Files
minne/retrieval-pipeline/src/lib.rs
T

276 lines
8.6 KiB
Rust

pub mod answer_retrieval;
pub mod pipeline;
pub mod query;
pub mod reranking;
pub mod scoring;
use std::sync::Arc;
pub use scoring::RetrievalCandidate;
use common::{
error::AppError,
storage::{
db::SurrealDbClient,
types::{knowledge_entity::KnowledgeEntity, text_chunk::TextChunk},
},
};
use reranking::RerankerLease;
use tracing::instrument;
/// Result of a retrieval run.
///
/// Chunk retrieval is always performed; entities are only present when the caller
/// requested entity resolution via [`RetrievalConfig::with_entities`].
#[derive(Debug)]
pub enum RetrievalOutput {
Chunks(Vec<RetrievedChunk>),
WithEntities {
chunks: Vec<RetrievedChunk>,
entities: Vec<RetrievedEntity>,
},
}
pub use pipeline::{
retrieved_entities_to_json, Diagnostics, RetrievalConfig, RetrievalParams, RetrievalTuning,
StageKind, StageTimings,
};
pub use query::normalize_fts_terms;
pub use scoring::{reciprocal_rank_fusion, RrfConfig, Scored};
/// Round a score to three decimal places for JSON output.
pub(crate) fn round_score(value: f32) -> f64 {
(f64::from(value) * 1000.0).round() / 1000.0
}
// Captures a supporting chunk plus its fused retrieval score for downstream prompts.
#[derive(Debug, Clone)]
pub struct RetrievedChunk {
pub chunk: Arc<TextChunk>,
pub score: f32,
}
// Knowledge entity resolved from retrieved chunks, enriched with its contributing chunks.
#[derive(Debug, Clone)]
pub struct RetrievedEntity {
pub entity: KnowledgeEntity,
pub score: f32,
pub chunks: Arc<Vec<RetrievedChunk>>,
}
/// Run chunk-first hybrid retrieval for `input_text`, optionally resolving owning entities.
#[instrument(skip_all, fields(user_id))]
pub async fn retrieve(
db_client: &SurrealDbClient,
embedding_provider: &common::utils::embedding::EmbeddingProvider,
input_text: &str,
user_id: &str,
config: RetrievalConfig,
reranker: Option<RerankerLease>,
) -> Result<RetrievalOutput, AppError> {
let params = pipeline::RetrievalParams {
db_client,
embedding_provider,
input_text,
user_id,
config,
reranker,
};
pipeline::execute(params).await
}
#[cfg(test)]
mod tests {
use super::*;
use anyhow::{self};
use common::storage::indexes::ensure_runtime;
use common::storage::types::knowledge_entity::{KnowledgeEntity, KnowledgeEntityType};
use common::storage::types::system_settings::SystemSettings;
use common::utils::embedding::EmbeddingProvider;
use uuid::Uuid;
fn test_embedding_provider() -> EmbeddingProvider {
EmbeddingProvider::new_hashed(3).unwrap_or_else(|_| unreachable!())
}
fn test_embedding() -> Vec<f32> {
vec![0.9, 0.1, 0.0]
}
fn chunk_embedding_primary() -> Vec<f32> {
vec![0.85, 0.15, 0.0]
}
fn chunk_embedding_secondary() -> Vec<f32> {
vec![0.2, 0.8, 0.0]
}
async fn configure_embedding_dimension(
db: &SurrealDbClient,
dimension: u32,
) -> anyhow::Result<()> {
let mut settings = SystemSettings::get_current(db).await?;
settings.embedding_dimensions = dimension;
SystemSettings::update(db, settings).await?;
Ok(())
}
async fn setup_test_db() -> anyhow::Result<SurrealDbClient> {
let namespace = "test_ns";
let database = &Uuid::new_v4().to_string();
let db = SurrealDbClient::memory(namespace, database).await?;
db.apply_migrations().await?;
configure_embedding_dimension(&db, 3).await?;
ensure_runtime(&db, 3).await?;
Ok(db)
}
#[tokio::test]
async fn test_chunk_retrieval_returns_chunks() -> anyhow::Result<()> {
let db = setup_test_db().await?;
let user_id = "test_user";
let chunk = TextChunk::new(
"source_1".into(),
"Tokio uses cooperative scheduling for fairness.".into(),
user_id.into(),
);
TextChunk::store_with_embedding(chunk.clone(), chunk_embedding_primary(), 3, &db).await?;
let embedding_provider = test_embedding_provider();
let params = pipeline::RetrievalParams {
db_client: &db,
embedding_provider: &embedding_provider,
input_text: "Rust concurrency async tasks",
user_id,
config: RetrievalConfig::default(),
reranker: None,
};
let results = pipeline::run_with_embedding(params, test_embedding()).await?;
let chunks = match results {
RetrievalOutput::Chunks(items) => items,
RetrievalOutput::WithEntities { .. } => {
anyhow::bail!("expected chunk results, got entities")
}
};
assert!(!chunks.is_empty(), "Expected at least one retrieval result");
assert!(
chunks
.first()
.is_some_and(|c| c.chunk.chunk.contains("Tokio")),
"Expected chunk about Tokio"
);
Ok(())
}
#[tokio::test]
async fn test_chunk_retrieval_returns_chunks_from_multiple_sources() -> anyhow::Result<()> {
let db = setup_test_db().await?;
let user_id = "multi_source_user";
let primary_chunk = TextChunk::new(
"primary_source".into(),
"Rust async tasks use Tokio's cooperative scheduler.".into(),
user_id.into(),
);
let secondary_chunk = TextChunk::new(
"secondary_source".into(),
"Tokio's scheduler manages task fairness across executors.".into(),
user_id.into(),
);
TextChunk::store_with_embedding(primary_chunk, chunk_embedding_primary(), 3, &db).await?;
TextChunk::store_with_embedding(secondary_chunk, chunk_embedding_secondary(), 3, &db)
.await?;
let embedding_provider = test_embedding_provider();
let params = pipeline::RetrievalParams {
db_client: &db,
embedding_provider: &embedding_provider,
input_text: "Rust concurrency async tasks",
user_id,
config: RetrievalConfig::default(),
reranker: None,
};
let results = pipeline::run_with_embedding(params, test_embedding()).await?;
let chunks = match results {
RetrievalOutput::Chunks(items) => items,
RetrievalOutput::WithEntities { .. } => {
anyhow::bail!("expected chunk results, got entities")
}
};
assert!(chunks.len() >= 2, "Expected chunks from multiple sources");
assert!(
chunks.iter().any(|c| c.chunk.source_id == "primary_source"),
"Should include primary source chunk"
);
assert!(
chunks
.iter()
.any(|c| c.chunk.source_id == "secondary_source"),
"Should include secondary source chunk"
);
Ok(())
}
#[tokio::test]
async fn test_with_entities_resolves_owning_entities() -> anyhow::Result<()> {
let db = setup_test_db().await?;
let user_id = "entity_user";
let chunk = TextChunk::new(
"entity_source".into(),
"Async Rust programming uses the Tokio runtime for concurrent tasks.".into(),
user_id.into(),
);
TextChunk::store_with_embedding(chunk.clone(), chunk_embedding_primary(), 3, &db).await?;
let entity = KnowledgeEntity::new(
"entity_source".into(),
"Tokio Runtime".into(),
"Async runtime for Rust".into(),
KnowledgeEntityType::Document,
None,
user_id.into(),
);
db.store_item(entity).await?;
let embedding_provider = test_embedding_provider();
let params = pipeline::RetrievalParams {
db_client: &db,
embedding_provider: &embedding_provider,
input_text: "async rust programming",
user_id,
config: RetrievalConfig::with_entities(),
reranker: None,
};
let results = pipeline::run_with_embedding(params, test_embedding()).await?;
let RetrievalOutput::WithEntities { chunks, entities } = results else {
anyhow::bail!("expected WithEntities output");
};
assert!(!chunks.is_empty(), "Should return chunks");
assert!(
entities.iter().any(|e| e.entity.name == "Tokio Runtime"),
"Should resolve the entity owning the retrieved chunk"
);
assert!(
entities
.iter()
.find(|e| e.entity.name == "Tokio Runtime")
.is_some_and(|e| !e.chunks.is_empty()),
"Resolved entity should carry its contributing chunks"
);
Ok(())
}
}