Merge pull request #34 from rust-scraper/serde

cfvescovo · web-flow · commit b3a9be133b25 · 2024-08-29T18:07:35.000+02:00
Implement serde traits for ego-tree
diff --git a/.gitignore b/.gitignore
@@ -1,3 +1,4 @@
 target
 Cargo.lock
-.vscode
+.vscode
+.idea
diff --git a/Cargo.toml b/Cargo.toml
@@ -11,3 +11,13 @@ authors = [
 license = "ISC"
 repository = "https://github.com/rust-scraper/ego-tree"
 readme = "README.md"
+
+[features]
+serde = ["dep:serde"]
+
+[dependencies]
+serde = { version = "1.0.209", optional = true }
+
+[dev-dependencies]
+serde = "1.0.209"
+serde_test = "1.0.177"
diff --git a/src/lib.rs b/src/lib.rs
@@ -38,6 +38,9 @@
 use std::fmt::{self, Debug, Display, Formatter};
 use std::num::NonZeroUsize;
 
+#[cfg(feature = "serde")]
+pub mod serde;
+
 /// Vec-backed ID-tree.
 ///
 /// Always contains at least a root node.
@@ -55,7 +58,7 @@ pub struct NodeId(NonZeroUsize);
 impl NodeId {
     // Safety: `n` must not equal `usize::MAX`.
     // (This is never the case for `Vec::len()`, that would mean it owns
-    // the entire address space without leaving space for even the its pointer.)
+    // the entire address space without leaving space even for its pointer.)
     unsafe fn from_index(n: usize) -> Self {
         NodeId(NonZeroUsize::new_unchecked(n + 1))
     }
@@ -75,7 +78,7 @@ struct Node<T> {
 }
 
 fn _static_assert_size_of_node() {
-    // "Instanciating" the generic `transmute` function without calling it
+    // "Instantiating" the generic `transmute` function without calling it
     // still triggers the magic compile-time check
     // that input and output types have the same `size_of()`.
     let _ = std::mem::transmute::<Node<()>, [usize; 5]>;
diff --git a/src/serde.rs b/src/serde.rs
@@ -0,0 +1,157 @@
+//! Implement `serde::Serialize` and `serde::Deserialize` traits for Tree
+//!
+//! # Warning
+//! Serialize and Deserialize implementations are recursive. They require an amount of stack memory
+//! proportional to the depth of the tree.
+
+use std::{fmt, marker::PhantomData};
+
+use serde::{
+    de::{self, MapAccess, Visitor},
+    ser::{Serialize, SerializeStruct},
+    Deserialize, Deserializer,
+};
+
+use crate::{NodeMut, NodeRef, Tree};
+
+#[derive(Debug)]
+struct SerNode<'a, T> {
+    value: &'a T,
+    children: Vec<SerNode<'a, T>>,
+}
+
+impl<'a, T> From<NodeRef<'a, T>> for SerNode<'a, T> {
+    fn from(node: NodeRef<'a, T>) -> Self {
+        let value = node.value();
+        let children = node.children().map(SerNode::from).collect();
+        Self { value, children }
+    }
+}
+
+impl<'a, T: Serialize> Serialize for SerNode<'a, T> {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: serde::Serializer,
+    {
+        let mut state = serializer.serialize_struct("Node", 2)?;
+        state.serialize_field("value", &self.value)?;
+        state.serialize_field("children", &self.children)?;
+        state.end()
+    }
+}
+
+impl<T: Serialize> Serialize for Tree<T> {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: serde::Serializer,
+    {
+        SerNode::from(self.root()).serialize(serializer)
+    }
+}
+
+#[derive(Debug)]
+struct DeserNode<T> {
+    value: T,
+    children: Vec<DeserNode<T>>,
+}
+
+impl<T> DeserNode<T> {
+    fn into_tree_node(self, parent: &mut NodeMut<T>) {
+        let mut node = parent.append(self.value);
+
+        for child in self.children {
+            child.into_tree_node(&mut node);
+        }
+    }
+}
+
+impl<T> From<DeserNode<T>> for Tree<T> {
+    fn from(root: DeserNode<T>) -> Self {
+        let mut tree: Tree<T> = Tree::new(root.value);
+        let mut tree_root = tree.root_mut();
+
+        for child in root.children {
+            child.into_tree_node(&mut tree_root);
+        }
+
+        tree
+    }
+}
+
+struct DeserNodeVisitor<T> {
+    marker: PhantomData<fn() -> DeserNode<T>>,
+}
+
+impl<T> DeserNodeVisitor<T> {
+    fn new() -> Self {
+        DeserNodeVisitor {
+            marker: PhantomData,
+        }
+    }
+}
+
+impl<'de, T> Visitor<'de> for DeserNodeVisitor<T>
+where
+    T: Deserialize<'de>,
+{
+    type Value = DeserNode<T>;
+
+    fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
+        formatter.write_str("struct Node")
+    }
+
+    fn visit_map<M>(self, mut map: M) -> Result<Self::Value, M::Error>
+    where
+        M: MapAccess<'de>,
+    {
+        let mut value = None;
+        let mut children = None;
+
+        while let Some(key) = map.next_key()? {
+            match key {
+                "value" => {
+                    if value.is_some() {
+                        return Err(de::Error::duplicate_field("value"));
+                    }
+                    value = Some(map.next_value()?);
+                }
+                "children" => {
+                    if children.is_some() {
+                        return Err(de::Error::duplicate_field("children"));
+                    }
+                    children = Some(map.next_value()?);
+                }
+                _ => {
+                    return Err(de::Error::unknown_field(key, &["value", "children"]));
+                }
+            }
+        }
+
+        let value = value.ok_or_else(|| de::Error::missing_field("value"))?;
+        let children = children.ok_or_else(|| de::Error::missing_field("children"))?;
+
+        Ok(DeserNode { value, children })
+    }
+}
+
+impl<'de, T> Deserialize<'de> for DeserNode<T>
+where
+    T: Deserialize<'de>,
+{
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        deserializer.deserialize_struct("Node", &["value", "children"], DeserNodeVisitor::new())
+    }
+}
+
+impl<'de, T: Deserialize<'de>> Deserialize<'de> for Tree<T> {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: serde::Deserializer<'de>,
+    {
+        let deser = DeserNode::<T>::deserialize(deserializer)?;
+        Ok(deser.into())
+    }
+}
diff --git a/tests/iter.rs b/tests/iter.rs
@@ -1,5 +1,4 @@
-#[macro_use]
-extern crate ego_tree;
+use ego_tree::tree;
 
 #[test]
 fn into_iter() {
diff --git a/tests/macro.rs b/tests/macro.rs
@@ -1,7 +1,4 @@
-#[macro_use]
-extern crate ego_tree;
-
-use ego_tree::Tree;
+use ego_tree::{tree, Tree};
 
 #[test]
 fn root() {
diff --git a/tests/node_mut.rs b/tests/node_mut.rs
@@ -1,7 +1,4 @@
-#[macro_use]
-extern crate ego_tree;
-
-use ego_tree::NodeRef;
+use ego_tree::{tree, NodeRef};
 
 #[test]
 fn value() {
diff --git a/tests/node_ref.rs b/tests/node_ref.rs
@@ -1,5 +1,4 @@
-#[macro_use]
-extern crate ego_tree;
+use ego_tree::tree;
 
 #[test]
 fn value() {
diff --git a/tests/serde.rs b/tests/serde.rs
@@ -0,0 +1,96 @@
+#![cfg(feature = "serde")]
+
+use ego_tree::tree;
+use serde_test::{assert_tokens, Token};
+
+#[test]
+fn test_internal_serde_repr_trivial() {
+    let tree = tree!("a");
+
+    assert_tokens(
+        &tree,
+        &[
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("a"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(0) },
+            Token::SeqEnd,
+            Token::StructEnd,
+        ],
+    );
+}
+
+#[test]
+fn test_internal_serde_repr() {
+    let tree = tree!("a" => {"b", "c" => {"d", "e"}, "f"});
+
+    assert_tokens(
+        &tree,
+        &[
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("a"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(3) },
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("b"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(0) },
+            Token::SeqEnd,
+            Token::StructEnd,
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("c"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(2) },
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("d"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(0) },
+            Token::SeqEnd,
+            Token::StructEnd,
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("e"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(0) },
+            Token::SeqEnd,
+            Token::StructEnd,
+            Token::SeqEnd,
+            Token::StructEnd,
+            Token::Struct {
+                name: "Node",
+                len: 2,
+            },
+            Token::BorrowedStr("value"),
+            Token::BorrowedStr("f"),
+            Token::BorrowedStr("children"),
+            Token::Seq { len: Some(0) },
+            Token::SeqEnd,
+            Token::StructEnd,
+            Token::SeqEnd,
+            Token::StructEnd,
+        ],
+    );
+}
diff --git a/tests/subtree.rs b/tests/subtree.rs
diff --git a/tests/tree.rs b/tests/tree.rs

-Original file line number
+Diff line change
 target
 Cargo.lock
 -.vscode
 +.vscode
 +.idea