fix: Postgres export_graph ignores pagination and filters

[Kwizii]

Bug Description

When GRAPH_DB_BACKEND=postgres, PostgresGraphDB.export_graph accepted page, page_size, memory_type, status, and filter only through **kwargs but silently ignored them. Its SQL had no LIMIT/OFFSET and no filtering beyond user_name, and total_nodes was set to the length of the returned page.

As a result, every consumer of POST /get_memory (which passes pagination and tag filters down to text_mem.get_all → export_graph) received the full memory list on every page, with an incorrect total — pagination and tag filtering were both broken.

How to Reproduce

1. Run MemOS Product API with GRAPH_DB_BACKEND=postgres.
2. Call POST /product/get_memory with page, page_size, and/or tag filters.
3. Observe that all memories are returned regardless of page, and total_nodes matches the page size instead of the filtered total.

Environment

• Python: 3.11
• Operating System: Linux (Docker)
• MemOS: main branch
• Graph backend: postgres + pgvector

Additional Context

• I am willing to implement this myself and have a patch ready on fork branch fix/postgres-export-graph-pagination (PR fix: support pagination and filters in Postgres export_graph #2268).

[Memtensor-AI]

🤖 AutoDev has picked up this issue and started working on it.

Task ID: c1bb13ac9253908e
Working branch: bugfix/autodev-2271-20260822151212558
Target branch: latest dev* branch
Workflow: opsp (analysis → coding → testing → PR)

I will post the PR link here once done. If I need more information, I will ask in the comments.

[Memtensor-AI]

✅ AutoDev task c1bb13ac9253908e completed.

Summary: Fixed issue #2271: PostgresGraphDB.export_graph was accepting page, page_size, memory_type, status, and filter only through **kwargs and silently ignoring them, so consumers of POST /product/get_memory (via TreeTextMemory.get_all) received the full memory list on every page with an incorrect total_nodes.

The fix promotes page / page_size / memory_type / status / filter to first-class parameters, mirroring the Neo4j and PolarDB signatures. The WHERE clause is now built from user_name + memory_type + status + filter (reusing _build_filter_where_clause for structured tag/property predicates); default status excludes soft-deleted nodes to match Neo4j. Node results are sorted by created_at DESC, id DESC, and paginated with LIMIT/OFFSET when both page and page_size are supplied. An independent COUNT(*) is issued so total_nodes reflects the filtered total rather than the page size. The edge query is tightened so only edges with both endpoints in the returned node page are returned, avoiding dangling references under pagination.

Tests: added tests/graph_dbs/test_postgres_export_graph.py with 7 cases covering LIMIT/OFFSET emission, memory_type filter propagation, default-status-excludes-deleted, explicit status list, tag filter reaching SQL, and the total_nodes regression. pytest tests/memories/textual/ tests/graph_dbs/ -q -> 98 passed / 3 skipped. ruff format + ruff check clean.

Base branch: main
Branch: bugfix/autodev-2271-20260822151212558
Commit: a1f8f47
PR: #2274
Assigned to: @CarltonXiang
Reviewers: @wustzdy