Move MeasureUnit to separate module and Implement UnitsTrieV1 Dat…

…a Provider (unicode-org#5348)

components/experimental/src/lib.rs

@@ -85,6 +85,7 @@ pub mod provider {
         impl_short_week_relative_time_format_data_v1_marker!(Baked);
         impl_short_year_relative_time_format_data_v1_marker!(Baked);
         impl_units_info_v1_marker!(Baked);
+        impl_units_trie_v1_marker!(Baked);
     };
 
     #[cfg(feature = "datagen")]
@@ -111,6 +112,7 @@ pub mod provider {
         super::displaynames::provider::RegionDisplayNamesV1Marker::INFO,
         super::displaynames::provider::ScriptDisplayNamesV1Marker::INFO,
         super::displaynames::provider::VariantDisplayNamesV1Marker::INFO,
+        super::measure::provider::trie::UnitsTrieV1Marker::INFO,
         super::personnames::provider::PersonNamesFormatV1Marker::INFO,
         super::relativetime::provider::LongDayRelativeTimeFormatDataV1Marker::INFO,
         super::relativetime::provider::LongHourRelativeTimeFormatDataV1Marker::INFO,

components/experimental/src/measure/measureunit.rs

@@ -0,0 +1,19 @@
+// This file is part of ICU4X. For terms of use, please see the file
+// called LICENSE at the top level of the ICU4X source tree
+// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
+
+use smallvec::SmallVec;
+
+use super::provider::single_unit::SingleUnit;
+
+// TODO NOTE: the MeasureUnitParser takes the trie and the ConverterFactory takes the full payload and an instance of MeasureUnitParser.
+/// MeasureUnit is a struct that contains a processed CLDR unit.
+///     For example, "meter-per-second".
+/// NOTE:
+///   - To construct a MeasureUnit from a cldr unit identifier, use the `MeasureUnitParser`.
+#[derive(Debug)]
+pub struct MeasureUnit {
+    // TODO: make this field private and add functions to use it.
+    /// Contains the processed units.
+    pub contained_units: SmallVec<[SingleUnit; 8]>,
+}

components/experimental/src/measure/mod.rs

@@ -4,3 +4,7 @@
 
 pub(crate) mod power;
 pub(crate) mod si_prefix;
+
+pub mod measureunit;
+pub mod parser;
+pub mod provider;

components/experimental/src/units/measureunit.rs → components/experimental/src/measure/parser.rs

@@ -3,16 +3,15 @@
 // (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
 
 use alloc::vec::Vec;
-use smallvec::SmallVec;
 use zerotrie::ZeroTrieSimpleAscii;
 
+use crate::measure::measureunit::MeasureUnit;
 use crate::measure::power::get_power;
 use crate::measure::si_prefix::get_si_prefix;
+use crate::units::InvalidUnitError;
 
-use crate::units::{
-    provider::{Base, MeasureUnitItem, SiPrefix},
-    InvalidUnitError,
-};
+use super::provider::si_prefix::{Base, SiPrefix};
+use super::provider::single_unit::SingleUnit;
 
 // TODO: add test cases for this parser after adding UnitsTest.txt to the test data.
 /// A parser for the CLDR unit identifier (e.g. `meter-per-square-second`)
@@ -97,7 +96,7 @@ impl<'data> MeasureUnitParser<'data> {
             return Err(InvalidUnitError);
         }
 
-        let mut measure_unit_items = Vec::<MeasureUnitItem>::new();
+        let mut measure_unit_items = Vec::<SingleUnit>::new();
         let mut sign = 1;
         while !code_units.is_empty() {
             // First: extract the power.
@@ -138,7 +137,7 @@ impl<'data> MeasureUnitParser<'data> {
                     }
                 };
 
-            measure_unit_items.push(MeasureUnitItem {
+            measure_unit_items.push(SingleUnit {
                 power: sign * power as i8,
                 si_prefix,
                 unit_id,
@@ -158,10 +157,3 @@ impl<'data> MeasureUnitParser<'data> {
         })
     }
 }
-
-// TODO NOTE: the MeasureUnitParser takes the trie and the ConverterFactory takes the full payload and an instance of MeasureUnitParser.
-#[derive(Debug)]
-pub struct MeasureUnit {
-    /// Contains the processed units.
-    pub contained_units: SmallVec<[MeasureUnitItem; 8]>,
-}

components/experimental/src/measure/provider/mod.rs

@@ -0,0 +1,7 @@
+// This file is part of ICU4X. For terms of use, please see the file
+// called LICENSE at the top level of the ICU4X source tree
+// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
+
+pub mod si_prefix;
+pub mod single_unit;
+pub mod trie;

components/experimental/src/measure/provider/si_prefix.rs

@@ -0,0 +1,47 @@
+// This file is part of ICU4X. For terms of use, please see the file
+// called LICENSE at the top level of the ICU4X source tree
+// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
+
+// Provider structs must be stable
+#![allow(clippy::exhaustive_structs, clippy::exhaustive_enums)]
+
+//! Data provider struct definitions for this ICU4X component.
+//!
+//! Read more about data providers: [`icu_provider`]
+
+/// Represents the base of an si prefix.
+#[zerovec::make_ule(BaseULE)]
+#[cfg_attr(
+    feature = "datagen",
+    derive(serde::Serialize, databake::Bake),
+    databake(path = icu_experimental::units::provider),
+)]
+#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
+#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
+#[repr(u8)]
+pub enum Base {
+    /// The base of the si prefix is 10.
+    #[default]
+    Decimal = 0,
+
+    /// The base of the si prefix is 2.
+    Binary = 1,
+}
+
+// TODO: Consider reducing the size of this struct while implementing the ULE.
+/// Represents the SI prefix.
+#[zerovec::make_ule(SiPrefixULE)]
+#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
+#[cfg_attr(
+    feature = "datagen",
+    derive(serde::Serialize, databake::Bake),
+    databake(path = icu_experimental::units::provider),
+)]
+#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
+pub struct SiPrefix {
+    /// The absolute value of the power of the si prefix.
+    pub power: i8,
+
+    /// The base of the si prefix.
+    pub base: Base,
+}

components/experimental/src/measure/provider/single_unit.rs

@@ -0,0 +1,28 @@
+// This file is part of ICU4X. For terms of use, please see the file
+// called LICENSE at the top level of the ICU4X source tree
+// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
+
+use super::si_prefix::SiPrefix;
+
+/// Represents a single unit in a measure unit.
+/// For example, the MeasureUnit `kilometer-per-square-second` contains two single units:
+///    1. `kilometer` with power 1 and prefix 3 with base 10.
+///    2. `second` with power -2 and prefix power equal to 0.
+#[zerovec::make_ule(SingleUnitULE)]
+#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
+#[cfg_attr(
+    feature = "datagen",
+    derive(serde::Serialize, databake::Bake),
+    databake(path = icu_experimental::units::provider),
+)]
+#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
+pub struct SingleUnit {
+    /// The power of the unit.
+    pub power: i8,
+
+    /// The si base of the unit.
+    pub si_prefix: SiPrefix,
+
+    /// The id of the unit.
+    pub unit_id: u16,
+}

components/experimental/src/measure/provider/trie.rs

@@ -0,0 +1,46 @@
+// This file is part of ICU4X. For terms of use, please see the file
+// called LICENSE at the top level of the ICU4X source tree
+// (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
+
+// Provider structs must be stable
+#![allow(clippy::exhaustive_structs, clippy::exhaustive_enums)]
+
+//! Data provider struct definitions for this ICU4X component.
+//!
+//! Read more about data providers: [`icu_provider`]
+
+use icu_provider::prelude::*;
+use zerotrie::ZeroTrieSimpleAscii;
+use zerovec::ZeroVec;
+
+#[cfg(feature = "compiled_data")]
+/// Baked data
+///
+/// <div class="stab unstable">
+/// 🚧 This code is considered unstable; it may change at any time, in breaking or non-breaking ways,
+/// including in SemVer minor releases. In particular, the `DataProvider` implementations are only
+/// guaranteed to match with this version's `*_unstable` providers. Use with caution.
+/// </div>
+pub use crate::provider::Baked;
+
+/// This type encapsulates all the constant data required for unit conversions.
+///
+/// <div class="stab unstable">
+/// 🚧 This code is considered unstable; it may change at any time, in breaking or non-breaking ways,
+/// including in SemVer minor releases. While the serde representation of data structs is guaranteed
+/// to be stable, their Rust representation might not be. Use with caution.
+/// </div>
+#[icu_provider::data_struct(marker(UnitsTrieV1Marker, "units/trie@1", singleton))]
+#[derive(Clone, PartialEq, Debug)]
+#[cfg_attr(
+    feature = "datagen",
+    derive(serde::Serialize, databake::Bake),
+    databake(path = icu_experimental::measure::provider::trie),
+)]
+#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
+pub struct UnitsTrieV1<'data> {
+    // TODO: remove this field from units/provider::UnitsInfoV1 once the `MeasureUnit` is fully used in the measurement units.
+    /// Maps from unit name (e.g. foot or meter) to its unit id. this id can be used to retrieve the conversion information from the `UnitsInfoV1`.
+    #[cfg_attr(feature = "serde", serde(borrow))]
+    pub trie: ZeroTrieSimpleAscii<ZeroVec<'data, u8>>,
+}

components/experimental/src/measure/si_prefix.rs

@@ -4,7 +4,7 @@
 
 use zerotrie::ZeroTrieSimpleAscii;
 
-use crate::units::provider::{Base, SiPrefix};
+use super::provider::si_prefix::{Base, SiPrefix};
 
 /// The offset of the SI prefixes.
 /// NOTE:

components/experimental/src/units/converter_factory.rs

@@ -2,15 +2,16 @@
 // called LICENSE at the top level of the ICU4X source tree
 // (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
 
+use crate::measure::measureunit::MeasureUnit;
+use crate::measure::parser::MeasureUnitParser;
+use crate::measure::provider::single_unit::SingleUnit;
 use crate::units::provider;
-use crate::units::provider::MeasureUnitItem;
 use crate::units::ratio::IcuRatio;
 use crate::units::{
     converter::{
         OffsetConverter, ProportionalConverter, ReciprocalConverter, UnitsConverter,
         UnitsConverterInner,
     },
-    measureunit::{MeasureUnit, MeasureUnitParser},
     provider::Sign,
 };
 
@@ -175,7 +176,7 @@ impl ConverterFactory {
         ///     For example, `newton` has the basic units:  `gram`, `meter`, and `second` (each one has it is own power and si prefix).
         fn insert_non_basic_units(
             factory: &ConverterFactory,
-            units: &[MeasureUnitItem],
+            units: &[SingleUnit],
             sign: i16,
             map: &mut LiteMap<u16, PowersInfo>,
         ) -> Option<()> {
@@ -200,7 +201,7 @@ impl ConverterFactory {
         ///     For example, `square-foot` , the base unit is `meter` with power 1.
         ///     Thus, the inserted power should be `1 * 2 = 2`.
         fn insert_base_units(
-            basic_units: &ZeroSlice<MeasureUnitItem>,
+            basic_units: &ZeroSlice<SingleUnit>,
             original_power: i16,
             sign: i16,
             map: &mut LiteMap<u16, PowersInfo>,
@@ -248,7 +249,7 @@ impl ConverterFactory {
         }
     }
 
-    fn compute_conversion_term(&self, unit_item: &MeasureUnitItem, sign: i8) -> Option<IcuRatio> {
+    fn compute_conversion_term(&self, unit_item: &SingleUnit, sign: i8) -> Option<IcuRatio> {
         let conversion_info = self
             .payload
             .get()

components/experimental/src/units/mod.rs

@@ -7,7 +7,6 @@ use displaydoc::Display;
 pub mod converter;
 pub mod converter_factory;
 pub mod convertible;
-pub mod measureunit;
 pub mod provider;
 pub mod ratio;
 

components/experimental/src/units/provider.rs

@@ -14,6 +14,7 @@ use num_bigint::BigInt;
 use zerotrie::ZeroTrieSimpleAscii;
 use zerovec::{ule::AsULE, VarZeroVec, ZeroVec};
 
+use crate::measure::provider::single_unit::SingleUnit;
 #[cfg(feature = "compiled_data")]
 /// Baked data
 ///
@@ -42,6 +43,7 @@ use super::ratio::IcuRatio;
 )]
 #[cfg_attr(feature = "serde", derive(serde::Deserialize))]
 pub struct UnitsInfoV1<'data> {
+    // TODO: remove this field once we are using this map from `measure/provider::UnitsTrie`.
     /// Maps from unit name (e.g. foot) to it is conversion information.
     #[cfg_attr(feature = "serde", serde(borrow))]
     pub units_conversion_trie: ZeroTrieSimpleAscii<ZeroVec<'data, u8>>,
@@ -75,7 +77,7 @@ pub struct UnitsInfoV1<'data> {
 pub struct ConversionInfo<'data> {
     /// Contains the base unit (after parsing) which what the unit is converted to.
     #[cfg_attr(feature = "serde", serde(borrow))]
-    pub basic_units: ZeroVec<'data, MeasureUnitItem>,
+    pub basic_units: ZeroVec<'data, SingleUnit>,
 
     /// Represents the numerator of the conversion factor.
     #[cfg_attr(feature = "serde", serde(borrow))]
@@ -137,65 +139,6 @@ pub enum Exactness {
     Approximate = 1,
 }
 
-#[zerovec::make_ule(BaseULE)]
-#[cfg_attr(
-    feature = "datagen",
-    derive(serde::Serialize, databake::Bake),
-    databake(path = icu_experimental::units::provider),
-)]
-#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
-#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
-#[repr(u8)]
-pub enum Base {
-    /// The base of the si prefix is 10.
-    #[default]
-    Decimal = 0,
-
-    /// The base of the si prefix is 2.
-    Binary = 1,
-}
-
-/// Represents an Item of a MeasureUnit.
-/// For example, the MeasureUnit `kilometer-per-square-second` contains two items:
-///    1. `kilometer` with power 1 and prefix 3 with base 10.
-///    2. `second` with power -2 and prefix `NotExist`.
-#[zerovec::make_ule(MeasureUnitItemULE)]
-#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
-#[cfg_attr(
-    feature = "datagen",
-    derive(serde::Serialize, databake::Bake),
-    databake(path = icu_experimental::units::provider),
-)]
-#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
-pub struct MeasureUnitItem {
-    /// The power of the unit.
-    pub power: i8,
-
-    /// The si base of the unit.
-    pub si_prefix: SiPrefix,
-
-    /// The id of the unit.
-    pub unit_id: u16,
-}
-
-// TODO: Consider reducing the size of this struct while implementing the ULE.
-/// Represents the SI prefix.
-#[zerovec::make_ule(SiPrefixULE)]
-#[derive(Copy, Clone, Debug, PartialOrd, Ord, PartialEq, Eq, Default)]
-#[cfg_attr(
-    feature = "datagen",
-    derive(serde::Serialize, databake::Bake),
-    databake(path = icu_experimental::units::provider),
-)]
-#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
-pub struct SiPrefix {
-    /// The absolute value of the power of the si prefix.
-    pub power: i8,
-
-    /// The base of the si prefix.
-    pub base: Base,
-}
-
 impl ConversionInfoULE {
     /// Extracts the conversion factor as [`super::ratio::IcuRatio`].
     pub(crate) fn factor_as_ratio(&self) -> IcuRatio {

components/experimental/src/units/ratio.rs

@@ -13,7 +13,7 @@ use num_rational::Ratio;
 use num_traits::Signed;
 use num_traits::{One, Pow, Zero};
 
-use super::provider::{Base, SiPrefix};
+use crate::measure::provider::si_prefix::{Base, SiPrefix};
 
 // TODO: add test cases for IcuRatio.
 // TODO: Make a decicion on whether to keep the `IcuRatio` public or not.