qdrant / qdrant-txtai

An integration of Qdrant ANN vector database backend with txtai
Apache License 2.0
23 stars 2 forks source link

Module doesn't work with txtai 7. #7

Closed dustyatx closed 6 months ago

dustyatx commented 7 months ago

Tried to install and it broken my python environment for txtai.. I realized after that this is pinned to v5. I tried bumping up pyproject.toml got it to install but it didn't write anything to Qdrant..

Any chance you'd revisit this and update it?

kacperlukawski commented 7 months ago

Hello @dustyatx We definitely need to have a closer look, but this integration was not actively maintained for a while. We are open to contributions if you were successful in using it with the most recent version of txtai.

miary commented 6 months ago

Confirming that qdrant-txtai does not work with version 7. Really too bad, it would have been so powerful to combine both!!

kacperlukawski commented 6 months ago

@dustyatx @miary We've just released 2.0.0. Could you please upgrade?

miary commented 6 months ago

Perfect!!! Excellent upgrade. I tested qdrant-txtai 2.0.0, qdrant , run qdrant latest through Docker, txtai 7.1.0. The out-of-the-box test script works really great: import txtai import os

os.environ["config"] = "app.yml"

embeddings = txtai.Embeddings() embeddings.index(["Correct", "Not what we hoped"]) result = embeddings.search("positive", 1) print(result)

However, I could not find any way to query an existing collection. I tried this as app.yml but it didn't work: embeddings: path: sentence-transformers/all-MiniLM-L6-v2 backend: qdrant_txtai.ann.qdrant.Qdrant metric: cosine qdrant: host: localhost port: 6333 grpc_port: 6334 prefer_grpc: true collection: startups search_params: hnsw_ef: 256

What is the way to query from an existing collection?