pkgs/store/vecid

pkgs/store/vecid Use locality-sensitive hashing and bucket clustering to establish a stable identity for vectors. The current main consumer is the audio voiceprint detector, which is used to classify similar speaker embeddings into existing identities or create new identities.

Go API References

Core structure and implementation

Symbol	Function
Config	Configure dimension, hash bits, distance and clustering behavior.
Registry / New	Register, match and maintain vector identities.
Bucket	Stores hash bucket and identity candidates.
Hasher	Use deterministic random hyperplanes to generate vector hash.
PlanesFile	Expresses hash planes that can be saved and restored.
NewHasher / NewHasherFromPlanes / NewHasherFromJSON	Create or restore Hasher.
Store	Define identity, bucket and compact state persistence.
MemoryStore	Provides in-process Registry store.

Ownership Boundary

VecID is not responsible for capturing audio, generating speaker embedding, or interpreting the user meaning of identity. It has different goals than vecstore: vecstore returns similar vector matches, vecid maintains a continuously updated identity registry.