use crate::errors::AppError;
use crate::paths::AppPaths;
use crate::pragmas::{apply_connection_pragmas, apply_init_pragmas, ensure_wal_mode};
use rusqlite::Connection;
use std::path::Path;
pub fn register_vec_extension() {}
pub fn open_rw(path: &Path) -> Result<Connection, AppError> {
let conn = Connection::open(path)?;
apply_connection_pragmas(&conn)?;
apply_secure_permissions(path);
adopt_embedding_dim(&conn);
Ok(conn)
}
fn adopt_embedding_dim(conn: &Connection) {
if crate::constants::embedding_dim_from_env().is_some() {
return;
}
if let Ok(value) = conn.query_row(
"SELECT value FROM schema_meta WHERE key = 'dim'",
[],
|row| row.get::<_, String>(0),
) {
if let Ok(dim) = value.parse::<usize>() {
crate::constants::set_active_embedding_dim(dim);
}
}
}
pub fn ensure_schema(conn: &mut Connection) -> Result<(), AppError> {
crate::migrations::runner()
.run(conn)
.map_err(|e| AppError::Internal(anyhow::anyhow!("migration failed: {e}")))?;
conn.execute_batch(&format!(
"PRAGMA user_version = {};",
crate::constants::SCHEMA_USER_VERSION
))?;
Ok(())
}
pub fn ensure_db_ready(paths: &AppPaths) -> Result<(), AppError> {
register_vec_extension();
paths.ensure_dirs()?;
let db_existed = paths.db.exists();
if !db_existed {
tracing::info!(target: "storage",
path = %paths.db.display(),
schema_version = crate::constants::CURRENT_SCHEMA_VERSION,
"creating database (auto-init)"
);
}
let mut conn = open_rw(&paths.db)?;
if !db_existed {
apply_init_pragmas(&conn)?;
}
let current_user_version: i64 = conn
.query_row("PRAGMA user_version", [], |row| row.get(0))
.unwrap_or(0);
let target_user_version = crate::constants::SCHEMA_USER_VERSION;
if current_user_version < target_user_version {
if db_existed {
tracing::warn!(target: "storage",
from = current_user_version,
to = target_user_version,
path = %paths.db.display(),
"auto-migrating database schema"
);
}
crate::migrations::runner()
.run(&mut conn)
.map_err(|e| AppError::Internal(anyhow::anyhow!("auto-migration failed: {e}")))?;
conn.execute_batch(&format!("PRAGMA user_version = {target_user_version};"))?;
if !db_existed {
insert_default_schema_meta(&conn)?;
}
ensure_wal_mode(&conn)?;
}
crate::commands::migrate::ensure_v013_tables_exist(&conn)?;
sync_embedding_dim_meta(&conn)?;
Ok(())
}
fn sync_embedding_dim_meta(conn: &Connection) -> Result<(), AppError> {
let db_dim: Option<usize> = conn
.query_row(
"SELECT value FROM schema_meta WHERE key = 'dim'",
[],
|row| row.get::<_, String>(0),
)
.ok()
.and_then(|v| v.parse::<usize>().ok());
if let Some(env_dim) = crate::constants::embedding_dim_from_env() {
if db_dim != Some(env_dim) {
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('dim', ?1)",
rusqlite::params![env_dim.to_string()],
)?;
}
return Ok(());
}
match db_dim {
Some(dim) => crate::constants::set_active_embedding_dim(dim),
None => {
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('dim', ?1)",
rusqlite::params![crate::constants::embedding_dim().to_string()],
)?;
}
}
Ok(())
}
fn insert_default_schema_meta(conn: &Connection) -> Result<(), AppError> {
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('schema_version', ?1)",
rusqlite::params![crate::constants::CURRENT_SCHEMA_VERSION.to_string()],
)?;
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('model', 'multilingual-e5-small')",
[],
)?;
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('dim', ?1)",
rusqlite::params![crate::constants::embedding_dim().to_string()],
)?;
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('created_at', CAST(unixepoch() AS TEXT))",
[],
)?;
conn.execute(
"INSERT OR REPLACE INTO schema_meta (key, value) VALUES ('sqlite-graphrag_version', ?1)",
rusqlite::params![crate::constants::SQLITE_GRAPHRAG_VERSION],
)?;
Ok(())
}
#[allow(unused_variables)]
fn apply_secure_permissions(path: &Path) {
#[cfg(unix)]
{
use std::os::unix::fs::PermissionsExt;
let candidates = [
path.to_path_buf(),
path.with_extension(format!(
"{}-wal",
path.extension()
.and_then(|e| e.to_str())
.unwrap_or("sqlite")
)),
path.with_extension(format!(
"{}-shm",
path.extension()
.and_then(|e| e.to_str())
.unwrap_or("sqlite")
)),
];
for file in candidates.iter() {
if file.exists() {
if let Ok(meta) = std::fs::metadata(file) {
let mut perms = meta.permissions();
perms.set_mode(0o600);
let _ = std::fs::set_permissions(file, perms);
}
}
}
}
#[cfg(windows)]
{
tracing::debug!(target: "storage",
path = %path.display(),
"skipping Unix mode 0o600 on Windows; NTFS DACL default is private-to-user"
);
}
}
pub fn open_ro(path: &Path) -> Result<Connection, AppError> {
let conn = Connection::open_with_flags(
path,
rusqlite::OpenFlags::SQLITE_OPEN_READ_ONLY | rusqlite::OpenFlags::SQLITE_OPEN_URI,
)?;
conn.execute_batch("PRAGMA foreign_keys = ON;")?;
adopt_embedding_dim(&conn);
Ok(conn)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
#[serial_test::serial(env)]
fn open_rw_adopts_schema_meta_dim() {
let dir = tempfile::tempdir().expect("tempdir");
let db = dir.path().join("g43.sqlite");
{
let conn = Connection::open(&db).expect("create seed db");
conn.execute_batch(
"CREATE TABLE schema_meta (key TEXT PRIMARY KEY, value TEXT);
INSERT INTO schema_meta VALUES ('dim', '128');",
)
.expect("seed schema_meta");
}
std::env::remove_var("SQLITE_GRAPHRAG_EMBEDDING_DIM");
let _conn = open_rw(&db).expect("open_rw");
let adopted = crate::constants::embedding_dim();
crate::constants::set_active_embedding_dim(crate::constants::DEFAULT_EMBEDDING_DIM);
assert_eq!(adopted, 128, "open_rw must adopt the recorded db dim (G43)");
}
#[test]
#[serial_test::serial(env)]
fn open_ro_adopts_schema_meta_dim() {
let dir = tempfile::tempdir().expect("tempdir");
let db = dir.path().join("g43-ro.sqlite");
{
let conn = Connection::open(&db).expect("create seed db");
conn.execute_batch(
"CREATE TABLE schema_meta (key TEXT PRIMARY KEY, value TEXT);
INSERT INTO schema_meta VALUES ('dim', '256');",
)
.expect("seed schema_meta");
}
std::env::remove_var("SQLITE_GRAPHRAG_EMBEDDING_DIM");
let _conn = open_ro(&db).expect("open_ro");
let adopted = crate::constants::embedding_dim();
crate::constants::set_active_embedding_dim(crate::constants::DEFAULT_EMBEDDING_DIM);
assert_eq!(adopted, 256, "open_ro must adopt the recorded db dim (G43)");
}
#[test]
#[serial_test::serial(env)]
fn env_override_wins_over_schema_meta_dim() {
let dir = tempfile::tempdir().expect("tempdir");
let db = dir.path().join("g43-env.sqlite");
{
let conn = Connection::open(&db).expect("create seed db");
conn.execute_batch(
"CREATE TABLE schema_meta (key TEXT PRIMARY KEY, value TEXT);
INSERT INTO schema_meta VALUES ('dim', '128');",
)
.expect("seed schema_meta");
}
std::env::set_var("SQLITE_GRAPHRAG_EMBEDDING_DIM", "96");
let _conn = open_rw(&db).expect("open_rw");
let adopted = crate::constants::embedding_dim();
std::env::remove_var("SQLITE_GRAPHRAG_EMBEDDING_DIM");
crate::constants::set_active_embedding_dim(crate::constants::DEFAULT_EMBEDDING_DIM);
assert_eq!(adopted, 96, "env override must win over the db dim (G43)");
}
#[test]
#[serial_test::serial(env)]
fn open_rw_on_virgin_db_is_a_noop() {
let dir = tempfile::tempdir().expect("tempdir");
let db = dir.path().join("g43-virgin.sqlite");
std::env::remove_var("SQLITE_GRAPHRAG_EMBEDDING_DIM");
crate::constants::set_active_embedding_dim(crate::constants::DEFAULT_EMBEDDING_DIM);
let _conn = open_rw(&db).expect("open_rw on virgin db must not fail");
assert_eq!(
crate::constants::embedding_dim(),
crate::constants::DEFAULT_EMBEDDING_DIM,
"virgin db must keep the compiled default (G43)"
);
}
}