fix: sync Qdrant payload on update_node in Neo4j community edition and add schedule dep

[octo-patch]

Fixes #1469

Problem

Two bugs affecting users of Neo4j Community Edition:

1. update_node not syncing Qdrant payload (High severity)

Neo4jCommunityGraphDB inherits update_node from the parent Neo4jGraphDB, which only updates Neo4j. The community edition stores vectors in an external Qdrant database, so key metadata fields (status, tags, memory_type, content, sources) were never propagated to Qdrant when a node was updated. This caused Neo4j and Qdrant to diverge — for example, archived memories would remain visible in vector searches because Qdrant still had the old status: activated payload.

2. Missing schedule dependency in docker/requirements.txt (Critical severity)

reorganizer.py (Thread-3, which handles conflict detection and memory archiving) imports the schedule library. While schedule>=1.2.2 is declared in pyproject.toml and present in requirements-full.txt, it was absent from docker/requirements.txt. This caused an ImportError on Thread-3 startup when deploying with the standard Docker requirements, silently disabling conflict detection and memory archiving.

Solution

• neo4j_community.py: Override update_node in Neo4jCommunityGraphDB to first call the parent's Neo4j update, then sync the relevant fields to Qdrant via vec_db.update(). Sync failures are logged as warnings and do not abort the Neo4j update.

• docker/requirements.txt: Add schedule==1.2.2 (matching the version already pinned in requirements-full.txt).

Testing

• The update_node override follows the same Qdrant interaction pattern used in add_node and add_nodes_batch.
• The schedule version matches the constraint in pyproject.toml (>=1.2.2,<2.0.0) and the pin in requirements-full.txt.