Make RdfSource generic over the dataset type

Cargo.lock

@@ -244,9 +244,9 @@ dependencies = [
 
 [[package]]
 name = "either"
-version = "1.15.0"
+version = "1.16.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
+checksum = "91622ff5e7162018101f2fea40d6ebf4a78bbe5a49736a2020649edf9693679e"
 
 [[package]]
 name = "errno"
@@ -687,6 +687,7 @@ dependencies = [
  "oxigraph",
  "parse_link_header",
  "reqwest-middleware",
+ "slotmap",
  "thiserror",
  "tracing",
 ]
@@ -1308,6 +1309,15 @@ version = "0.4.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0c790de23124f9ab44544d7ac05d60440adc586479ce501c1d6d7da3cd8c9cf5"
 
+[[package]]
+name = "slotmap"
+version = "1.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bdd58c3c93c3d278ca835519292445cb4b0d4dc59ccfdf7ceadaab3f8aeb4038"
+dependencies = [
+ "version_check",
+]
+
 [[package]]
 name = "smallvec"
 version = "1.15.1"

Cargo.toml

@@ -17,6 +17,7 @@ color-eyre = "0.6"
 oxigraph = "0.5"
 parse_link_header = "0.4"
 reqwest-middleware = { version = "0.5", features = ["stream"] }
+slotmap = "1.1"
 thiserror = "2"
 tokio = { version = "1", features = ["full"] }
 tracing = "0.1"

ldctl/src/main.rs

@@ -1,5 +1,6 @@
 use ldp::ResourceRequestBuilder;
 use ldp::middleware::BasicAuthMiddleware;
+use ldp::oxigraph::model::Dataset;
 use ldp::reqwest::{Client, Url};
 use ldp::reqwest_middleware::ClientBuilder;
 use tracing::debug;
@@ -28,9 +29,9 @@ async fn main() -> color_eyre::Result<()> {
         .build();
 
     let resource = request.send().await?;
-    let rdf_source = resource.into_rdf_source().await?;
+    let rdf_source = resource.into_rdf_source::<Dataset>().await?;
     for quad in rdf_source.dataset() {
-        debug!(?quad);
+        println!("{:#?}", quad);
     }
 
     Ok(())

ldp/Cargo.toml

@@ -8,6 +8,10 @@ license = "GPL-3.0-only"
 keywords = ["ldp", "rdf", "sparql"]
 categories = ["database", "web-programming::http-client"]
 
+[features]
+default = ["keyed"]
+keyed = ["dep:slotmap"]
+
 [dependencies]
 async-trait.workspace = true
 base64.workspace = true
@@ -17,5 +21,6 @@ http.workspace = true
 oxigraph.workspace = true
 parse_link_header.workspace = true
 reqwest-middleware.workspace = true
+slotmap = { optional = true, workspace = true }
 thiserror.workspace = true
 tracing.workspace = true

ldp/src/error.rs

@@ -22,4 +22,8 @@ pub enum Error {
     /// The response from the server was not in a format we understand.
     #[error("Response was not in a supported RDF format")]
     UnsupportedFormat,
+
+    /// The RDF data failed to parse.
+    #[error(transparent)]
+    InvalidRdfSyntax(#[from] oxigraph::io::RdfSyntaxError),
 }

ldp/src/keyed.rs

@@ -0,0 +1,59 @@
+use oxigraph::model::{Quad, QuadRef};
+use slotmap::{SecondaryMap, SlotMap, new_key_type};
+
+new_key_type! { pub struct QuadKey; }
+
+#[derive(Clone, Debug, Default)]
+pub struct KeyedDataset<T> {
+    pub quads: SlotMap<QuadKey, Quad>,
+    pub associated_data: SecondaryMap<QuadKey, T>,
+}
+
+impl<T> FromIterator<Quad> for KeyedDataset<T> {
+    fn from_iter<U: IntoIterator<Item = Quad>>(iter: U) -> Self {
+        let mut quads = SlotMap::with_key();
+        for quad in iter {
+            quads.insert(quad);
+        }
+        Self {
+            quads,
+            associated_data: SecondaryMap::new(),
+        }
+    }
+}
+
+impl<'a, T> IntoIterator for &'a KeyedDataset<T> {
+    type Item = QuadRef<'a>;
+    type IntoIter = std::iter::Map<slotmap::basic::Values<'a, QuadKey, Quad>, fn(&Quad) -> QuadRef>;
+
+    fn into_iter(self) -> Self::IntoIter {
+        self.quads.values().map(|quad| quad.as_ref())
+    }
+}
+
+impl<T> KeyedDataset<T> {
+    /// Extend the dataset with the provided quads, returning a QuadKey for each insertion.
+    pub fn extend(&mut self, quads: impl Iterator<Item = Quad>) -> impl Iterator<Item = QuadKey> {
+        quads.map(|quad| self.quads.insert(quad))
+    }
+
+    /// Remove a quad and its associated value, if present.
+    pub fn remove(&mut self, key: QuadKey) {
+        self.quads.remove(key);
+        self.associated_data.remove(key);
+    }
+
+    /// Remove all quads and all associated data.
+    pub fn clear(&mut self) {
+        self.quads.clear();
+        self.associated_data.clear();
+    }
+
+    /// Iterate over both the quads and the associated data. Quads with no corresponding associated
+    /// data are filtered out.
+    pub fn iter_both(&self) -> impl Iterator<Item = (QuadKey, &Quad, &T)> {
+        self.quads
+            .iter()
+            .filter_map(|(key, quad)| self.associated_data.get(key).map(|ad| (key, quad, ad)))
+    }
+}

ldp/src/lib.rs

@@ -7,11 +7,16 @@ mod rdf_source;
 mod resource;
 pub mod vocab;
 
+#[cfg(feature = "keyed")]
+pub mod keyed;
+
 pub use http;
 pub use oxigraph;
 pub use reqwest_middleware;
 pub use reqwest_middleware::reqwest;
 
 pub use error::{Error, Result};
-pub use rdf_source::{RdfSource, RdfSourceUpdateRequest, RdfSourceUpdateResponse};
+pub use rdf_source::{
+    RdfSource, RdfSourceUpdateRequest, RdfSourceUpdateResponse, SerializationOptions,
+};
 pub use resource::{Resource, ResourceRequest, ResourceRequestBuilder, ResponseFormat};

ldp/src/rdf_source.rs

@@ -1,20 +1,36 @@
 use bytes::BufMut;
 use http::{StatusCode, header};
 use oxigraph::io::{RdfFormat, RdfSerializer};
-use oxigraph::model::Dataset;
+use oxigraph::model::{
+    GraphName, Literal, NamedNode, NamedOrBlankNode, Quad, QuadRef, Term, Triple, TripleRef, vocab,
+};
 use reqwest_middleware::ClientWithMiddleware;
 use reqwest_middleware::reqwest::Url;
 
 /// A LDP [RDF Source](https://www.w3.org/TR/ldp/#ldprs).
 #[derive(Clone, Debug)]
-pub struct RdfSource {
+pub struct RdfSource<D> {
     pub(crate) origin: Url,
     pub(crate) described_by: Option<Url>,
     pub(crate) state_token: Option<String>,
-    pub(crate) dataset: Dataset,
+    pub(crate) dataset: D,
 }
 
-impl RdfSource {
+impl<D: Default> RdfSource<D> {
+    /// Create a new, empty, RdfSource with the given URL.
+    ///
+    /// The subject of new quads will use this URL.
+    pub fn new(origin: Url) -> Self {
+        Self {
+            origin,
+            described_by: None,
+            state_token: None,
+            dataset: Default::default(),
+        }
+    }
+}
+
+impl<D> RdfSource<D> {
     /// The original URL used to procure this RDF Source.
     pub fn origin(&self) -> &Url {
         &self.origin
@@ -34,23 +50,104 @@ impl RdfSource {
         self.state_token.as_deref()
     }
 
-    /// The underlying Dataset.
+    /// The underlying dataset.
-    pub fn dataset(&self) -> &Dataset {
+    pub fn dataset(&self) -> &D {
         &self.dataset
     }
 
-    /// Serializes the Dataset in to the provided format.
+    /// A mutable reference to the underlying dataset.
-    pub fn serialize(&self, format: RdfFormat) -> crate::Result<bytes::Bytes> {
+    pub fn dataset_mut(&mut self) -> &mut D {
-        let writer = bytes::BytesMut::new().writer();
+        &mut self.dataset
-        let mut serializer = RdfSerializer::from_format(format).for_writer(writer);
+    }
 
-        if format.supports_datasets() {
+    /// Create a new quad, using the origin as the subject.
+    ///
+    /// The graph name is the `describedby` value, if present. If not present, the graph name is the
+    /// origin.
+    pub fn new_quad(&self) -> Quad {
+        let graph_name = self
+            .described_by
+            .as_ref()
+            .map(|db| db.as_str())
+            .unwrap_or(self.origin().as_str());
+        Quad::new(
+            NamedOrBlankNode::NamedNode(NamedNode::new_unchecked(self.origin.clone())),
+            vocab::rdf::VALUE,
+            Term::Literal(Literal::new_simple_literal("")),
+            GraphName::NamedNode(NamedNode::new_unchecked(graph_name)),
+        )
+    }
+
+    /// Create a new quad, using the given triple as a template.
+    ///
+    /// The graph name is the `describedby` value, if present. If not present, the graph name is the
+    /// origin.
+    pub fn quad_from_triple(&self, triple: Triple) -> Quad {
+        let graph_name = GraphName::NamedNode(NamedNode::new_unchecked(
+            self.described_by
+                .as_ref()
+                .map(|db| db.as_str())
+                .unwrap_or(self.origin().as_str()),
+        ));
+
+        Quad::new(
+            triple.subject,
+            triple.predicate,
+            triple.object,
+            graph_name.clone(),
+        )
+    }
+}
+
+/// Holds options related to serialization.
+pub struct SerializationOptions {
+    format: RdfFormat,
+    filter: Box<dyn Fn(TripleRef<'_>) -> bool>,
+}
+
+impl SerializationOptions {
+    /// Create a new set of serialization options with the provided format.
+    pub fn from_format(format: RdfFormat) -> Self {
+        Self {
+            format,
+            filter: Box::new(|_| true),
+        }
+    }
+
+    /// Filter triples that match the predicate.
+    ///
+    /// A return value of `true` means that it the triple ought to be included in the serialization.
+    pub fn with_filter<F>(self, filter: F) -> Self
+    where
+        F: Fn(TripleRef<'_>) -> bool + 'static,
+    {
+        Self {
+            format: self.format,
+            filter: Box::new(filter),
+        }
+    }
+}
+
+impl<'a, D: 'a> RdfSource<D>
+where
+    &'a D: IntoIterator<Item = QuadRef<'a>>,
+{
+    /// Serializes the dataset in to the provided format.
+    pub fn serialize(&'a self, options: SerializationOptions) -> crate::Result<bytes::Bytes> {
+        let writer = bytes::BytesMut::new().writer();
+        let mut serializer = RdfSerializer::from_format(options.format).for_writer(writer);
+
+        if options.format.supports_datasets() {
             for quad in &self.dataset {
-                serializer.serialize_quad(quad)?;
+                if (options.filter)(TripleRef::from(quad)) {
+                    serializer.serialize_quad(quad)?;
+                }
             }
         } else {
             for quad in &self.dataset {
-                serializer.serialize_triple(quad)?;
+                if (options.filter)(TripleRef::from(quad)) {
+                    serializer.serialize_triple(quad)?;
+                }
             }
         }
 
@@ -59,10 +156,13 @@ impl RdfSource {
     }
 
     /// Prepare an update request.
-    pub fn to_update(&self, format: RdfFormat) -> crate::Result<RdfSourceUpdateRequest> {
+    pub fn to_update(
+        &'a self,
+        options: SerializationOptions,
+    ) -> crate::Result<RdfSourceUpdateRequest> {
         let url = self.described_by.clone().unwrap_or(self.origin.clone());
-        let media_type = format.media_type().to_string();
+        let media_type = options.format.media_type().to_string();
-        let body = self.serialize(format)?;
+        let body = self.serialize(options)?;
 
         Ok(RdfSourceUpdateRequest {
             url,

ldp/src/resource.rs

@@ -3,7 +3,7 @@ use crate::vocab;
 use bytes::Bytes;
 use futures::Stream;
 use oxigraph::io::{RdfFormat, RdfParser};
-use oxigraph::model::{Dataset, GraphNameRef, NamedNodeRef};
+use oxigraph::model::{GraphNameRef, NamedNodeRef, Quad};
 use reqwest_middleware::reqwest::{Client, Response, StatusCode, Url, header};
 use reqwest_middleware::{ClientBuilder, ClientWithMiddleware, RequestBuilder};
 use tracing::error;
@@ -12,17 +12,22 @@ use tracing::error;
 ///
 /// # Example
 /// ```rust
+/// use ldp::ResourceRequestBuilder;
+/// use ldp::oxigraph::model::Dataset;
 /// use ldp::reqwest::{Client, Url};
 /// use ldp::reqwest_middleware::ClientBuilder;
-/// use ldp::ResourceRequestBuilder;
 ///
-/// let client = ClientBuilder::new(Client::new()).build();
 /// let url = Url::parse("http://server/resource")?;
-/// let resource = ResourceRequestBuilder::with_client_and_url(client.clone(), url)
+/// let request = ResourceRequestBuilder::new(url)
 ///     .follow_described_by(true)
 ///     .accept_all_rdf_formats()
-///     .send();
+///     .build();
-///     .await?;
+///
+/// let resource = request.send().await?;
+/// let rdf_source = resource.into_rdf_source::<Dataset>().await?;
+/// for quad in rdf_source.dataset() {
+///     println!("{:?}", quad);
+/// }
 /// ```
 #[derive(Clone, Debug)]
 pub struct ResourceRequestBuilder {
@@ -253,14 +258,17 @@ impl Resource {
     }
 
     /// Parse the response.
-    pub async fn into_rdf_source(self) -> crate::Result<RdfSource> {
+    pub async fn into_rdf_source<D>(self) -> crate::Result<RdfSource<D>>
+    where
+        D: FromIterator<Quad>,
+    {
         if let ResponseFormat::RdfFormat(format) = self.format {
             let graph_url = self.described_by.as_ref().unwrap_or(&self.origin);
             let graph = GraphNameRef::NamedNode(NamedNodeRef::new_unchecked(graph_url.as_str()));
             let parser = RdfParser::from_format(format).with_default_graph(graph);
             let body = self.response.bytes().await?;
             let quads = parser.for_slice(&body);
-            let dataset = quads.filter_map(Result::ok).collect::<Dataset>();
+            let dataset = quads.collect::<Result<_, _>>()?;
             Ok(RdfSource {
                 origin: self.origin,
                 described_by: self.described_by,