fix(memory): add mem0ai 2.0 API compatibility for Mem0LongTermMemory

[octo-patch]

Fixes #1511

Problem

mem0ai 2.0 introduced a breaking change to both search() and add(): entity identifiers (user_id, agent_id, run_id) are no longer accepted as top-level parameters. They must now be passed inside a filters dict:

Version	Call Style
v1.x	client.search(query, user_id="alice")
v2.0	client.search(query, filters={"user_id": "alice"})

This caused Mem0LongTermMemory to raise:

Error retrieving memory: Top-level entity parameters frozenset({'user_id', 'run_id', 'agent_id'}) are not supported in search(). Use filters={'user_id': '...'} instead.

Solution

• Add _detect_mem0_v2() static method to detect the installed mem0ai version.
• Add _build_entity_filters() instance method that returns the correct kwargs format:
  • mem0ai >= 2.0.0: {"filters": {"user_id": ..., ...}} (with AND for multiple identifiers)
  • mem0ai < 2.0.0: {"user_id": ..., "agent_id": ..., "run_id": ...}
• Update all three call sites (retrieve(), retrieve_from_memory(), _mem0_record()) to use _build_entity_filters().

This preserves full backward compatibility with older mem0ai versions.

Testing

Verified the logic manually against the mem0ai 2.0 documentation. The version check uses the existing packaging dependency already imported in this module.