feat: Better serialization and perf. (#133)

- Improve serialization of the DataClass type. It's
now serialized as a simple string rather than a clumsy
map.

- USe serde_core instead of serde for improve build times.

- Change the Classified trait's data_class method
to return a &DataClass rather than a DataClass.
This avoids the need to clone the DataClass in many places,
improving performance (since DataClass has two Cow to clone).

- Use FxHashMap to hold the redactors to get much better
hash lookup perf. We don't need DoS protection since
the hash keys are entirely internal.

Author: geeknoid

Cargo.lock

@@ -25,9 +25,9 @@ repository = "https://github.com/microsoft/oxidizer"
 
 # local dependencies
 bytesbuf = { path = "crates/bytesbuf", default-features = false, version = "0.1.2" }
-data_privacy = { path = "crates/data_privacy", default-features = false, version = "0.9.0" }
-data_privacy_macros = { path = "crates/data_privacy_macros", default-features = false, version = "0.8.0" }
-data_privacy_macros_impl = { path = "crates/data_privacy_macros_impl", default-features = false, version = "0.8.0" }
+data_privacy = { path = "crates/data_privacy", default-features = false, version = "0.10.0" }
+data_privacy_macros = { path = "crates/data_privacy_macros", default-features = false, version = "0.9.0" }
+data_privacy_macros_impl = { path = "crates/data_privacy_macros_impl", default-features = false, version = "0.9.0" }
 fundle = { path = "crates/fundle", default-features = false, version = "0.3.0" }
 fundle_macros = { path = "crates/fundle_macros", default-features = false, version = "0.3.0" }
 fundle_macros_impl = { path = "crates/fundle_macros_impl", default-features = false, version = "0.3.0" }
@@ -68,8 +68,9 @@ quote = { version = "1.0.42", default-features = false }
 rand = { version = "0.9.2", default-features = false }
 rapidhash = { version = "4.1.1", default-features = false }
 regex = { version = "1.12.2", default-features = false }
+rustc-hash = { version = "2.1.0", default-features = false }
 serde = { version = "1.0.228", default-features = false }
-serde_core = { version = "1.0.224", default-features = false }
+serde_core = { version = "1.0.228", default-features = false }
 serde_json = { version = "1.0.145", default-features = false }
 smallvec = { version = "1.15.1", default-features = false }
 static_assertions = { version = "1.1.0", default-features = false }

Cargo.toml

@@ -1,10 +1,14 @@
 # Changelog
 
-## [0.9.0] - 2025-12-16
+## [0.10.0] - 2025-12-16
+
+- ✨ Features
+
+  - Better serialization and perf.
 
 - 🧩 Miscellaneous
 
-  - Make it to_redacted_string to avoid annoying downstream conflicts.
+  - Make it to_redacted_string to avoid annoying downstream conflicts. ([#143](https://github.com/microsoft/oxidizer/pull/143))
 
 ## [0.8.0] - 2025-12-10
 

crates/data_privacy/CHANGELOG.md

@@ -4,7 +4,7 @@
 [package]
 name = "data_privacy"
 description = "Data annotation and redaction system providing a robust way to manipulate sensitive information."
-version = "0.9.0"
+version = "0.10.0"
 readme = "README.md"
 keywords = ["oxidizer", "compliance", "privacy", "redaction", "scrubbing"]
 categories = ["data-structures"]
@@ -22,7 +22,6 @@ rustdoc-args = ["--cfg", "docsrs"]
 
 [package.metadata.cargo_check_external_types]
 allowed_external_types = [
-    "serde::*",
     "serde_core::de::*",
     "serde_core::ser::*",
     "data_privacy_macros::*",
@@ -31,21 +30,22 @@ allowed_external_types = [
 
 [features]
 default = ["serde"]
-serde = ["dep:serde"]
-xxh3 = ["dep:xxhash-rust"]
 rapidhash = ["dep:rapidhash"]
+serde = ["dep:serde_core"]
+xxh3 = ["dep:xxhash-rust"]
 
 [dependencies]
 data_privacy_macros.workspace = true
 rapidhash = { workspace = true, optional = true }
-serde = { workspace = true, optional = true, default-features = false, features = ["derive", "std"] }
+rustc-hash = { workspace = true, features = ["std"] }
+serde_core = { workspace = true, optional = true, default-features = false, features = ["std"] }
 xxhash-rust = { workspace = true, optional = true, features = ["xxh3"] }
 
 [dev-dependencies]
 derive_more = { workspace = true, features = ["constructor", "from"] }
 mutants.workspace = true
 once_cell = { workspace = true, features = ["std"] }
-serde = { workspace = true, features = ["derive", "std"] }
+serde = { workspace = true, features = ["std", "derive"] }
 serde_json = { workspace = true, features = ["std"] }
 
 [lints]

crates/data_privacy/Cargo.toml

@@ -51,8 +51,9 @@ use crate::DataClass;
 /// }
 ///
 /// impl Classified for ClassifiedPerson {
-///     fn data_class(&self) -> DataClass {
-///         DataClass::new("example_taxonomy", "classified_person")
+///     fn data_class(&self) -> &DataClass {
+///         static DATA_CLASS: DataClass = DataClass::new("example_taxonomy", "classified_person");
+///         &DATA_CLASS
 ///     }
 /// }
 ///
@@ -64,5 +65,5 @@ use crate::DataClass;
 pub trait Classified {
     /// Returns the data class of the classified data.
     #[must_use]
-    fn data_class(&self) -> DataClass;
+    fn data_class(&self) -> &DataClass;
 }