use query result cache

Signed-off-by: Connor Tsui <connor.tsui20@gmail.com>

Author: connortsui20

Cargo.lock

@@ -26,11 +26,12 @@ path = "src/main.rs"
 anyhow = { workspace = true }
 axum = "0.8"
 base64 = "0.22"
+dashmap = { workspace = true }
 # track vortex-duckdb's bundled engine version (build.rs)
 duckdb = { version = "1.10502", features = ["bundled"] }
 maud = { version = "0.27", features = ["axum"] }
 parking_lot = { workspace = true }
-serde = { workspace = true, features = ["derive"] }
+serde = { workspace = true, features = ["derive", "rc"] }
 serde_json = { workspace = true }
 subtle = "2.6"
 thiserror = { workspace = true }

benchmarks-website/server/Cargo.toml

@@ -9,6 +9,7 @@
 //! returns a [`ChartResponse`].
 
 use std::collections::BTreeMap;
+use std::sync::Arc;
 
 use anyhow::Context as _;
 use anyhow::Result;
@@ -100,7 +101,7 @@ pub(crate) fn collect_group_charts(
         charts.push(NamedChartResponse {
             name: link.name,
             slug: link.slug,
-            chart,
+            chart: Arc::new(chart),
         });
     }
     if charts.is_empty() {

benchmarks-website/server/src/api/charts.rs

@@ -10,6 +10,7 @@
 //! ingest side, see [`crate::records`]).
 
 use std::collections::BTreeMap;
+use std::sync::Arc;
 
 use serde::Serialize;
 use serde_json::Value as JsonValue;
@@ -54,17 +55,21 @@ pub fn group_sort_key(name: &str) -> (usize, &str) {
 }
 
 /// Body of `GET /api/groups`: every group with its chart links and summary.
+///
+/// The inner [`Vec`] is held in an [`Arc`] so [`crate::query_cache::QueryCache`]
+/// can serve the same allocation to every concurrent reader without cloning.
+/// `Arc<T>` serialises through to `T`, so the wire shape is unchanged.
 #[derive(Debug, Serialize)]
 pub struct GroupsResponse {
     /// Every group surfaced by the discovery passes, in canonical order.
-    pub groups: Vec<Group>,
+    pub groups: Arc<Vec<Group>>,
 }
 
 /// One group: a display name, a slug for the group permalink, and the chart
 /// links inside it. Optionally carries a v2-compatible rollup summary and a
 /// short editorial description (rendered as a hover tooltip on the
 /// disclosure title).
-#[derive(Debug, Serialize)]
+#[derive(Debug, Clone, Serialize)]
 pub struct Group {
     /// Human-readable group label rendered in the disclosure header.
     pub name: String,
@@ -98,7 +103,7 @@ pub struct GroupChartsResponse {
 }
 
 /// Server-computed group summary, matching the v2 metadata contract.
-#[derive(Debug, Serialize)]
+#[derive(Debug, Clone, Serialize)]
 #[serde(tag = "type")]
 pub enum Summary {
     /// Random-access format ranking for the latest populated random-access chart.
@@ -161,7 +166,7 @@ pub enum Summary {
 }
 
 /// One random-access summary row.
-#[derive(Debug, Serialize)]
+#[derive(Debug, Clone, Serialize)]
 pub struct RandomAccessRanking {
     /// Series name, normally the physical format.
     pub name: String,
@@ -172,7 +177,7 @@ pub struct RandomAccessRanking {
 }
 
 /// One query benchmark summary row.
-#[derive(Debug, Serialize)]
+#[derive(Debug, Clone, Serialize)]
 pub struct QueryRanking {
     /// Series name, normally `engine:format`.
     pub name: String,
@@ -186,6 +191,10 @@ pub struct QueryRanking {
 /// A single chart inside a [`GroupChartsResponse`]. `name` is the chart's
 /// short label inside the group (e.g. `Q1`); `slug` round-trips through
 /// `/api/chart/{slug}`.
+///
+/// `chart` is held in an [`Arc`] so the cache and the landing-page builder
+/// share the same allocation; `Arc<T>` serialises as `T`, so the wire shape
+/// is identical to a plain `ChartResponse`.
 #[derive(Debug, Serialize)]
 pub struct NamedChartResponse {
     /// Chart label rendered in the chart-card title (e.g. `Q1`).
@@ -194,12 +203,12 @@ pub struct NamedChartResponse {
     pub slug: String,
     /// Inlined chart payload — same shape as `/api/chart/{slug}`.
     #[serde(flatten)]
-    pub chart: ChartResponse,
+    pub chart: Arc<ChartResponse>,
 }
 
 /// One chart's short label inside a group (e.g. `Q1`) plus the slug that
 /// resolves to its `/api/chart/{slug}` payload.
-#[derive(Debug, Serialize)]
+#[derive(Debug, Clone, Serialize)]
 pub struct ChartLink {
     /// Chart label rendered in the chart-card title (e.g. `Q1`).
     pub name: String,

benchmarks-website/server/src/api/dto.rs

@@ -67,7 +67,7 @@ use crate::slug::GroupKey;
 
 /// Handler for `GET /api/groups`.
 pub async fn groups(State(state): State<AppState>) -> Result<impl IntoResponse, ApiError> {
-    let groups = db::run_blocking(&state.db, |conn| collect_groups(conn)).await?;
+    let groups = cached_groups(&state).await?;
     Ok(Json(GroupsResponse { groups }))
 }
 
@@ -80,8 +80,7 @@ pub async fn chart(
     let key = ChartKey::from_slug(&slug)
         .map_err(|e| ApiError::BadRequest(format!("invalid slug: {e}")))?;
     let window = q.window();
-    let response =
-        db::run_blocking(&state.db, move |conn| chart_payload(conn, &key, &window)).await?;
+    let response = cached_chart_payload(&state, &slug, &key, &window).await?;
     let response =
         response.ok_or_else(|| ApiError::NotFound(format!("no data for slug {slug:?}")))?;
     Ok(Json(response))
@@ -96,15 +95,74 @@ pub async fn group(
     let key = GroupKey::from_slug(&slug)
         .map_err(|e| ApiError::BadRequest(format!("invalid group slug: {e}")))?;
     let window = q.window();
-    let response = db::run_blocking(&state.db, move |conn| {
-        collect_group_charts(conn, &key, &window)
-    })
-    .await?;
+    let response = cached_group_charts(&state, &slug, &key, &window).await?;
     let response =
         response.ok_or_else(|| ApiError::NotFound(format!("no data for group slug {slug:?}")))?;
     Ok(Json(response))
 }
 
+/// Cache-aware wrapper around `collect_groups`.
+pub async fn cached_groups(state: &AppState) -> Result<std::sync::Arc<Vec<Group>>> {
+    let db = state.db.clone();
+    state
+        .cache
+        .groups(move || async move { db::run_blocking(&db, |conn| collect_groups(conn)).await })
+        .await
+}
+
+/// Cache-aware wrapper around [`collect_filter_universe`].
+pub async fn cached_filter_universe(state: &AppState) -> Result<std::sync::Arc<FilterUniverse>> {
+    let db = state.db.clone();
+    state
+        .cache
+        .filter_universe(move || async move {
+            db::run_blocking(&db, |conn| collect_filter_universe(conn)).await
+        })
+        .await
+}
+
+/// Cache-aware wrapper around `chart_payload`.
+pub async fn cached_chart_payload(
+    state: &AppState,
+    slug: &str,
+    key: &ChartKey,
+    window: &CommitWindow,
+) -> Result<Option<std::sync::Arc<ChartResponse>>> {
+    let db = state.db.clone();
+    let key_for_compute = key.clone();
+    let window_for_compute = *window;
+    state
+        .cache
+        .chart_payload(slug, window, move || async move {
+            db::run_blocking(&db, move |conn| {
+                chart_payload(conn, &key_for_compute, &window_for_compute)
+            })
+            .await
+        })
+        .await
+}
+
+/// Cache-aware wrapper around `collect_group_charts`.
+pub async fn cached_group_charts(
+    state: &AppState,
+    slug: &str,
+    key: &GroupKey,
+    window: &CommitWindow,
+) -> Result<Option<std::sync::Arc<GroupChartsResponse>>> {
+    let db = state.db.clone();
+    let key_for_compute = key.clone();
+    let window_for_compute = *window;
+    state
+        .cache
+        .group_charts(slug, window, move || async move {
+            db::run_blocking(&db, move |conn| {
+                collect_group_charts(conn, &key_for_compute, &window_for_compute)
+            })
+            .await
+        })
+        .await
+}
+
 /// Handler for `GET /health`.
 pub async fn health(State(state): State<AppState>) -> Result<impl IntoResponse, ApiError> {
     let path = state.db_path.display().to_string();

benchmarks-website/server/src/api/mod.rs

@@ -17,7 +17,7 @@ use super::dto::DEFAULT_COMMIT_WINDOW;
 ///
 /// `Last(n)` keeps the most recent `n` commits by `commits.timestamp`; `All`
 /// returns every commit ever ingested.
-#[derive(Debug, Clone, Copy)]
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
 pub enum CommitWindow {
     /// Keep the most recent `n` commits.
     Last(NonZeroU32),

benchmarks-website/server/src/api/window.rs

@@ -30,6 +30,7 @@ use crate::db::DbHandle;
 use crate::db::{self};
 use crate::html;
 use crate::ingest;
+use crate::query_cache::QueryCache;
 
 /// Shared state for all handlers. Cheap to clone (everything is `Arc`-shaped
 /// or a small `String`).
@@ -41,6 +42,9 @@ pub struct AppState {
     pub bearer_token: Arc<String>,
     /// On-disk path of the DuckDB file. Surfaced on `/health`.
     pub db_path: Arc<PathBuf>,
+    /// In-memory cache of every read-side query result. Cleared by
+    /// [`crate::ingest`] after a successful commit. See [`crate::query_cache`].
+    pub cache: Arc<QueryCache>,
 }
 
 impl AppState {
@@ -52,6 +56,7 @@ impl AppState {
             db,
             bearer_token: Arc::new(bearer_token),
             db_path: Arc::new(path),
+            cache: Arc::new(QueryCache::new()),
         })
     }
 }