introduce LruCache for transpilation

Author: KSDaemon

packages/cubejs-backend-native/Cargo.lock

@@ -48,6 +48,7 @@ tokio = { version = "1", features = ["full", "rt"] }
 uuid = { version = "1", features = ["v4"] }
 bytes = "1.5.0"
 regex = "1.10.2"
+lru = "0.13.0"
 
 [dependencies.neon]
 version = "=1"

packages/cubejs-backend-native/Cargo.toml

@@ -2,12 +2,15 @@ use crate::node_obj_deserializer::JsValueDeserializer;
 use crate::node_obj_serializer::NodeObjSerializer;
 use anyhow::anyhow;
 use cubetranspilers::{run_transpilers, TransformConfig, Transpilers};
+use lru::LruCache;
 use neon::context::{Context, FunctionContext, ModuleContext};
 use neon::prelude::{JsPromise, JsResult, JsValue, NeonResult};
 use neon::types::JsString;
 use serde::Deserialize;
 use std::collections::{HashMap, HashSet};
-use std::sync::{LazyLock, RwLock};
+use std::env;
+use std::num::NonZeroUsize;
+use std::sync::{LazyLock, Mutex};
 
 #[derive(Deserialize, Default, Clone, Debug)]
 #[serde(rename_all = "camelCase")]
@@ -26,8 +29,21 @@ pub struct TransformRequestConfig {
     pub meta_data: Option<TransformMetaData>,
 }
 
-static METADATA_CACHE: LazyLock<RwLock<HashMap<String, TransformMetaData>>> =
-    LazyLock::new(|| RwLock::new(HashMap::new()));
+/// It should be equal or more then number of internal libuv threads used by Neon
+/// By 01.2025 it defaults to 4. But maybe changed via `UV_THREADPOOL_SIZE` env var.
+/// `CUBEJS_TRANSPILER_METADATA_CACHE_SIZE` env var is provided for fine tuning.
+/// @see https://docs.libuv.org/en/v1.x/threadpool.html
+/// @see https://nodejs.org/api/cli.html#cli_uv_threadpool_size_size
+static DEFAULT_CACHE_SIZE: usize = 16;
+
+static METADATA_CACHE: LazyLock<Mutex<LruCache<String, TransformMetaData>>> = LazyLock::new(|| {
+    let cache_size = env::var("CUBEJS_TRANSPILER_METADATA_CACHE_SIZE")
+        .ok()
+        .and_then(|s| s.parse::<usize>().ok())
+        .and_then(NonZeroUsize::new)
+        .unwrap_or(NonZeroUsize::new(DEFAULT_CACHE_SIZE).unwrap());
+    Mutex::new(LruCache::new(cache_size))
+});
 
 pub fn register_module(cx: &mut ModuleContext) -> NeonResult<()> {
     cx.export_function("transpileJs", transpile_js)?;
@@ -46,7 +62,7 @@ pub fn transpile_js(mut cx: FunctionContext) -> JsResult<JsPromise> {
             let transform_config: TransformConfig = match transform_request_config {
                 Ok(data) => match data.meta_data {
                     Some(meta_data) => {
-                        let mut config_lock = METADATA_CACHE.write().unwrap();
+                        let mut config_lock = METADATA_CACHE.lock().unwrap();
                         let cache = TransformMetaData {
                             cube_names: meta_data.cube_names,
                             cube_symbols: meta_data.cube_symbols,
@@ -59,13 +75,13 @@ pub fn transpile_js(mut cx: FunctionContext) -> JsResult<JsPromise> {
                             cube_symbols: cache.cube_symbols.clone(),
                             context_symbols: cache.context_symbols.clone(),
                         };
-                        config_lock.insert(data.compiler_id.clone(), cache);
+                        config_lock.put(data.compiler_id.clone(), cache);
                         cfg
                     }
                     None => {
-                        let cache = METADATA_CACHE.read().unwrap();
+                        let mut config_lock = METADATA_CACHE.lock().unwrap();
 
-                        match cache.get(&data.compiler_id) {
+                        match config_lock.get(&data.compiler_id) {
                             Some(cached) => TransformConfig {
                                 file_name: data.file_name,
                                 transpilers: data.transpilers,

packages/cubejs-backend-native/src/transpilers.rs

@@ -10,7 +10,10 @@ homepage = "https://cube.dev"
 
 [dependencies]
 arc-swap = "1"
-datafusion = { git = 'https://github.com/cube-js/arrow-datafusion.git', rev = "dcf3e4aa26fd112043ef26fa4a78db5dbd443c86", default-features = false, features = ["regex_expressions", "unicode_expressions"] }
+datafusion = { git = 'https://github.com/cube-js/arrow-datafusion.git', rev = "dcf3e4aa26fd112043ef26fa4a78db5dbd443c86", default-features = false, features = [
+    "regex_expressions",
+    "unicode_expressions",
+] }
 thiserror = "2"
 cubeclient = { path = "../cubeclient" }
 pg-srv = { path = "../pg-srv" }
@@ -25,7 +28,7 @@ futures = "0.3.23"
 rand = "0.8.3"
 hashbrown = "0.14.3"
 log = "0.4.21"
-rust_decimal = { version = "1.25", features = ["c-repr", "db-postgres"]}
+rust_decimal = { version = "1.25", features = ["c-repr", "db-postgres"] }
 postgres-types = "0.2.3"
 # Locked, because starting from 1.15 this crate switch from chrono to time
 # which panic with Could not determine the UTC offset on this system.
@@ -38,10 +41,12 @@ uuid = { version = "1", features = ["serde", "v4"] }
 bincode = "1.3.1"
 chrono = "0.4.31"
 chrono-tz = "0.6"
-tokio-util = { version = "0.7", features=["compat"] }
+tokio-util = { version = "0.7", features = ["compat"] }
 comfy-table = "7.1.0"
 bitflags = "1.3.2"
-egg = { rev = "952f8c2a1033e5da097d23c523b0d8e392eb532b", git = "https://github.com/cube-js/egg.git", features = ["serde-1"] }
+egg = { rev = "952f8c2a1033e5da097d23c523b0d8e392eb532b", git = "https://github.com/cube-js/egg.git", features = [
+    "serde-1",
+] }
 paste = "1.0.6"
 tracing = { version = "0.1.40", features = ["async-await"] }
 async-stream = "0.3.3"
@@ -50,7 +55,7 @@ futures-util = "0.3.23"
 sha1_smol = "1.0.0"
 tera = { version = "1", default-features = false }
 minijinja = { version = "1", features = ["json", "loader"] }
-lru = "0.12.1"
+lru = "0.13.0"
 sha2 = "0.10.8"
 bigdecimal = "0.4.2"
 indexmap = "1.9.3"
@@ -60,7 +65,10 @@ indexmap = "1.9.3"
 pretty_assertions = "1.0.0"
 insta = "1.12"
 portpicker = "0.1.1"
-tokio-postgres = { version = "0.7.7", features = ["with-chrono-0_4", "runtime"] }
+tokio-postgres = { version = "0.7.7", features = [
+    "with-chrono-0_4",
+    "runtime",
+] }
 rust_decimal = { version = "1.23", features = ["db-tokio-postgres"] }
 pg_interval = "0.4.1"
 criterion = { version = "0.4.0", features = ["html_reports"] }

rust/cubenativeutils/Cargo.lock

@@ -5306,18 +5306,6 @@
     uuid "9.0.0"
     winston "3.8.2"
 
-"@cubejs-backend/[email protected]":
-  version "1.2.4"
-  resolved "https://registry.yarnpkg.com/@cubejs-backend/linter/-/linter-1.2.4.tgz#b38551ada107d94a3a12bb3f82e582ef6e184096"
-  integrity sha512-V150xWxBgRptCO4GAdZTuCwY7+FoxlJ1rBYM379LLEX8sLFC4NG0AN8S3vy0YSaUaC3P07ne3CrEILIJLt0JAg==
-  dependencies:
-    "@typescript-eslint/eslint-plugin" "^6.12.0"
-    "@typescript-eslint/parser" "^6.12.0"
-    eslint "^8.54.0"
-    eslint-config-airbnb-base "^14.2.1"
-    eslint-plugin-import "^2.22.1"
-    eslint-plugin-node "^9.2.0"
-
 "@cubejs-backend/[email protected]":
   version "0.33.20"
   resolved "https://registry.yarnpkg.com/@cubejs-backend/shared/-/shared-0.33.20.tgz#3d9fa60041599cca9fe4c04df05daa4b8ab8675f"
@@ -5338,26 +5326,6 @@
     throttle-debounce "^3.0.1"
     uuid "^8.3.2"
 
-"@cubejs-backend/[email protected]":
-  version "1.2.4"
-  resolved "https://registry.yarnpkg.com/@cubejs-backend/shared/-/shared-1.2.4.tgz#dcdc8a0537cfe9df97f3502bea983204634fbdfd"
-  integrity sha512-C/BA0ogCRhhtYwYyQkhXTTKLIUtJ1020aybbQ9tbKtlkrd3ab+a166jmtlAD5kDeoIhB08DgMTQjaQGDYh3gBg==
-  dependencies:
-    "@oclif/color" "^0.1.2"
-    bytes "^3.1.0"
-    cli-progress "^3.9.0"
-    cross-spawn "^7.0.3"
-    decompress "^4.2.1"
-    env-var "^6.3.0"
-    fs-extra "^9.1.0"
-    http-proxy-agent "^4.0.1"
-    moment-range "^4.0.1"
-    moment-timezone "^0.5.46"
-    node-fetch "^2.6.1"
-    shelljs "^0.8.5"
-    throttle-debounce "^3.0.1"
-    uuid "^8.3.2"
-
 "@cubejs-infra/post-installer@^0.0.7":
   version "0.0.7"
   resolved "https://registry.yarnpkg.com/@cubejs-infra/post-installer/-/post-installer-0.0.7.tgz#a28d2d03e5b7b69a64020d75194a7078cf911d2d"