Topoteretes
Replies: 31 comments
-
|
hey <@605959891273908243> , could you try removing the vector_db_name parameter from your vector database configuration:
|
Beta Was this translation helpful? Give feedback.
-
|
You could also check our community code example here for the correct Milvus adaptor usage. |
Beta Was this translation helpful? Give feedback.
-
|
let us know if this helps. |
Beta Was this translation helpful? Give feedback.
-
|
i try this but 2026-01-15T16:53:05.132089 [info ] Switching to TikToken default tokenizer. [LiteLLMEmbeddingEngine] |
Beta Was this translation helpful? Give feedback.
-
|
the same |
Beta Was this translation helpful? Give feedback.
-
|
Oh nice and i have a question |
Beta Was this translation helpful? Give feedback.
-
|
For falkor? This versión of cognee work wirh falkro ? |
Beta Was this translation helpful? Give feedback.
-
|
Falkor?? |
Beta Was this translation helpful? Give feedback.
-
|
. |
Beta Was this translation helpful? Give feedback.
-
|
yes, falkor works with cognee 0.5.1 |
Beta Was this translation helpful? Give feedback.
-
|
you can use falkor as a hybrid store, both as a vector and a graph database, with Cognee, and it supports multi-user |
Beta Was this translation helpful? Give feedback.
-
|
So, in the latest version, it would not be possible to use both Falkor and Milvus. |
Beta Was this translation helpful? Give feedback.
-
|
Depends on what you need. You can run both with cognee 0.3.4 versions. But if you need multi user, then it's not possible to use Milvus |
Beta Was this translation helpful? Give feedback.
-
|
thanks <@506030217710665738> it works now |
Beta Was this translation helpful? Give feedback.
-
|
I just have a question about how to define the ontology because I define it but it doesn't respect it. It creates other nodes that have nothing to do with the ontology I defined. |
Beta Was this translation helpful? Give feedback.
-
|
<@605959891273908243> could you give me more details about how you added the ontology, and what is the outcome? In Cognee you can add an ontology in two different ways, either through and RDF/OWL file, or by definining it on a lower lvl with pydantic models. |
Beta Was this translation helpful? Give feedback.
-
|
When we get something back from the llm, we check if those entities exist in the ontology. If they do, we traverse the ontology graph for more nodes/information about said entitiy. Those entities that exist in the ontology are deemed "ontology-valid", and others are not (i.e. have that field false) |
Beta Was this translation helpful? Give feedback.
-
|
So, it is a possibility to have nodes that are not from the ontology, even if you pass the ontology |
Beta Was this translation helpful? Give feedback.
-
|
this |
Beta Was this translation helpful? Give feedback.
-
|
. |
Beta Was this translation helpful? Give feedback.
-
|
Yes, so it works kind of in the way I explained it. We query the llm, then check if the entities from the response exist in the ontology. If so, we traverse the ontology and mark the entities as ontology valid. If not, we just add them in the graph, and they are marked as not ontology valid |
Beta Was this translation helpful? Give feedback.
-
|
this is my graph |
Beta Was this translation helpful? Give feedback.
-
|
and I don't see the ontology working, so based on what you're telling me, it means that the LLM hasn't understood my ontology correctly. |
Beta Was this translation helpful? Give feedback.
-
|
check the details of those nodes, e.g. take one entity you have in the ontology and see if it's in the graph, and if that node has onotology_valid set to true |
Beta Was this translation helpful? Give feedback.
-
|
is false |
Beta Was this translation helpful? Give feedback.
-
|
oh no |
Beta Was this translation helpful? Give feedback.
-
|
🙁 |
Beta Was this translation helpful? Give feedback.
-
|
How do you see my ontology? |
Beta Was this translation helpful? Give feedback.
-
|
<@506030217710665738> please |
Beta Was this translation helpful? Give feedback.
-
|
or how I would implement it with pydantic? |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
=Context: I am using cognee v0.5.1 and trying to use Milvus (Zilliz Cloud) as the vector database and FalkorDB as the graph database. I am registering the community adapters as required by the new version.
The Problem: When calling cognee.add(), the system fails while trying to initialize the vector engine. Even though I have set VECTOR_DB_PROVIDER = "milvus", the traceback shows that Cognee is internally calling create_pgvector_db_and_tables and then failing with a TypeError because it passes database_name to the MilvusAdapter, which doesn't support it.
traceback:
❌ Error durante la prueba: MilvusAdapter.init() got an unexpected keyword argument 'database_name'
Traceback (most recent call last):
File "/home/cristofer/test/main.py", line 71, in test_system
await cognee.add(linea)
File "/home/cristofer/test/lib/python3.12/site-packages/cognee/api/v1/add/add.py", line 188, in add
await setup()
File "/home/cristofer/test/lib/python3.12/site-packages/cognee/modules/engine/operations/setup.py", line 17, in setup
await create_pgvector_db_and_tables()
File "/home/cristofer/test/lib/python3.12/site-packages/cognee/infrastructure/databases/vector/pgvector/create_db_and_tables.py", line 8, in create_db_and_tables
vector_engine = get_vector_engine()
^^^^^^^^^^^^^^^^^^^
File "/home/cristofer/test/lib/python3.12/site-packages/cognee/infrastructure/databases/vector/get_vector_engine.py", line 7, in get_vector_engine
return create_vector_engine(**get_vectordb_context_config())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/cristofer/test/lib/python3.12/site-packages/cognee/infrastructure/databases/vector/create_vector_engine.py", line 48, in create_vector_engine
return adapter(
^^^^^^^^
TypeError: MilvusAdapter.init() got an unexpected keyword argument 'database_name'
This discussion was automatically pulled from Discord.
Beta Was this translation helpful? Give feedback.
All reactions