+:[^,\s]+(?:,\s*\d+:[^,\s]+)*\s*", content_between_brackets
    ):

After:

# At the top of the file, alongside the existing import:
import re
from memos.memories.textual.item import (
    TextualMemoryItem,
)

# Then inside split_continuous_references, remove the local import re and keep:
    if not re.fullmatch(
        r"\s*\d+:[^,\s]+(?:,\s*\d+:[^,\s]+)*\s*", content_between_brackets
    ):

Generated by cloud-assistant via Open Code Review.

[Memtensor-AI]

✅ Automated Test Results: PASSED

All tests passed (7/7 executed). memos_python_core/changed-repo-python: 7/7. Duration: 8s [advisory, non-gating] AI-generated tests on branch test/auto-gen-5669ea117d4e38f7-20261001000157: 54/54 passed — these do NOT affect the PR verdict; review the branch manually.

Branch: sq/ref-split-shape