Module handler 

HTTP handlers for the embedding service.

Submodules:

• common: shared input validation and readiness helpers.
• dense: POST /v1/embeddings (OpenAI-compatible dense embeddings).
• sparse: POST /v1/sparse-embeddings (BGE-M3 SPLADE-style sparse embeddings).
• both: POST /v1/embeddings:both (paired dense + sparse output in one pass).
• health: GET /health (readiness + tuning details).
• models: GET /v1/models (fleet discovery).

Modules

both 🔒

POST /v1/embeddings:both handler — dense + sparse embeddings in one pass.

common 🔒

Shared input validation and service-readiness helpers used by all handlers.

dense 🔒

POST /v1/embeddings handler — OpenAI-compatible dense embeddings.

health 🔒

GET /health handler — readiness status, worker counts, and tuning diagnostics.

models 🔒

GET /v1/models handler — OpenAI-compatible fleet discovery endpoint.

sparse 🔒

POST /v1/sparse-embeddings handler — BGE-M3 SPLADE-style sparse embeddings.

Functions

both_embeddings

Handles POST /v1/embeddings:both — returns dense and sparse embeddings in one pass.

dense_embeddings

Handles POST /v1/embeddings — returns dense (float32) embeddings.

health

Handles GET /health — returns readiness status, worker counts, and tuning diagnostics.

models

Returns an OpenAI-compatible models list confirming BGE-M3 is resident.

sparse_embeddings

Handles POST /v1/sparse-embeddings — returns sparse (SPLADE-style) embeddings.