Skip to content

[Bug]: FalkorDB adapter missing graph_database_port parameter in create_graph_engine() #2107

@tsepmud1-cloud

Description

@tsepmud1-cloud

Bug Description

Bug Description

When using FalkorDB with Cognee, the prune_system() operation fails with redis.exceptions.ResponseError: unknown command 'GRAPH.LIST' because the create_graph_engine() function doesn't pass the graph_database_port parameter to FalkorDB adapters.

Root Cause

In cognee/infrastructure/databases/graph/get_graph_engine.py, the create_graph_engine() function creates FalkorDB adapters using:

return adapter(
    graph_database_url=graph_database_url,
    graph_database_username=graph_database_username,
    graph_database_password=graph_database_password,
    database_name=graph_database_name,
)

But it's missing the graph_database_port parameter. The FalkorDBAdapter.__init__() defaults to port 6379, but users may configure different ports.

Workaround

Skip the prune operations or modify Docker port mapping to use 6379.

Steps to Reproduce

  1. Set up FalkorDB in Docker with port mapping (e.g., 6380:6379)
  2. Configure Cognee to use FalkorDB with custom port:
    config.set_graph_db_config({
        "graph_database_provider": "falkor",
        "graph_database_url": "host.docker.internal",
        "graph_database_port": "6380",  # Custom port
        # ... other config
    })
  3. Call await cognee.prune.prune_system(metadata=True)

Expected Behavior

The prune operation should work correctly and clean up graphs using the configured port.

Actual Behavior

Fails with:

redis.exceptions.ResponseError: unknown command 'GRAPH.LIST', with args beginning with:

Environment

  • Cognee version: 0.5.2
  • FalkorDB version: 1.6.7
  • Python version: 3.11
  • Docker setup: FalkorDB container mapped to host port 6380

Logs/Error Messages

Additional Context

Direct FalkorDB connection works fine:

from falkordb import FalkorDB
client = FalkorDB(host="host.docker.internal", port=6380)
print(client.list_graphs())  # Works: ['doc_graph_rag']

The issue occurs specifically during prune operations when a new graph engine is created for dataset deletion.

Pre-submission Checklist

  • I have searched existing issues to ensure this bug hasn't been reported already
  • I have provided a clear and detailed description of the bug
  • I have included steps to reproduce the issue
  • I have included my environment details

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions