Overview
dsh-semantic-search
A DeepSeek Harness (dsh) plugin for local semantic code search. Builds a fragment-level index of a workspace (symbol-aware chunking for many languages), computes local embeddings (lightweight lexical TF-IDF feature vectors by default, or a configurable OpenAI-compatible embedding endpoint), and answers natural-language queries with hybrid retrieval (vector cosine + BM25, fused with reciprocal-rank fusion). Ships dsh tools sema_search / sema_reindex / sema_stats plus a stand-alone CLI.
README / EN
Package documentation
Registry summary
A DeepSeek Harness (dsh) plugin for local semantic code search. Builds a fragment-level index of a workspace (symbol-aware chunking for many languages), computes local embeddings (lightweight lexical TF-IDF feature vectors by default, or a configurable OpenAI-compatible embedding endpoint), and answers natural-language queries with hybrid retrieval (vector cosine + BM25, fused with reciprocal-rank fusion). Ships dsh tools sema_search / sema_reindex / sema_stats plus a stand-alone CLI.
dsh.pub verifies the pinned bundle contract, runtime facts, and distribution semantics. The complete README remains in the source repository.
Read the full README on GitHub