ory / lumen / 23916073376

70%

push

github

feat(config): add LUMEN_EMBED_DIMS/LUMEN_EMBED_CTX escape hatch for custom models (#88)

Allows users to run unlisted embedding models (quantized variants, local
fine-tunes, models on non-standard ports) without forking or modifying
KnownModels. Both variables are independently overridable — LUMEN_EMBED_DIMS
is required for unknown models; LUMEN_EMBED_CTX overrides context length for
any model and defaults to 8192 for unlisted ones.

Adds table-driven TestLoad covering: known model (no overrides, each variable
independently, both), unknown model (DIMS only, DIMS+CTX, neither → error,
CTX-only → error).

Documents new env vars and custom model usage in README.

Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

17 of 17 new or added lines in 1 file covered. (100.0%)

2655 of 3845 relevant lines covered (69.05%)

46.54 hits per line