Merge pull request #330 from korpling/add-facet

Add facet support to some types

Author: thomaskrause

CHANGELOG.md

@@ -14,6 +14,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   change also increases performance for several queries in the test set by up to
   30%.
 
+### Added
+
+- Added the support for [Facet](https://facet.rs/) to some of the core types
+  (like `Annotation`, `Component`, etc.) to make it possible to use Facet in the
+  [Annatto](https://github.com/korpling/annatto/)
+
 ## [3.8.3] - 2025-08-12
 
 ### Fixed

core/Cargo.toml

@@ -14,6 +14,7 @@ version = "3.8.3"
 binary-layout = "4.0.1"
 bincode = "1.2"
 clru = "0.6.1"
+facet = "0.28.0"
 itertools = "0.10"
 lazy_static = "1.4"
 log = "0.4"
@@ -32,8 +33,6 @@ serde_bytes = "0.11"
 serde_derive = "1.0"
 smallvec = "1.6"
 sstable = "0.11"
-strum = "0.21"
-strum_macros = "0.21"
 tempfile = "3.1"
 thiserror = "1"
 toml = "0.8"

core/src/types.rs

@@ -1,3 +1,4 @@
+use facet::Facet;
 use num_traits::{Bounded, FromPrimitive, Num, ToPrimitive};
 use serde::Serialize;
 use serde::de::DeserializeOwned;
@@ -6,8 +7,6 @@ use std::fmt;
 use std::ops::AddAssign;
 
 use std::{convert::TryInto, str::FromStr};
-use strum::IntoEnumIterator;
-use strum_macros::{EnumIter, EnumString};
 
 use super::serializer::{FixedSizeKeySerializer, KeySerializer};
 use crate::serializer::KeyVec;
@@ -22,7 +21,9 @@ use std::result::Result as StdResult;
 pub type NodeID = u64;
 
 /// The fully qualified name of an annotation.
-#[derive(Serialize, Deserialize, Default, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash)]
+#[derive(
+    Facet, Serialize, Deserialize, Default, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash,
+)]
 pub struct AnnoKey {
     /// Name of the annotation.
     pub name: String,
@@ -31,7 +32,9 @@ pub struct AnnoKey {
 }
 
 /// An annotation with a qualified name and a value.
-#[derive(Serialize, Deserialize, Default, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash)]
+#[derive(
+    Facet, Serialize, Deserialize, Default, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash,
+)]
 pub struct Annotation {
     /// Qualified name or unique "key" for the annotation
     pub key: AnnoKey,
@@ -40,7 +43,9 @@ pub struct Annotation {
 }
 
 /// Directed edge between a source and target node which are identified by their ID.
-#[derive(Serialize, Deserialize, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash, Default)]
+#[derive(
+    Facet, Serialize, Deserialize, Eq, PartialEq, PartialOrd, Ord, Clone, Debug, Hash, Default,
+)]
 #[repr(C)]
 pub struct Edge {
     pub source: NodeID,
@@ -135,7 +140,8 @@ pub trait ComponentType:
 }
 
 /// A simplified implementation of a `ComponentType` that only has one type of edges.
-#[derive(Clone, Eq, PartialEq, PartialOrd, Ord, EnumString, EnumIter, Debug)]
+#[derive(Facet, Clone, Eq, PartialEq, PartialOrd, Ord, Debug)]
+#[repr(u16)]
 pub enum DefaultComponentType {
     Edge,
 }
@@ -160,7 +166,7 @@ impl fmt::Display for DefaultComponentType {
 
 pub struct DefaultGraphIndex;
 
-#[derive(Serialize, Deserialize)]
+#[derive(Facet, Serialize, Deserialize)]
 pub struct DefaultGlobalStatistics;
 
 impl ComponentType for DefaultComponentType {
@@ -173,15 +179,27 @@ impl ComponentType for DefaultComponentType {
         Ok(DefaultGraphIndex {})
     }
     fn all_component_types() -> Vec<Self> {
-        DefaultComponentType::iter().collect()
+        vec![DefaultComponentType::Edge]
     }
     fn calculate_global_statistics(_graph: &mut Graph<Self>) -> StdResult<(), ComponentTypeError> {
         Ok(())
     }
 }
 
+impl FromStr for DefaultComponentType {
+    type Err = GraphAnnisCoreError;
+
+    fn from_str(s: &str) -> StdResult<Self, Self::Err> {
+        if s == "Edge" {
+            Ok(Self::Edge)
+        } else {
+            Err(GraphAnnisCoreError::InvalidComponentType(s.to_string()))
+        }
+    }
+}
+
 /// Identifies an edge component of the graph.
-#[derive(Serialize, Deserialize, Eq, PartialEq, PartialOrd, Ord, Hash, Clone, Debug)]
+#[derive(Facet, Serialize, Deserialize, Eq, PartialEq, PartialOrd, Ord, Hash, Clone, Debug)]
 pub struct Component<CT: ComponentType> {
     /// Type of the component
     ctype: u16,
@@ -257,3 +275,20 @@ impl NumValue for u64 {}
 impl NumValue for u32 {}
 impl NumValue for u16 {}
 impl NumValue for u8 {}
+
+#[cfg(test)]
+mod tests {
+    use std::str::FromStr;
+
+    use crate::{errors::GraphAnnisCoreError, types::DefaultComponentType};
+
+    #[test]
+    fn parse_invalid_component_type() {
+        let result = DefaultComponentType::from_str("doesnotexist");
+        assert!(result.is_err());
+        assert!(matches!(
+            result.unwrap_err(),
+            GraphAnnisCoreError::InvalidComponentType(_)
+        ));
+    }
+}

graphannis/Cargo.toml

@@ -24,6 +24,7 @@ regex = "1"
 [dependencies]
 boolean_expression = "0.4"
 csv = "1"
+facet = "0.28.0"
 fs2 = "0.4"
 graphannis-core = { path = "../core/", version = "^3" }
 itertools = "0.10"

graphannis/src/annis/db/aql/model.rs

@@ -3,6 +3,7 @@ use crate::{
     errors::GraphAnnisError,
     graph::{Edge, EdgeContainer, GraphStorage, NodeID},
 };
+use facet::Facet;
 use graphannis_core::{
     annostorage::ValueSearch,
     dfs::CycleSafeDFS,
@@ -44,7 +45,18 @@ lazy_static! {
 
 /// Specifies the type of component of the annotation graph. The types of this enum carray certain semantics about the edges of the graph components their are used in.
 #[derive(
-    Serialize, Deserialize, Eq, PartialEq, PartialOrd, Ord, Hash, Clone, Debug, EnumIter, EnumString,
+    Facet,
+    Serialize,
+    Deserialize,
+    Eq,
+    PartialEq,
+    PartialOrd,
+    Ord,
+    Hash,
+    Clone,
+    Debug,
+    EnumIter,
+    EnumString,
 )]
 #[repr(C)]
 pub enum AnnotationComponentType {

graphannis/src/annis/errors.rs

@@ -1,6 +1,7 @@
 use std::{fmt::Display, sync::PoisonError};
 
 use crate::annis::types::LineColumnRange;
+
 use graphannis_core::{
     errors::{ComponentTypeError, GraphAnnisCoreError},
     types::NodeID,

verify.sh

@@ -18,7 +18,7 @@ cargo llvm-cov report --ignore-filename-regex '(tests?\.rs)|(capi/.*)' --release
 
 # Use diff-cover (https://github.com/Bachmann1234/diff_cover) and output code coverage compared to main branch
 mkdir -p target/llvm-cov/html/
-OUTPUT="$(diff-cover target/llvm-cov/tests.lcov --html-report target/llvm-cov/html/patch.html)"
+OUTPUT="$(diff-cover target/llvm-cov/tests.lcov --format html:target/llvm-cov/html/patch.html)"
 echo "$OUTPUT"
 if [ -z "${CI}" ]; then
     echo "HTML report available at $PWD/target/llvm-cov/html/patch.html"