torvalds-dev[bot]
commented
8 months ago Based on the provided information, the issue is about vectorizing data using Weaviate. Weaviate is an open-source vector database that allows you to turn your text, images, and more into a searchable vector database using state-of-the-art ML models.
To vectorize your data using Weaviate, you have two options:
-
Vectorize your data at import time: This means that as you import your data into Weaviate, it will automatically be converted into vectors. This is done using Weaviate's modules, which can use popular services and model hubs such as OpenAI, Cohere, or HuggingFace.
-
Upload your own vectors: If you have pre-computed vectors, you can upload them directly to Weaviate.
The relevant file to this issue is things_schema.json which seems to be a schema for the data that you want to vectorize. This schema includes classes such as "City", "Person", "Country", "Airport", "Airline", "Plane", and "Aircraft", and properties for each class.
To vectorize your data, you would need to follow these steps:
-
Install Weaviate: You can do this by following the instructions in the quickstart tutorial.
-
Import your data: You can do this by following the instructions in the Weaviate documentation. Make sure that your data matches the schema in
things_schema.json. -
Vectorize your data: If you want to vectorize your data at import time, you can do this by configuring Weaviate to use one of its modules. If you want to upload your own vectors, you can do this by following the instructions in the Weaviate documentation.
-
Query your data: Once your data is vectorized and stored in Weaviate, you can query it using Weaviate's GraphQL API.
If you encounter any issues during this process, you can ask for help in the Weaviate community on Slack or the forum.
torvalds.dev is analyzing the ticket