Add REDIS vector search

[danellecline]

This will be helpful for multiple projects to revise training or processed data, e.g. CFE (SINKER and Planktivore), i2MAP, Biodiversity, UAV.

• Implement for the gridview panel only
• Create Algorithm https://github.com/mbari-org/argo-workflows/blob/main/workflows/retinanet_tator.yml* and https://github.com/mbari-org/argo-workflows/blob/main/src/inference/detectron_infer_tator.py serve as a simple example.
• Deploy algorithm using fixed naming convention to simplify integration, e.g. VSS
• Use model defaults as defined in respective model yaml files here: https://docs.mbari.org/internal/ai/projects/

vss:
  project: "901902-UAV"
  threshold: 0.1
  model: "/mnt/DeepSea-AI/models/UAV/mbari-uav-vits-b8-20251002/"

• Deploy FastAPI-VSS to a maximilian which has capacity both GPU and disk. Copy REDIS docker volumes from doris since most projects already have populated databases.