Skip to main content

Crate vera2_core

Crate vera2_core 

Source
Expand description

Vera Core — parsing, indexing, retrieval, storage, and embedding engine.

This crate provides the core functionality for Vera’s code indexing and retrieval pipeline. It is consumed by vera-cli (CLI interface) and vera-mcp (MCP server).

§Module overview

  • parsing — Tree-sitter based source code parsing and AST extraction.
  • indexing — Building and maintaining the search index (BM25, vectors).
  • retrieval — Hybrid search, fusion, and reranking pipeline.
  • storage — Persistent storage backends (SQLite + sqlite-vec, Tantivy).
  • embedding — Embedding generation via external API providers.

Modules§

chunk_text
Shared text builders for semantic and lexical indexing.
config
Configuration types and defaults. Configuration types and defaults for Vera’s pipeline.
corpus
Corpus-aware file classification used by ranking and filtering.
discovery
File discovery: walk directory tree, respect .gitignore, exclude defaults.
embedding
Embedding generation via external API providers.
git_scope
Git-aware path scoping for changed-file workflows.
indexing
Index construction and maintenance.
local_models
parsing
Source code parsing using tree-sitter.
presentation
Output presentation helpers shared by the CLI and MCP frontends.
retrieval
Hybrid retrieval pipeline: BM25 + vector search, RRF fusion, reranking.
stats
Index statistics collection. Index statistics collection.
storage
Persistent storage backends for Vera’s index.
types
Shared types used across multiple modules. Shared types used across Vera’s core modules.

Structs§

CancellationToken
Cooperative cancellation shared across indexing pipeline stages.

Functions§

init_tls
Install the rustls crypto provider (ring). Safe to call multiple times.