Struct CodeIndexer 

pub struct CodeIndexer { /* private fields */ }

Orchestrates code indexing over a project tree.

CodeIndexer is the primary driver of the indexing pipeline. It walks the file tree, delegates per-file work to FileIndexWorker, and coordinates the Qdrant + SQLite writes via CodeStore.

Cloning and concurrency

CodeIndexer is not Clone — it is typically wrapped in an Arc and shared between the initial indexing task and the file watcher.

Examples

use std::sync::Arc;
use zeph_index::indexer::{CodeIndexer, IndexerConfig};
use zeph_index::store::CodeStore;

let indexer = CodeIndexer::new(store, provider, IndexerConfig::default());
let report = indexer.index_project(std::path::Path::new("."), None).await?;
println!("indexed {} files in {}ms", report.files_indexed, report.duration_ms);

Implementations

impl CodeIndexer

pub fn new( store: CodeStore, provider: Arc<AnyProvider>, config: IndexerConfig, ) -> Self

Create a new CodeIndexer.

The store and provider are cloned cheaply (reference-counted) across the concurrent file-processing tasks.

pub fn with_spawner(self, spawner: Arc<dyn BlockingSpawner>) -> Self

Attach a supervised blocking spawner for chunk_file tasks.

When set, each chunk_file call is routed through the spawner so it is visible in supervisor snapshots and subject to graceful shutdown.

Examples

use std::sync::Arc;
use zeph_index::indexer::{CodeIndexer, IndexerConfig};
use zeph_index::store::CodeStore;
use zeph_common::BlockingSpawner;

let indexer = CodeIndexer::new(store, provider, IndexerConfig::default())
    .with_spawner(spawner);

pub async fn index_project( &self, root: &Path, progress_tx: Option<&Sender<IndexProgress>>, ) -> Result<IndexReport>

Full project indexing with incremental change detection.

Errors

Returns an error if the embedding probe or collection setup fails.

pub async fn reindex_file(&self, root: &Path, abs_path: &Path) -> Result<usize>

Re-index a specific file (for file watcher).

Errors

Returns an error if reading, chunking, or embedding fails.