Self-hosted Hive deployments can configure S3 object storage and a vector database

Operators running the Hive server from ghcr.io/tuist/hive:0.5.0 can now point Hive at their own S3-compatible object store and at a separate vector database service, instead of relying on any bundled defaults. The server ships an S3 client supporting signed put, get, head, delete, public URL and bucket helpers, and exposes runtime configuration for both pieces through the HIVE_S3_* environment variables (endpoint, region, bucket, credentials) and HIVE_OPENDATA_VECTOR_URL. A new optional Helm subchart deploys opendata-vector alongside Hive, reusing the same S3 bucket and credentials and placing the index under a configurable vector.storage.prefix with a PVC cache, so a full self-hosted setup no longer requires bringing those pieces up by hand.