huawei-cloud-openviking-embedding-switch
Switch OpenViking to a local embedding model.
What it does
Switches OpenViking's embedding configuration to a local endpoint, rebuilds incompatible vector indexes, and verifies the restarted server.
When to use it
Use when changing OpenViking's embedding model or dimension, fixing a dimension mismatch, or rebuilding its vector index after a model change.
How to use it
Give it an embedding model name, llama-server port, and vector dimension; it updates the local OpenViking configuration and restarts its sandboxed server.
What you provide
- the embedding model name
- the llama-server port
- the vector dimension
Uses
Good to know
- cannot be undone: ${SANDBOX_DIR}/data/vectordb/context; /workspace/process_dir/openviking-server.log
- may misreport its own work
What you need · 6
Requires job-env-manager running at http://127.0.0.1:8090.
Requires an already deployed OpenViking environment whose state is running.
Access · 7
This skill
${SANDBOX_DIR}/data/vectordb/context
Deletes incompatible OpenViking vector indexes (irreversible)
/workspace/process_dir/openviking-server.log
Overwrites the OpenViking server log (irreversible)