Make EncodingID Copy

bench-vortex/src/lib.rs

@@ -9,18 +9,10 @@ use log::{info, warn};
 use parquet::arrow::arrow_reader::ParquetRecordBatchReaderBuilder;
 use parquet::arrow::ProjectionMask;
 
-use vortex::array::bool::BoolEncoding;
-use vortex::array::chunked::{ChunkedArray, ChunkedEncoding};
-use vortex::array::composite::CompositeEncoding;
-use vortex::array::constant::ConstantEncoding;
+use vortex::array::chunked::ChunkedArray;
 use vortex::array::downcast::DowncastArrayBuiltin;
-use vortex::array::primitive::PrimitiveEncoding;
-use vortex::array::sparse::SparseEncoding;
-use vortex::array::struct_::StructEncoding;
-use vortex::array::varbin::VarBinEncoding;
-use vortex::array::varbinview::VarBinViewEncoding;
-use vortex::array::IntoArray;
-use vortex::array::{Array, ArrayRef, Encoding};
+use vortex::array::{Array, ArrayRef};
+use vortex::array::{EncodingId, IntoArray};
 use vortex::arrow::FromArrowType;
 use vortex::compress::{CompressConfig, CompressCtx};
 use vortex::formatter::display_tree;
@@ -32,40 +24,25 @@ use vortex_ree::REEEncoding;
 use vortex_roaring::RoaringBoolEncoding;
 use vortex_schema::DType;
 
-pub fn enumerate_arrays() -> Vec<&'static dyn Encoding> {
+pub fn enumerate_arrays() -> Vec<EncodingId> {
     vec![
-        // TODO(ngates): fix https://github.com/fulcrum-so/vortex/issues/35
-        // Builtins
-        &BoolEncoding,
-        &ChunkedEncoding,
-        &CompositeEncoding,
-        &ConstantEncoding,
-        &PrimitiveEncoding,
-        &SparseEncoding,
-        &StructEncoding,
-        &VarBinEncoding,
-        &VarBinViewEncoding,
-        // Encodings
-        &ALPEncoding,
-        &DictEncoding,
-        &BitPackedEncoding,
-        &FoREncoding,
-        &DateTimeEncoding,
-        // &DeltaEncoding,
-        // &FFoREncoding,
-        &REEEncoding,
-        &RoaringBoolEncoding,
-        // &RoaringIntEncoding,
+        ALPEncoding::ID,
+        DictEncoding::ID,
+        BitPackedEncoding::ID,
+        FoREncoding::ID,
+        DateTimeEncoding::ID,
+        // DeltaEncoding::ID,
+        // FFoREncoding::ID,
+        REEEncoding::ID,
+        RoaringBoolEncoding::ID,
+        // RoaringIntEncoding::ID,
         // Doesn't offer anything more than FoR really
-        // &ZigZagEncoding,
+        // ZigZagEncoding::ID,
     ]
 }
 
 pub fn compress_ctx() -> CompressCtx {
-    let cfg = CompressConfig::new(
-        HashSet::from_iter(enumerate_arrays().iter().map(|e| (*e).id())),
-        HashSet::default(),
-    );
+    let cfg = CompressConfig::new(HashSet::from_iter(enumerate_arrays()), HashSet::default());
     info!("Compression config {cfg:?}");
     CompressCtx::new(Arc::new(cfg))
 }

pyvortex/src/array.rs

@@ -111,7 +111,7 @@ impl PyArray {
                 inner.into_any().downcast::<VarBinViewArray>().unwrap(),
             )?
             .extract(py),
-            ArrayKind::Other(other) => match *other.encoding().id() {
+            ArrayKind::Other(other) => match other.encoding().id() {
                 // PyEnc chooses to expose certain encodings as first-class objects.
                 // For the remainder, we should have a generic EncArray implementation that supports basic functions.
                 ALPEncoding::ID => {

vortex-alp/src/array.rs

@@ -51,7 +51,7 @@ impl ALPArray {
     pub fn encode(array: &dyn Array) -> VortexResult<ArrayRef> {
         match ArrayKind::from(array) {
             ArrayKind::Primitive(p) => Ok(alp_encode(p)?.into_array()),
-            _ => Err(VortexError::InvalidEncoding(array.encoding().id().clone())),
+            _ => Err(VortexError::InvalidEncoding(array.encoding().id())),
         }
     }
 
@@ -131,8 +131,8 @@ impl ALPEncoding {
 }
 
 impl Encoding for ALPEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn compression(&self) -> Option<&dyn EncodingCompression> {

vortex-array/src/array/bool/mod.rs

@@ -139,8 +139,8 @@ impl BoolEncoding {
 static ENCODINGS_BOOL: EncodingRef = &BoolEncoding;
 
 impl Encoding for BoolEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn serde(&self) -> Option<&dyn EncodingSerde> {

vortex-array/src/array/chunked/mod.rs

@@ -183,8 +183,8 @@ impl ChunkedEncoding {
 static ENCODINGS_CHUNKED: EncodingRef = &ChunkedEncoding;
 
 impl Encoding for ChunkedEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn serde(&self) -> Option<&dyn EncodingSerde> {

vortex-array/src/array/composite/array.rs

@@ -140,8 +140,8 @@ impl CompositeEncoding {
 static ENCODINGS_COMPOSITE: EncodingRef = &CompositeEncoding;
 
 impl Encoding for CompositeEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn compression(&self) -> Option<&dyn EncodingCompression> {

vortex-array/src/array/composite/compress.rs

@@ -14,7 +14,7 @@ impl EncodingCompression for CompositeEncoding {
         array: &dyn Array,
         _config: &CompressConfig,
     ) -> Option<&dyn EncodingCompression> {
-        (array.encoding().id() == &Self::ID).then_some(self)
+        (array.encoding().id() == Self::ID).then_some(self)
     }
 
     fn compress(

vortex-array/src/array/constant/mod.rs

@@ -109,8 +109,8 @@ impl ConstantEncoding {
 static ENCODINGS_CONSTANT: EncodingRef = &ConstantEncoding;
 
 impl Encoding for ConstantEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn serde(&self) -> Option<&dyn EncodingSerde> {

vortex-array/src/array/mod.rs

@@ -247,7 +247,7 @@ pub fn check_validity_buffer(validity: Option<&ArrayRef>, expected_len: usize) -
     Ok(())
 }
 
-#[derive(Clone, Debug, Eq, PartialEq, Hash)]
+#[derive(Clone, Copy, Debug, Eq, PartialEq, Hash)]
 pub struct EncodingId(&'static str);
 
 impl EncodingId {
@@ -268,7 +268,7 @@ impl Display for EncodingId {
 }
 
 pub trait Encoding: Debug + Send + Sync + 'static {
-    fn id(&self) -> &EncodingId;
+    fn id(&self) -> EncodingId;
 
     /// Whether this encoding provides a compressor.
     fn compression(&self) -> Option<&dyn EncodingCompression> {
@@ -308,7 +308,7 @@ pub enum ArrayKind<'a> {
 
 impl<'a> From<&'a dyn Array> for ArrayKind<'a> {
     fn from(value: &'a dyn Array) -> Self {
-        match *value.encoding().id() {
+        match value.encoding().id() {
             BoolEncoding::ID => ArrayKind::Bool(value.as_bool()),
             ChunkedEncoding::ID => ArrayKind::Chunked(value.as_chunked()),
             CompositeEncoding::ID => ArrayKind::Composite(value.as_composite()),

vortex-array/src/array/primitive/compute/patch.rs

@@ -11,7 +11,7 @@ use crate::{compute, match_each_native_ptype};
 impl PatchFn for PrimitiveArray {
     fn patch(&self, patch: &dyn Array) -> VortexResult<ArrayRef> {
         match patch.encoding().id() {
-            &SparseEncoding::ID => patch_with_sparse(self, patch.as_sparse()),
+            SparseEncoding::ID => patch_with_sparse(self, patch.as_sparse()),
             // TODO(ngates): support a default implementation based on iter_arrow?
             _ => Err(VortexError::MissingKernel(
                 "patch",

vortex-array/src/array/primitive/mod.rs

@@ -229,8 +229,8 @@ impl PrimitiveEncoding {
 static ENCODINGS_PRIMITIVE: EncodingRef = &PrimitiveEncoding;
 
 impl Encoding for PrimitiveEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn serde(&self) -> Option<&dyn EncodingSerde> {

vortex-array/src/array/sparse/compress.rs

@@ -14,7 +14,7 @@ impl EncodingCompression for SparseEncoding {
         array: &dyn Array,
         _config: &CompressConfig,
     ) -> Option<&dyn EncodingCompression> {
-        (array.encoding().id() == &Self::ID).then_some(self)
+        (array.encoding().id() == Self::ID).then_some(self)
     }
 
     fn compress(

vortex-array/src/array/sparse/mod.rs

@@ -163,8 +163,8 @@ impl SparseEncoding {
 static ENCODINGS_SPARSE: EncodingRef = &SparseEncoding;
 
 impl Encoding for SparseEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn compression(&self) -> Option<&dyn EncodingCompression> {

vortex-array/src/array/struct_/compress.rs

@@ -11,7 +11,7 @@ impl EncodingCompression for StructEncoding {
         array: &dyn Array,
         _config: &CompressConfig,
     ) -> Option<&dyn EncodingCompression> {
-        (array.encoding().id() == &Self::ID).then_some(self)
+        (array.encoding().id() == Self::ID).then_some(self)
     }
 
     fn compress(

vortex-array/src/array/struct_/mod.rs

@@ -125,8 +125,8 @@ impl StructEncoding {
 static ENCODINGS_STRUCT: EncodingRef = &StructEncoding;
 
 impl Encoding for StructEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn compression(&self) -> Option<&dyn EncodingCompression> {

vortex-array/src/array/varbin/compress.rs

@@ -14,7 +14,7 @@ impl EncodingCompression for VarBinEncoding {
         array: &dyn Array,
         _config: &CompressConfig,
     ) -> Option<&dyn EncodingCompression> {
-        (array.encoding().id() == &Self::ID).then_some(self)
+        (array.encoding().id() == Self::ID).then_some(self)
     }
 
     fn compress(

vortex-array/src/array/varbin/mod.rs

@@ -276,8 +276,8 @@ impl VarBinEncoding {
 static ENCODINGS_VARBIN: EncodingRef = &VarBinEncoding;
 
 impl Encoding for VarBinEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn compression(&self) -> Option<&dyn EncodingCompression> {

vortex-array/src/array/varbinview/mod.rs

@@ -266,8 +266,8 @@ impl VarBinViewEncoding {
 static ENCODINGS_VARBINVIEW: EncodingRef = &VarBinViewEncoding;
 
 impl Encoding for VarBinViewEncoding {
-    fn id(&self) -> &EncodingId {
-        &Self::ID
+    fn id(&self) -> EncodingId {
+        Self::ID
     }
 
     fn serde(&self) -> Option<&dyn EncodingSerde> {

vortex-array/src/compress.rs

@@ -4,9 +4,12 @@ use std::sync::Arc;
 
 use log::{debug, info, warn};
 
-use crate::array::chunked::ChunkedArray;
+use crate::array::chunked::{ChunkedArray, ChunkedEncoding};
+use crate::array::composite::CompositeEncoding;
 use crate::array::constant::{ConstantArray, ConstantEncoding};
-use crate::array::struct_::StructArray;
+use crate::array::sparse::SparseEncoding;
+use crate::array::struct_::{StructArray, StructEncoding};
+use crate::array::varbin::VarBinEncoding;
 use crate::array::{Array, ArrayKind, ArrayRef, Encoding, EncodingId, ENCODINGS};
 use crate::compute;
 use crate::compute::scalar_at::scalar_at;
@@ -47,8 +50,8 @@ pub struct CompressConfig {
     max_depth: u8,
     // TODO(ngates): can each encoding define their own configs?
     pub ree_average_run_threshold: f32,
-    encodings: HashSet<&'static EncodingId>,
-    disabled_encodings: HashSet<&'static EncodingId>,
+    encodings: HashSet<EncodingId>,
+    disabled_encodings: HashSet<EncodingId>,
 }
 
 impl Default for CompressConfig {
@@ -68,12 +71,23 @@ impl Default for CompressConfig {
 }
 
 impl CompressConfig {
+    const DEFAULT_ENCODINGS: [EncodingId; 5] = [
+        ChunkedEncoding::ID,
+        CompositeEncoding::ID,
+        SparseEncoding::ID,
+        StructEncoding::ID,
+        VarBinEncoding::ID,
+    ];
+
     pub fn new(
-        encodings: HashSet<&'static EncodingId>,
-        mut disabled_encodings: HashSet<&'static EncodingId>,
+        mut encodings: HashSet<EncodingId>,
+        mut disabled_encodings: HashSet<EncodingId>,
     ) -> Self {
+        Self::DEFAULT_ENCODINGS.iter().for_each(|e| {
+            encodings.insert(*e);
+        });
         // Always disable constant encoding, it's handled separately
-        disabled_encodings.insert(&ConstantEncoding::ID);
+        disabled_encodings.insert(ConstantEncoding::ID);
         Self {
             encodings,
             disabled_encodings,
@@ -91,9 +105,9 @@ impl CompressConfig {
         )
     }
 
-    pub fn is_enabled(&self, kind: &EncodingId) -> bool {
-        (self.encodings.is_empty() || self.encodings.contains(kind))
-            && !self.disabled_encodings.contains(kind)
+    pub fn is_enabled(&self, kind: EncodingId) -> bool {
+        (self.encodings.is_empty() || self.encodings.contains(&kind))
+            && !self.disabled_encodings.contains(&kind)
     }
 }
 
@@ -236,7 +250,7 @@ pub fn sampled_compression(array: &dyn Array, ctx: &CompressCtx) -> VortexResult
     let mut candidates: Vec<&dyn EncodingCompression> = ENCODINGS
         .iter()
         .filter(|encoding| ctx.options().is_enabled(encoding.id()))
-        .filter(|encoding| !ctx.disabled_encodings.contains(encoding.id()))
+        .filter(|encoding| !ctx.disabled_encodings.contains(&encoding.id()))
         .filter_map(|encoding| encoding.compression())
         .filter(|compression| {
             if compression

vortex-array/src/error.rs

@@ -17,10 +17,10 @@ pub enum VortexError {
     InvalidArgument(ErrString),
     // Used when a function is not implemented for a given array type.
     #[error("function {0} not implemented for {1}")]
-    NotImplemented(&'static str, &'static EncodingId),
+    NotImplemented(&'static str, EncodingId),
     // Used when a function is implemented for an array type, but the RHS is not supported.
     #[error("missing kernel {0} for {1} and {2:?}")]
-    MissingKernel(&'static str, &'static EncodingId, Vec<&'static EncodingId>),
+    MissingKernel(&'static str, EncodingId, Vec<EncodingId>),
     #[error("invalid data type: {0}")]
     InvalidDType(DType),
     #[error("invalid physical type: {0:?}")]

vortex-array/src/serde/mod.rs

@@ -35,7 +35,7 @@ where
 
 pub struct ReadCtx<'a> {
     schema: &'a DType,
-    encodings: Vec<&'static EncodingId>,
+    encodings: Vec<EncodingId>,
     r: &'a mut dyn Read,
 }
 
@@ -167,7 +167,7 @@ impl<'a> ReadCtx<'a> {
 
 pub struct WriteCtx<'a> {
     w: &'a mut dyn Write,
-    available_encodings: Vec<&'static EncodingId>,
+    available_encodings: Vec<EncodingId>,
 }
 
 impl<'a> WriteCtx<'a> {

vortex-datetime/src/compress.rs

@@ -17,7 +17,7 @@ impl EncodingCompression for DateTimeEncoding {
         array: &dyn Array,
         _config: &CompressConfig,
     ) -> Option<&dyn EncodingCompression> {
-        if array.encoding().id() != &CompositeEncoding::ID {
+        if array.encoding().id() != CompositeEncoding::ID {
             return None;
         }