how to load later a graph / vectors after creation with cognify?

[LStromann]

=After some steps with cognee in python (local installation), I am wondereing how to re-use graphs / vectors I created with "cognify". I understand the result is stored in LanceDB / Kuzu automatically.
How can I use this later in other programs or even after a restart of the machine?
Sorry, if this is obvious to do, but unfortunately I did not find any hint on that topic yet.

This discussion was automatically pulled from Discord.

[LStromann]

hey <@1475220624849899593> after cognify() everything is already persisted LanceDB + Kuzu by default. On restart or in another program you just need to point to the same config and data directories.

As long as DATA_ROOT_DIRECTORY, SYSTEM_ROOT_DIRECTORY, and the same DB providers are set Cognee will automatically reconnect to the existing storage on first use.

[LStromann]

You can also check the docs for storage/config details: https://docs.cognee.ai/setup-configuration/graph-stores#data-location

[LStromann]

Thanks for the hints. Got it. So still a very practical example I cannot figure out yet: How do I get access to the graph etc. without running cognify() again?

After a first run to create and use the graph from a text (add, cognify, search):
Starting a new program then:
e.g. this does NOT work: await visualize_graph() --> html shows nothing
but:
await cognee.cognify()
await visualize_graph()
works.

[LStromann]

could you share the script for the new program where you call visualize_graph()?
It might be a config/path issue between runs.

[LStromann]

If you set data and system root directories in one program, make sure you set the same ones in another program. Cognee defaults to certain values in general

[LStromann]

<@1475220624849899593> just checking in were you able to resolve this using Andrej's suggestion?