feat(onpair): promote OnPair to a stable read-only encoding

Cargo.toml

@@ -59,8 +59,7 @@ members = [
     "encodings/zstd",
     "encodings/bytebool",
     "encodings/parquet-variant",
-    # Experimental encodings
-    "encodings/experimental/onpair",
+    "encodings/onpair",
     # Benchmarks
     "benchmarks/lance-bench",
     "benchmarks/compress-bench",
@@ -307,7 +306,7 @@ vortex-json = { version = "0.1.0", path = "./vortex-json", default-features = fa
 vortex-layout = { version = "0.1.0", path = "./vortex-layout", default-features = false }
 vortex-mask = { version = "0.1.0", path = "./vortex-mask", default-features = false }
 vortex-metrics = { version = "0.1.0", path = "./vortex-metrics", default-features = false }
-vortex-onpair = { version = "0.1.0", path = "./encodings/experimental/onpair", default-features = false }
+vortex-onpair = { version = "0.1.0", path = "./encodings/onpair", default-features = false }
 vortex-parquet-variant = { version = "0.1.0", path = "./encodings/parquet-variant" }
 vortex-pco = { version = "0.1.0", path = "./encodings/pco", default-features = false }
 vortex-proto = { version = "0.1.0", path = "./vortex-proto", default-features = false }

vortex-btrblocks/src/schemes/string/scheme_selection_tests.rs

@@ -48,9 +48,9 @@ fn test_dict_compressed() -> VortexResult<()> {
 
 #[cfg(feature = "unstable_encodings")]
 #[test]
-fn test_onpair_in_default_scheme_list() {
+fn test_unstable_all_schemes_includes_onpair() {
     use crate::SchemeExt;
-    use crate::schemes::string::OnPairScheme;
+    use crate::schemes::string::onpair::OnPairScheme;
 
     let ids: Vec<_> = crate::ALL_SCHEMES.iter().map(|s| s.id()).collect();
     assert!(
@@ -61,10 +61,11 @@ fn test_onpair_in_default_scheme_list() {
 
 #[cfg(feature = "unstable_encodings")]
 #[test]
-fn test_onpair_compressed() -> VortexResult<()> {
+fn test_unstable_default_btrblocks_compressor_selects_onpair() -> VortexResult<()> {
     // Dictionary-style string corpus: high lexical overlap, short rows.
     // OnPair beats FSST on this corpus, so it wins the sample-based
-    // comparison even though both are registered by default.
+    // comparison even though both are registered when `unstable_encodings`
+    // is enabled.
     let mut strings = Vec::with_capacity(1000);
     for i in 0..1000 {
         strings.push(Some(format!(
@@ -83,8 +84,8 @@ fn test_onpair_compressed() -> VortexResult<()> {
     Ok(())
 }
 
-/// FSST is registered in the default scheme list (alongside OnPair), and an
-/// FSST-only builder still produces an FSST array.
+/// FSST is registered in the default scheme list, and an FSST-only builder
+/// still produces an FSST array.
 #[test]
 fn test_fsst_in_default_scheme_list() -> VortexResult<()> {
     use crate::BtrBlocksCompressorBuilder;

vortex-file/Cargo.toml

@@ -47,7 +47,7 @@ vortex-io = { workspace = true }
 vortex-layout = { workspace = true }
 vortex-mask = { workspace = true }
 vortex-metrics = { workspace = true }
-vortex-onpair = { workspace = true, optional = true }
+vortex-onpair = { workspace = true }
 vortex-pco = { workspace = true }
 vortex-runend = { workspace = true }
 vortex-scan = { workspace = true }
@@ -81,7 +81,6 @@ tokio = [
 zstd = ["dep:vortex-zstd", "vortex-btrblocks/zstd", "vortex-btrblocks/pco"]
 # This feature enables unstable encodings for which we don't guarantee stability.
 unstable_encodings = [
-    "dep:vortex-onpair",
     "dep:vortex-tensor",
     "vortex-zstd?/unstable_encodings",
     "vortex-btrblocks/unstable_encodings",

vortex-file/src/lib.rs

@@ -165,7 +165,6 @@ mod forever_constant {
 pub fn register_default_encodings(session: &VortexSession) {
     vortex_bytebool::initialize(session);
     vortex_fsst::initialize(session);
-    #[cfg(feature = "unstable_encodings")]
     vortex_onpair::initialize(session);
     vortex_zigzag::initialize(session);
 
@@ -201,6 +200,7 @@ mod default_encoding_tests {
     use vortex_array::optimizer::kernels::ArrayKernelsExt as _;
     use vortex_array::session::ArraySessionExt as _;
     use vortex_fsst::FSST;
+    use vortex_onpair::OnPair;
 
     use crate::register_default_encodings;
 
@@ -215,5 +215,17 @@ mod default_encoding_tests {
 
         assert!(session.arrays().registry().find(&FSST.id()).is_some());
         assert!(session.kernels().has_execute_parent(Filter.id(), FSST.id()));
+        assert!(session.arrays().registry().find(&OnPair.id()).is_some());
+        assert!(
+            session
+                .kernels()
+                .has_execute_parent(Filter.id(), OnPair.id())
+        );
+    }
+
+    #[cfg(not(feature = "unstable_encodings"))]
+    #[test]
+    fn default_writer_does_not_allow_onpair() {
+        assert!(!crate::ALLOWED_ENCODINGS.contains(&OnPair.id()));
     }
 }