chore: remove prefilter, stand-alone-test and default feature (#505)

* chore: remove prefilter, stand-alone-test and default feature Signed-off-by: usamoi <usamoi@outlook.com> * fix: simplify hnsw entry Signed-off-by: usamoi <usamoi@outlook.com> --------- Signed-off-by: usamoi <usamoi@outlook.com>
tensorchord · Jun 20, 2024 · 4360cac · 4360cac
1 parent 1a9e0b6
commit 4360cac
Show file tree

Hide file tree

Showing 29 changed files with 127 additions and 1,208 deletions.
diff --git a/.github/workflows/psql.yml b/.github/workflows/psql.yml
@@ -95,8 +95,8 @@ jobs:
           echo PGRX_TARGET_INFO_PATH_PG$VERSION=$HOME/.pgrx_binding >> "$GITHUB_ENV"
       - name: Build Release
         run: |
-          cargo build --lib --no-default-features --features "pg$VERSION" --release
-          ./tools/schema.sh --no-default-features --features "pg$VERSION" --release | expand -t 4 > ./target/vectors--$SEMVER.sql
+          cargo build --lib --features "pg$VERSION" --release
+          ./tools/schema.sh --features "pg$VERSION" --release | expand -t 4 > ./target/vectors--$SEMVER.sql
       - name: Set up PostgreSQL
         run: |
           sudo sh -c 'echo "deb http://apt.postgresql.org/pub/repos/apt $(lsb_release -cs)-pgdg main" >> /etc/apt/sources.list.d/pgdg.list'

diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml
@@ -92,8 +92,8 @@ jobs:
           echo PGRX_TARGET_INFO_PATH_PG$VERSION=$HOME/.pgrx_binding >> "$GITHUB_ENV"
       - name: Build
         run: |
-          cargo build --lib --no-default-features --features pg$VERSION --release --target $ARCH-unknown-linux-gnu
-          ./tools/schema.sh --no-default-features --features pg$VERSION --release --target $ARCH-unknown-linux-gnu | expand -t 4 > ./target/vectors--$SEMVER.sql
+          cargo build --lib --features pg$VERSION --release --target $ARCH-unknown-linux-gnu
+          ./tools/schema.sh --features pg$VERSION --release --target $ARCH-unknown-linux-gnu | expand -t 4 > ./target/vectors--$SEMVER.sql
       - name: Package
         run: |
           export PLATFORM=$(echo $ARCH | sed 's/aarch64/arm64/; s/x86_64/amd64/')

diff --git a/.github/workflows/rust.yml b/.github/workflows/rust.yml
@@ -98,9 +98,9 @@ jobs:
           cp ./vendor/pgrx_binding/pg${VERSION}_$(uname --machine)-unknown-linux-gnu.rs ~/.pgrx_binding/pg${VERSION}_raw_bindings.rs
           echo PGRX_TARGET_INFO_PATH_PG$VERSION=$HOME/.pgrx_binding >> "$GITHUB_ENV"
       - name: Clippy
-        run: cargo clippy --no-default-features --features "pg$VERSION" --target $ARCH-unknown-linux-gnu
+        run: cargo clippy --features "pg$VERSION" --target $ARCH-unknown-linux-gnu
       - name: Build
-        run: cargo build --lib --no-default-features --features "pg$VERSION" --target $ARCH-unknown-linux-gnu
+        run: cargo build --lib --features "pg$VERSION" --target $ARCH-unknown-linux-gnu
       - name: Post Set up Cache
         uses: actions/cache/save@v4
         if: ${{ !steps.cache.outputs.cache-hit }}
@@ -163,14 +163,14 @@ jobs:
           cp ./vendor/pgrx_binding/pg${VERSION}_$(uname --machine)-unknown-linux-gnu.rs ~/.pgrx_binding/pg${VERSION}_raw_bindings.rs
           echo PGRX_TARGET_INFO_PATH_PG$VERSION=$HOME/.pgrx_binding >> "$GITHUB_ENV"
       - name: Test
-        run: cargo test --all --no-fail-fast --no-default-features --features "pg$VERSION" --target $ARCH-unknown-linux-gnu -- --nocapture
+        run: cargo test --all --no-fail-fast --features "pg$VERSION" --target $ARCH-unknown-linux-gnu -- --nocapture
       - name: Test (x86_64)
         if: matrix.arch == 'x86_64'
         run: |
             ASSETS=$(mktemp -d)
             wget https://downloadmirror.intel.com/813591/sde-external-9.33.0-2024-01-07-lin.tar.xz -O $ASSETS/sde-external.tar.xz
             tar -xf $ASSETS/sde-external.tar.xz -C $ASSETS
-            cargo --config "target.x86_64-unknown-linux-gnu.runner = [\"$ASSETS/sde-external-9.33.0-2024-01-07-lin/sde64\", \"-spr\", \"--\"]" test "_v4" --all --no-fail-fast --no-default-features --features "pg$VERSION" --target $ARCH-unknown-linux-gnu -- --nocapture
+            cargo --config "target.x86_64-unknown-linux-gnu.runner = [\"$ASSETS/sde-external-9.33.0-2024-01-07-lin/sde64\", \"-spr\", \"--\"]" test "_v4" --all --no-fail-fast --features "pg$VERSION" --target $ARCH-unknown-linux-gnu -- --nocapture
       - name: Post Set up Cache
         uses: actions/cache/save@v4
         if: ${{ !steps.cache.outputs.cache-hit }}

diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -11,7 +11,7 @@ name = "pgrx_embed_vectors"
 path = "./src/bin/pgrx_embed.rs"
 
 [features]
-default = ["pg15"]
+default = []
 pg14 = ["pgrx/pg14"]
 pg15 = ["pgrx/pg15"]
 pg16 = ["pgrx/pg16"]

diff --git a/crates/base/src/index.rs b/crates/base/src/index.rs
@@ -495,7 +495,6 @@ impl Default for ProductQuantizationOptionsRatio {
 
 #[derive(Debug, Clone, Serialize, Deserialize, Validate)]
 pub struct SearchOptions {
-    pub prefilter_enable: bool,
     #[validate(range(min = 1, max = 65535))]
     pub hnsw_ef_search: u32,
     #[validate(range(min = 1, max = 1_000_000))]

diff --git a/crates/base/src/search.rs b/crates/base/src/search.rs
@@ -74,10 +74,6 @@ pub struct Element {
     pub payload: Payload,
 }
 
-pub trait Filter: Clone {
-    fn check(&mut self, payload: Payload) -> bool;
-}
-
 pub trait Collection<O: Operator> {
     fn dims(&self) -> u32;
     fn len(&self) -> u32;

diff --git a/crates/base/src/worker.rs b/crates/base/src/worker.rs
@@ -28,20 +28,18 @@ pub trait WorkerOperations {
 }
 
 pub trait ViewBasicOperations {
-    fn basic<'a, F: Fn(Pointer) -> bool + Clone + 'a>(
+    fn basic<'a>(
         &'a self,
         vector: &'a OwnedVector,
         opts: &'a SearchOptions,
-        filter: F,
     ) -> Result<Box<dyn Iterator<Item = Pointer> + 'a>, BasicError>;
 }
 
 pub trait ViewVbaseOperations {
-    fn vbase<'a, F: FnMut(Pointer) -> bool + Clone + 'a>(
+    fn vbase<'a>(
         &'a self,
         vector: &'a OwnedVector,
         opts: &'a SearchOptions,
-        filter: F,
     ) -> Result<Box<dyn Iterator<Item = Pointer> + 'a>, VbaseError>;
 }
 

diff --git a/crates/flat/src/lib.rs b/crates/flat/src/lib.rs
@@ -39,18 +39,16 @@ impl<O: OperatorFlat> Flat<O> {
         &self,
         vector: Borrowed<'_, O>,
         _opts: &SearchOptions,
-        filter: impl Filter,
     ) -> BinaryHeap<Reverse<Element>> {
-        basic(&self.mmap, vector, filter)
+        basic(&self.mmap, vector)
     }
 
     pub fn vbase<'a>(
         &'a self,
         vector: Borrowed<'a, O>,
         _opts: &'a SearchOptions,
-        filter: impl Filter + 'a,
     ) -> (Vec<Element>, Box<(dyn Iterator<Item = Element> + 'a)>) {
-        vbase(&self.mmap, vector, filter)
+        vbase(&self.mmap, vector)
     }
 
     pub fn len(&self) -> u32 {
@@ -129,31 +127,25 @@ pub fn open<O: OperatorFlat>(path: &Path, options: IndexOptions) -> FlatMmap<O>
 pub fn basic<O: OperatorFlat>(
     mmap: &FlatMmap<O>,
     vector: Borrowed<'_, O>,
-    mut filter: impl Filter,
 ) -> BinaryHeap<Reverse<Element>> {
     let mut result = BinaryHeap::new();
     for i in 0..mmap.storage.len() {
         let distance = mmap.quantization.distance(vector, i);
         let payload = mmap.storage.payload(i);
-        if filter.check(payload) {
-            result.push(Reverse(Element { distance, payload }));
-        }
+        result.push(Reverse(Element { distance, payload }));
     }
     result
 }
 
 pub fn vbase<'a, O: OperatorFlat>(
     mmap: &'a FlatMmap<O>,
     vector: Borrowed<'a, O>,
-    mut filter: impl Filter + 'a,
 ) -> (Vec<Element>, Box<dyn Iterator<Item = Element> + 'a>) {
     let mut result = Vec::new();
     for i in 0..mmap.storage.len() {
         let distance = mmap.quantization.distance(vector, i);
         let payload = mmap.storage.payload(i);
-        if filter.check(payload) {
-            result.push(Element { distance, payload });
-        }
+        result.push(Element { distance, payload });
     }
     (result, Box::new(std::iter::empty()))
 }
diff --git a/crates/hnsw/Cargo.toml b/crates/hnsw/Cargo.toml
@@ -3,9 +3,6 @@ name = "hnsw"
 version.workspace = true
 edition.workspace = true
 
-[features]
-stand-alone-test = []
-
 [dependencies]
 bytemuck.workspace = true
 parking_lot.workspace = true