use std::path::Path;
use surrealdb::engine::local::SurrealKv;
use surrealdb::Surreal;
use crate::error::GraphError;
pub type Db = surrealdb::engine::local::Db;
const NAMESPACE: &str = "recall";
const DATABASE: &str = "graph";
pub async fn open(path: &Path) -> Result<Surreal<Db>, GraphError> {
let surreal_path = path.join("surreal");
std::fs::create_dir_all(&surreal_path)?;
let db: Surreal<Db> = Surreal::new::<SurrealKv>(surreal_path.to_str().unwrap()).await?;
db.use_ns(NAMESPACE).use_db(DATABASE).await?;
Ok(db)
}
pub async fn init_schema(db: &Surreal<Db>) -> Result<(), GraphError> {
db.query(
r#"
DEFINE TABLE IF NOT EXISTS entity SCHEMAFULL;
DEFINE FIELD IF NOT EXISTS name ON entity TYPE string;
DEFINE FIELD IF NOT EXISTS entity_type ON entity TYPE string;
DEFINE FIELD IF NOT EXISTS abstract ON entity TYPE string;
DEFINE FIELD IF NOT EXISTS overview ON entity TYPE string;
DEFINE FIELD IF NOT EXISTS content ON entity TYPE option<string>;
DEFINE FIELD IF NOT EXISTS attributes ON entity TYPE option<object> FLEXIBLE;
DEFINE FIELD IF NOT EXISTS embedding ON entity TYPE option<array<float>>;
DEFINE FIELD IF NOT EXISTS mutable ON entity TYPE bool DEFAULT true;
DEFINE FIELD IF NOT EXISTS access_count ON entity TYPE int DEFAULT 0;
DEFINE FIELD IF NOT EXISTS created_at ON entity TYPE datetime DEFAULT time::now();
DEFINE FIELD IF NOT EXISTS updated_at ON entity TYPE datetime DEFAULT time::now();
DEFINE FIELD IF NOT EXISTS source ON entity TYPE option<string>;
DEFINE INDEX IF NOT EXISTS entity_name ON entity FIELDS name;
DEFINE INDEX IF NOT EXISTS entity_type ON entity FIELDS entity_type;
DEFINE INDEX IF NOT EXISTS entity_vector ON entity FIELDS embedding HNSW DIMENSION 384 DIST COSINE;
-- Pipeline attribute indexes
DEFINE INDEX IF NOT EXISTS entity_pipeline_stage ON entity FIELDS attributes.pipeline_stage;
DEFINE INDEX IF NOT EXISTS entity_pipeline_status ON entity FIELDS attributes.pipeline_status;
DEFINE TABLE IF NOT EXISTS relates_to SCHEMAFULL TYPE RELATION;
DEFINE FIELD IF NOT EXISTS rel_type ON relates_to TYPE string;
DEFINE FIELD IF NOT EXISTS description ON relates_to TYPE option<string>;
DEFINE FIELD IF NOT EXISTS valid_from ON relates_to TYPE datetime DEFAULT time::now();
DEFINE FIELD IF NOT EXISTS valid_until ON relates_to TYPE option<datetime>;
DEFINE FIELD IF NOT EXISTS confidence ON relates_to TYPE float DEFAULT 1.0;
DEFINE FIELD IF NOT EXISTS source ON relates_to TYPE option<string>;
DEFINE INDEX IF NOT EXISTS rel_type_idx ON relates_to FIELDS rel_type;
DEFINE TABLE IF NOT EXISTS episode SCHEMAFULL;
DEFINE FIELD IF NOT EXISTS session_id ON episode TYPE string;
DEFINE FIELD IF NOT EXISTS timestamp ON episode TYPE datetime DEFAULT time::now();
DEFINE FIELD IF NOT EXISTS abstract ON episode TYPE string;
DEFINE FIELD IF NOT EXISTS overview ON episode TYPE option<string>;
DEFINE FIELD IF NOT EXISTS content ON episode TYPE option<string>;
DEFINE FIELD IF NOT EXISTS embedding ON episode TYPE option<array<float>>;
DEFINE FIELD IF NOT EXISTS log_number ON episode TYPE option<int>;
DEFINE FIELD IF NOT EXISTS extracted ON episode TYPE bool DEFAULT false;
-- Backfill: set extracted = false on episodes that predate the field
UPDATE episode SET extracted = false WHERE extracted IS NONE;
DEFINE INDEX IF NOT EXISTS episode_session ON episode FIELDS session_id;
DEFINE INDEX IF NOT EXISTS episode_time ON episode FIELDS timestamp;
DEFINE INDEX IF NOT EXISTS episode_vector ON episode FIELDS embedding HNSW DIMENSION 384 DIST COSINE;
"#,
)
.await?
.check()?;
Ok(())
}