Lint

Cargo.toml

@@ -9,14 +9,15 @@ members = [
 ldp = { path = "ldp" }
 
 async-trait = "0.1"
-base64 = "0.22"
-bytes = "1.11"
+base64 = "0.23"
+bytes = "1.12"
 futures = "0.3"
-http = "1.4"
+http = "1.5"
 color-eyre = "0.6"
 oxigraph = "0.5"
 parse_link_header = "0.4"
 reqwest-middleware = { version = "0.5", features = ["stream"] }
+sfv = "0.15"
 slotmap = "1.1"
 thiserror = "2"
 tokio = { version = "1", features = ["full"] }

ldctl/Cargo.toml

@@ -4,9 +4,12 @@ version = "0.0.0"
 edition = "2024"
 publish = false
 
+[[bin]]
+name = "ldctl"
+doc = false
+
 [dependencies]
 ldp.workspace = true
-futures.workspace = true
 color-eyre.workspace = true
 tokio.workspace = true
 tracing.workspace = true

ldp/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "ldp"
-version = "0.0.0"
+version = "0.1.0"
 edition = "2024"
 description = "A library to assist with the creation and maintenance of remote RDF data via LDP"
 readme = "README.md"
@@ -8,7 +8,7 @@ license = "GPL-3.0-only"
 keywords = ["ldp", "rdf", "sparql"]
 categories = ["database", "web-programming::http-client"]
 homepage = "https://graphofliberty.org"
-repository = "https://code.graphofliberty.org/graphofliberty/ldp"
+repository = "https://code.graphofliberty.org/gl/ldp"
 
 [features]
 keyed = ["dep:slotmap"]
@@ -22,6 +22,7 @@ http.workspace = true
 oxigraph.workspace = true
 parse_link_header.workspace = true
 reqwest-middleware.workspace = true
+sfv.workspace = true
 slotmap = { optional = true, workspace = true }
 thiserror.workspace = true
 tracing.workspace = true

ldp/src/container.rs

@@ -0,0 +1,29 @@
+use crate::vocab;
+use oxigraph::model::NamedNodeRef;
+
+/// The type of [Container](https://www.w3.org/TR/ldp/#ldpc).
+#[derive(Clone, Debug)]
+pub enum ContainerType {
+    Basic,
+    Direct,
+    Indirect,
+}
+
+impl ContainerType {
+    pub fn as_named_node_ref(&self) -> NamedNodeRef<'_> {
+        match self {
+            ContainerType::Basic => vocab::ldp::BASIC_CONTAINER,
+            ContainerType::Direct => vocab::ldp::DIRECT_CONTAINER,
+            ContainerType::Indirect => vocab::ldp::INDIRECT_CONTAINER,
+        }
+    }
+
+    pub fn from_named_node_ref(node: NamedNodeRef<'_>) -> Option<Self> {
+        match node {
+            vocab::ldp::BASIC_CONTAINER => Some(ContainerType::Basic),
+            vocab::ldp::DIRECT_CONTAINER => Some(ContainerType::Direct),
+            vocab::ldp::INDIRECT_CONTAINER => Some(ContainerType::Indirect),
+            _ => None,
+        }
+    }
+}

ldp/src/error.rs

@@ -26,4 +26,7 @@ pub enum Error {
     /// The RDF data failed to parse.
     #[error(transparent)]
     InvalidRdfSyntax(#[from] oxigraph::io::RdfSyntaxError),
+
+    #[error(transparent)]
+    ToStr(#[from] http::header::ToStrError),
 }

ldp/src/header.rs

@@ -3,4 +3,3 @@ use reqwest_middleware::reqwest::header::HeaderName;
 pub const X_STATE_TOKEN: HeaderName = HeaderName::from_static("x-state-token");
 pub const X_IF_STATE_TOKEN: HeaderName = HeaderName::from_static("x-if-state-token");
 pub const PREFER: HeaderName = HeaderName::from_static("prefer");
-pub const PREFERENCE_APPLIED: HeaderName = HeaderName::from_static("preference-applied");

ldp/src/lib.rs

@@ -1,23 +1,27 @@
 #![cfg_attr(doc, doc = include_str!("../README.md"))]
 
+mod container;
 mod error;
-pub mod header;
+mod header;
 pub mod middleware;
+pub mod model;
+mod prefer;
 mod rdf_source;
 mod resource;
-pub mod traverse;
+mod traverse;
 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 container::ContainerType;
 pub use error::{Error, Result};
+pub use prefer::{Preference, header_for_preferences};
 pub use rdf_source::{
     RdfSource, RdfSourceUpdateRequest, RdfSourceUpdateResponse, SerializationOptions,
 };
-pub use resource::{Resource, ResourceRequest, ResourceRequestBuilder, ResponseFormat};
+pub use resource::{
+    Resource, ResourceRequest, ResourceRequestBuilder, ResourceType, ResponseFormat,
+};
+pub use traverse::Traverse;

ldp/src/model/dataset.rs

@@ -0,0 +1,17 @@
+use crate::RdfSource;
+use oxigraph::model::vocab::rdf;
+use oxigraph::model::{Dataset, NamedNodeRef, NamedOrBlankNodeRef, TermRef};
+
+impl RdfSource<Dataset> {
+    /// Return all the classes associated with this document.
+    pub fn classes(&self) -> impl Iterator<Item = NamedNodeRef<'_>> {
+        let subject =
+            NamedOrBlankNodeRef::NamedNode(NamedNodeRef::new_unchecked(self.origin.as_str()));
+        self.dataset
+            .quads_for_pattern(Some(subject), Some(rdf::TYPE), None, None)
+            .filter_map(|quad| match quad.object {
+                TermRef::NamedNode(node) => Some(node),
+                _ => None,
+            })
+    }
+}

ldp/src/model/keyed_dataset.rs

@@ -1,11 +1,21 @@
 use oxigraph::model::{Quad, QuadRef};
 use slotmap::{SecondaryMap, SlotMap, new_key_type};
 
-new_key_type! { pub struct QuadKey; }
+new_key_type! {
+    /// A type which uniquely identifies a quad.
+    pub struct QuadKey;
+}
 
+/// A set of quads stored in a [`SlotMap`].
+///
+/// This exists as a convenience for those who want to uniquely identify a quad and associate
+/// application-specific data with it.
 #[derive(Clone, Debug, Default)]
 pub struct KeyedDataset<T> {
+    /// The underlying SlotMap.
     pub quads: SlotMap<QuadKey, Quad>,
+
+    /// Application-specific information to be associated with the quads.
     pub associated_data: SecondaryMap<QuadKey, T>,
 }
 

ldp/src/model/mod.rs

@@ -0,0 +1,8 @@
+//! This module contains functionality related to the content of RDF documents.
+
+mod dataset;
+#[cfg(feature = "keyed")]
+mod keyed_dataset;
+
+#[cfg(feature = "keyed")]
+pub use keyed_dataset::{KeyedDataset, QuadKey};

ldp/src/prefer.rs

@@ -0,0 +1,59 @@
+use crate::vocab;
+use http::{HeaderMap, HeaderValue};
+use oxigraph::model::{NamedNode, NamedNodeRef};
+
+/// [Request preferences](https://www.w3.org/TR/ldp/#prefer-parameters)
+#[derive(Clone, Debug)]
+pub enum Preference {
+    Containment,
+    Membership,
+    MinimalContainer,
+    Other(NamedNode),
+}
+
+impl Preference {
+    pub fn as_named_node_ref(&self) -> NamedNodeRef<'_> {
+        match self {
+            Preference::Containment => vocab::ldp::PREFER_CONTAINMENT,
+            Preference::Membership => vocab::ldp::PREFER_MEMBERSHIP,
+            Preference::MinimalContainer => vocab::ldp::PREFER_MINIMAL_CONTAINER,
+            Preference::Other(preference) => preference.as_ref(),
+        }
+    }
+}
+
+pub fn header_for_preferences<'a, I, J>(include: I, omit: J) -> crate::Result<HeaderMap>
+where
+    I: IntoIterator<Item = &'a Preference>,
+    J: IntoIterator<Item = &'a Preference>,
+{
+    let include_str = include
+        .into_iter()
+        .map(|preference| preference.as_named_node_ref().as_str().to_string())
+        .collect::<Vec<_>>()
+        .join(" ");
+
+    let omit_str = omit
+        .into_iter()
+        .map(|preference| preference.as_named_node_ref().as_str().to_string())
+        .collect::<Vec<_>>()
+        .join(" ");
+
+    let mut value = String::from("return=representation");
+
+    if !include_str.is_empty() {
+        value = format!("{value}; include=\"{include_str}\"");
+    }
+
+    if !omit_str.is_empty() {
+        value = format!("{value}; omit=\"{omit_str}\"");
+    }
+
+    let header_value =
+        HeaderValue::from_str(&value).expect("Failed to produce Prefer header value");
+
+    Ok(HeaderMap::from_iter([(
+        crate::header::PREFER,
+        header_value,
+    )]))
+}

ldp/src/rdf_source.rs

@@ -1,3 +1,4 @@
+use crate::resource::ResourceType;
 use bytes::BufMut;
 use http::{StatusCode, header};
 use oxigraph::io::{RdfFormat, RdfSerializer};
@@ -11,8 +12,11 @@ use reqwest_middleware::reqwest::Url;
 #[derive(Clone, Debug)]
 pub struct RdfSource<D> {
     pub(crate) origin: Url,
-    pub(crate) described_by: Option<Url>,
+    pub(crate) origin_type: Option<ResourceType>,
     pub(crate) state_token: Option<String>,
+    pub(crate) described_by: Option<Url>,
+    pub(crate) file_name: Option<String>,
+    pub(crate) size: Option<usize>,
     pub(crate) dataset: D,
 }
 
@@ -23,8 +27,11 @@ impl<D: Default> RdfSource<D> {
     pub fn new(origin: Url) -> Self {
         Self {
             origin,
-            described_by: None,
+            origin_type: None,
             state_token: None,
+            described_by: None,
+            file_name: None,
+            size: None,
             dataset: Default::default(),
         }
     }
@@ -36,11 +43,9 @@ impl<D> RdfSource<D> {
         &self.origin
     }
 
-    /// The URL at which the RDF description of the resource is located.
-    ///
-    /// This is the URL at which updates are submitted.
-    pub fn described_by(&self) -> Option<&Url> {
-        self.described_by.as_ref()
+    /// The type of Resource that was originally fetched.
+    pub fn origin_type(&self) -> Option<&ResourceType> {
+        self.origin_type.as_ref()
     }
 
     /// The state token of the resource.
@@ -50,6 +55,23 @@ impl<D> RdfSource<D> {
         self.state_token.as_deref()
     }
 
+    /// The URL at which the RDF description of the resource is located.
+    ///
+    /// This is the URL at which updates are submitted.
+    pub fn described_by(&self) -> Option<&Url> {
+        self.described_by.as_ref()
+    }
+
+    /// The name of the file that was originally fetched.
+    pub fn origin_file_name(&self) -> Option<&str> {
+        self.file_name.as_deref()
+    }
+
+    /// The size of the file that was originally fetched.
+    pub fn origin_size(&self) -> Option<usize> {
+        self.size
+    }
+
     /// The underlying dataset.
     pub fn dataset(&self) -> &D {
         &self.dataset
@@ -72,7 +94,7 @@ impl<D> RdfSource<D> {
             .unwrap_or(self.origin().as_str());
         Quad::new(
             NamedOrBlankNode::NamedNode(NamedNode::new_unchecked(self.origin.clone())),
-            vocab::rdf::VALUE,
+            vocab::rdf::TYPE,
             Term::Literal(Literal::new_simple_literal("")),
             GraphName::NamedNode(NamedNode::new_unchecked(graph_name)),
         )
@@ -100,12 +122,12 @@ impl<D> RdfSource<D> {
 }
 
 /// Holds options related to serialization.
-pub struct SerializationOptions {
+pub struct SerializationOptions<'a> {
     format: RdfFormat,
-    filter: Box<dyn Fn(TripleRef<'_>) -> bool>,
+    filter: Box<dyn Fn(TripleRef<'a>) -> bool>,
 }
 
-impl SerializationOptions {
+impl<'a> SerializationOptions<'a> {
     /// Create a new set of serialization options with the provided format.
     pub fn from_format(format: RdfFormat) -> Self {
         Self {
@@ -120,7 +142,7 @@ impl SerializationOptions {
     #[must_use]
     pub fn with_filter<F>(self, filter: F) -> Self
     where
-        F: Fn(TripleRef<'_>) -> bool + 'static,
+        F: Fn(TripleRef<'a>) -> bool + 'static,
     {
         Self {
             format: self.format,
@@ -134,7 +156,7 @@ 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> {
+    pub fn serialize(&'a self, options: SerializationOptions<'a>) -> crate::Result<bytes::Bytes> {
         let writer = bytes::BytesMut::new().writer();
         let mut serializer = RdfSerializer::from_format(options.format).for_writer(writer);
 
@@ -159,7 +181,7 @@ where
     /// Prepare an update request.
     pub fn to_update(
         &'a self,
-        options: SerializationOptions,
+        options: SerializationOptions<'a>,
     ) -> crate::Result<RdfSourceUpdateRequest> {
         let url = self.described_by.clone().unwrap_or(self.origin.clone());
         let media_type = options.format.media_type().to_string();

ldp/src/resource.rs

@@ -1,11 +1,16 @@
+use crate::container::ContainerType;
 use crate::rdf_source::RdfSource;
-use crate::vocab;
+use crate::{Preference, prefer, vocab};
 use bytes::Bytes;
 use futures::Stream;
+use http::Method;
 use oxigraph::io::{RdfFormat, RdfParser};
 use oxigraph::model::{GraphNameRef, NamedNodeRef, Quad};
 use reqwest_middleware::reqwest::{Client, Response, StatusCode, Url, header};
 use reqwest_middleware::{ClientBuilder, ClientWithMiddleware, RequestBuilder};
+use sfv::{Item, TokenRef};
+use std::collections::BTreeSet;
+use std::str::FromStr;
 use tracing::error;
 
 /// Builds an HTTP request for a [Resource](https://www.w3.org/TR/ldp/#ldpr).
@@ -34,7 +39,10 @@ pub struct ResourceRequestBuilder {
     client: ClientWithMiddleware,
     url: Url,
     follow_described_by: bool,
+    validate_support: bool,
     formats: Vec<RdfFormat>,
+    include_preferences: Vec<Preference>,
+    omit_preferences: Vec<Preference>,
 }
 
 impl ResourceRequestBuilder {
@@ -52,17 +60,37 @@ impl ResourceRequestBuilder {
             client,
             url,
             follow_described_by: true,
+            validate_support: true,
             formats: Vec::new(),
+            include_preferences: Vec::new(),
+            omit_preferences: Vec::new(),
         }
     }
 
-    /// If the HTTP response includes a [describedby](https://www.w3.org/TR/ldp/#link-relation-describedby) link rel, then it will be followed.
+    /// If the HTTP response includes a [describedby](https://www.w3.org/TR/ldp/#link-relation-describedby)
+    /// link rel, then it will be followed. The original URL may be accessed using [`Resource::origin`],
+    /// and the link rel may be accessed using [`Resource::described_by`].
+    ///
+    /// Note: If your intention is to fetch a NonRDFSource (e.g. a PDF, video, etc.), this value
+    /// must be `false`.
+    ///
+    /// Default value: `true`.
     #[must_use]
     pub fn follow_described_by(mut self, value: bool) -> Self {
         self.follow_described_by = value;
         self
     }
 
+    /// Require the presence of a `Link: <http://www.w3.org/ns/ldp#Resource>; rel="type"` response header.
+    /// This is described in [section 4.2.1.4](https://www.w3.org/TR/ldp/#ldpr-resource) of the spec.
+    ///
+    /// Default value: `true`.
+    #[must_use]
+    pub fn validate_support(mut self, value: bool) -> Self {
+        self.validate_support = value;
+        self
+    }
+
     /// Restrict the request to the given RDF format.
     ///
     /// Repeated calls are additive. By default, all formats — even non-RDF formats — are accepted.
@@ -87,6 +115,18 @@ impl ResourceRequestBuilder {
         self
     }
 
+    /// Include a representation preference according to [section 7.2](https://www.w3.org/TR/ldp/#prefer-parameters).
+    pub fn include_preference(mut self, preference: Preference) -> Self {
+        self.include_preferences.push(preference);
+        self
+    }
+
+    /// Omit a representation preference according to [section 7.2](https://www.w3.org/TR/ldp/#prefer-parameters).
+    pub fn omit_preference(mut self, preference: Preference) -> Self {
+        self.omit_preferences.push(preference);
+        self
+    }
+
     /// Build the request.
     pub fn build(self) -> ResourceRequest {
         ResourceRequest { builder: self }
@@ -125,7 +165,7 @@ impl ResourceRequest {
         if response.status() == StatusCode::OK {
             let headers = response.headers();
             for link in headers.get_all(header::LINK) {
-                let link = link.to_str().unwrap_or("");
+                let link = link.to_str().unwrap_or_default();
                 match parse_link_header::parse(link) {
                     Ok(link_map) => {
                         if let Some(metadata_url) = link_map.get(&Some("type".to_string())) {
@@ -142,6 +182,34 @@ impl ResourceRequest {
         Err(crate::Error::LDPUnsupported)
     }
 
+    fn resource_type(response: &Response) -> Option<ResourceType> {
+        let headers = response.headers();
+        let links = headers
+            .get_all(header::LINK)
+            .iter()
+            .map(|hv| hv.to_str().unwrap_or_default())
+            .filter_map(|link| parse_link_header::parse_with_rel(link).ok())
+            .filter_map(|link_map| link_map.get("type").map(|item| item.raw_uri.clone()))
+            .collect::<BTreeSet<_>>();
+
+        // NB: The comparisons below must be done in this order, because multiple Link headers may
+        // be present with differing degrees of specificity. E.g. BasicContainer is more specific
+        // than RDFSource, but both are valid and may be present concomitantly.
+        if links.contains(vocab::ldp::BASIC_CONTAINER.as_str()) {
+            Some(ResourceType::RdfSource(Some(ContainerType::Basic)))
+        } else if links.contains(vocab::ldp::DIRECT_CONTAINER.as_str()) {
+            Some(ResourceType::RdfSource(Some(ContainerType::Direct)))
+        } else if links.contains(vocab::ldp::INDIRECT_CONTAINER.as_str()) {
+            Some(ResourceType::RdfSource(Some(ContainerType::Indirect)))
+        } else if links.contains(vocab::ldp::RDF_SOURCE.as_str()) {
+            Some(ResourceType::RdfSource(None))
+        } else if links.contains(vocab::ldp::NON_RDF_SOURCE.as_str()) {
+            Some(ResourceType::NonRdfSource)
+        } else {
+            None
+        }
+    }
+
     fn add_media_types(&self, mut request_builder: RequestBuilder) -> RequestBuilder {
         let media_types = self.builder.formats.iter().map(|f| f.media_type());
         for media_type in media_types {
@@ -150,19 +218,80 @@ impl ResourceRequest {
         request_builder
     }
 
+    fn is_method_allowed(response: &Response, method: Method) -> crate::Result<bool> {
+        for header in response.headers().get_all(header::ALLOW) {
+            for value in header.to_str()?.replace(" ", "").split(',') {
+                if value == method {
+                    return Ok(true);
+                }
+            }
+        }
+        Ok(false)
+    }
+
     /// Send the request.
     ///
-    /// There are two stages to this process. First, a HEAD request is made to determine whether
+    /// There are two stages to this process. First, a `HEAD` request is made to determine whether
     /// the requested resource is described by an RDF graph at another location. If it is, and if
-    /// the user permits it, then the URL in the
+    /// the user [permits it](`ResourceRequestBuilder::follow_described_by`), then the URL in the
     /// [describedby](https://www.w3.org/TR/ldp/#link-relation-describedby) header is used.
     /// Otherwise, the original URL is used.
     ///
-    /// During the second stage, a GET request is made. No parsing occurs at this time.
+    /// During the second stage, a `GET` request is made. **The response body is not consumed yet.**
+    /// To consume the response body, you must call either [`Resource::into_stream`] to receive the
+    /// raw bytes or [`Resource::into_rdf_source`] to parse the content and receive a [`RdfSource`].
+    ///
+    /// If the `HEAD` method is not allowed on this URL (a violation of the LDP spec), the `Allow`
+    /// header is inspected for `GET`. If absent, this method will return an error.
     pub async fn send(&self) -> crate::Result<Resource> {
         let request_builder = self.builder.client.head(self.builder.url.clone());
-        let mut response = request_builder.send().await?.error_for_status()?;
-        Self::ensure_ldp_support(&response)?;
+        let mut response = request_builder.send().await?;
+
+        match response.error_for_status_ref() {
+            Ok(response) => {
+                if self.builder.validate_support {
+                    Self::ensure_ldp_support(response)?;
+                }
+            }
+            Err(err) if err.status() == Some(StatusCode::METHOD_NOT_ALLOWED) => {
+                if !Self::is_method_allowed(&response, Method::GET)? {
+                    return Err(err.into());
+                }
+            }
+            err => {
+                err?;
+            }
+        }
+
+        let size = response
+            .headers()
+            .get(header::CONTENT_LENGTH)
+            .and_then(|hv| hv.to_str().ok())
+            .and_then(|et| usize::from_str(et).ok());
+
+        let content_disposition = response
+            .headers()
+            .get(header::CONTENT_DISPOSITION)
+            .and_then(|hv| hv.to_str().ok());
+
+        let file_name = if let Some(content_disposition) = content_disposition {
+            sfv::Parser::new(content_disposition)
+                .parse::<Item>()
+                .ok()
+                .and_then(|item| {
+                    if item.bare_item.as_token() == Some(TokenRef::constant("attachment")) {
+                        item.params
+                            .get("filename")
+                            .and_then(|item| item.as_string().map(|s| s.to_string()))
+                    } else {
+                        None
+                    }
+                })
+        } else {
+            None
+        };
+
+        let resource_type = Self::resource_type(&response);
 
         let url_to_get;
         let described_by = Self::extract_described_by(&response);
@@ -177,8 +306,20 @@ impl ResourceRequest {
         let mut request_builder = self.builder.client.get(url_to_get);
         request_builder = self.add_media_types(request_builder);
 
+        if !self.builder.include_preferences.is_empty() || !self.builder.omit_preferences.is_empty()
+        {
+            let header = prefer::header_for_preferences(
+                &self.builder.include_preferences,
+                &self.builder.omit_preferences,
+            )?;
+            request_builder = request_builder.headers(header);
+        }
+
         response = request_builder.send().await?.error_for_status()?;
-        Self::ensure_ldp_support(&response)?;
+
+        if self.builder.validate_support {
+            Self::ensure_ldp_support(&response)?;
+        }
 
         let state_token = response
             .headers()
@@ -202,12 +343,47 @@ impl ResourceRequest {
             origin: self.builder.url.clone(),
             described_by,
             state_token,
+            resource_type,
             format,
+            file_name,
+            size,
             response,
         })
     }
 }
 
+/// The type of [Resource](https://www.w3.org/TR/ldp/#ldpr-resource).
+#[derive(Clone, Debug)]
+pub enum ResourceType {
+    NonRdfSource,
+    RdfSource(Option<ContainerType>),
+}
+
+impl ResourceType {
+    pub fn as_named_node_ref(&self) -> NamedNodeRef<'_> {
+        match self {
+            ResourceType::NonRdfSource => vocab::ldp::NON_RDF_SOURCE,
+            ResourceType::RdfSource(None) => vocab::ldp::RDF_SOURCE,
+            ResourceType::RdfSource(Some(container_type)) => container_type.as_named_node_ref(),
+        }
+    }
+
+    pub fn from_named_node_ref(node: NamedNodeRef<'_>) -> Option<Self> {
+        match node {
+            vocab::ldp::NON_RDF_SOURCE => Some(ResourceType::NonRdfSource),
+            vocab::ldp::RDF_SOURCE => Some(ResourceType::RdfSource(None)),
+            _ => {
+                let container_type = ContainerType::from_named_node_ref(node);
+                if container_type.is_some() {
+                    Some(ResourceType::RdfSource(container_type))
+                } else {
+                    None
+                }
+            }
+        }
+    }
+}
+
 /// The format of the response, as determined by the `Content-Type` HTTP header.
 pub enum ResponseFormat {
     RdfFormat(RdfFormat),
@@ -217,17 +393,20 @@ pub enum ResponseFormat {
 
 /// A LDP [Resource](https://www.w3.org/TR/ldp/#ldpr).
 ///
-/// A Resource on its own is not very useful, as it is a very abstract concept. If the intention is
-/// to treat the response as opaque data (such as a media file), call [`Resource::into_stream`]. In
-/// the spec, this is equivalent to a [NonRDFSource](https://www.w3.org/TR/ldp/#ldpnr).
+/// A Resource on its own is not very useful, as it is a very abstract concept. If your intention is
+/// to treat the response as opaque data (such as a PDF or video file), call [`Resource::into_stream`].
+/// In the spec, this is equivalent to a [NonRDFSource](https://www.w3.org/TR/ldp/#ldpnr).
 ///
-/// If the intention is to treat the response as RDF data to be parsed, call
+/// If your intention is to treat the response as RDF data to be parsed, call
 /// [`Resource::into_rdf_source`].
 pub struct Resource {
     origin: Url,
     described_by: Option<Url>,
     state_token: Option<String>,
+    resource_type: Option<ResourceType>,
     format: ResponseFormat,
+    file_name: Option<String>,
+    size: Option<usize>,
     response: Response,
 }
 
@@ -244,24 +423,59 @@ impl Resource {
         self.described_by.as_ref()
     }
 
-    /// The format of the response.
+    /// The format of the response, as reported by the `Content-Type` HTTP header.
     pub fn format(&self) -> &ResponseFormat {
         &self.format
     }
 
+    /// The size of the content, as reported by the `Content-Length` HTTP header.
+    ///
+    /// Note: This value is extracted from the response headers of the HEAD request for the origin
+    /// URL.
+    pub fn size(&self) -> Option<usize> {
+        self.size
+    }
+
+    /// The file name of the content, as reported by the `Content-Disposition` HTTP header.
+    /// Example:
+    /// ```
+    /// Content-Disposition: attachment; filename="new-king-james-version-en.pdf"
+    /// ```
+    ///
+    /// Note: This value is extracted from the response headers of the HEAD request for the origin
+    /// URL.
+    pub fn file_name(&self) -> Option<&str> {
+        self.file_name.as_deref()
+    }
+
+    /// The type of Resource, as reported by the `Link: <...> rel="type"` HTTP header.
+    ///
+    /// Note: This value is extracted from the response headers of the HEAD request for the origin
+    /// URL.
+    pub fn resource_type(&self) -> Option<&ResourceType> {
+        self.resource_type.as_ref()
+    }
+
     /// The state token, as extracted from the `X-State-Token` header.
     ///
-    /// Note that this is a feature [specific to Fedora](https://fedora.info/2021/05/01/spec/#state-tokens).
+    /// <div class="warning">
+    /// This is a feature specific to <a href="https://fedora.info/2021/05/01/spec/#state-tokens">Fedora</a>.
+    /// </div>
     pub fn state_token(&self) -> Option<&String> {
         self.state_token.as_ref()
     }
 
-    /// Extract the response as a stream of unparsed bytes.
+    /// Provide the response body as a stream of bytes.
     pub fn into_stream(self) -> impl Stream<Item = reqwest_middleware::reqwest::Result<Bytes>> {
         self.response.bytes_stream()
     }
 
     /// Parse the response.
+    ///
+    /// The `D` type parameter can be any type capable of storing [`Quad`]s. If you intend to query
+    /// the dataset, use [`Dataset`](`oxigraph::model::Dataset`). If you want a stable ordering of the
+    /// quads, a [`KeyedDataset`](`crate::model::KeyedDataset`), [`Vec`], or similar data structure
+    /// may be used.
     pub async fn into_rdf_source<D>(self) -> crate::Result<RdfSource<D>>
     where
         D: FromIterator<Quad>,
@@ -275,8 +489,11 @@ impl Resource {
             let dataset = quads.collect::<Result<_, _>>()?;
             Ok(RdfSource {
                 origin: self.origin,
-                described_by: self.described_by,
+                origin_type: self.resource_type,
                 state_token: self.state_token,
+                described_by: self.described_by,
+                file_name: self.file_name,
+                size: self.size,
                 dataset,
             })
         } else {

ldp/src/traverse.rs

@@ -7,6 +7,7 @@ use reqwest_middleware::reqwest::Url;
 use std::collections::HashSet;
 use std::pin::Pin;
 use std::task::{Context, Poll, ready};
+use tracing::debug_span;
 
 /// Recursively traverse an entire repository.
 ///
@@ -21,6 +22,7 @@ pub struct Traverse<'a> {
     visited_urls: HashSet<Url>,
     pending_urls: HashSet<Url>,
     parallelism: Option<usize>,
+    span: tracing::Span,
 }
 
 impl<'a> Traverse<'a> {
@@ -35,6 +37,7 @@ impl<'a> Traverse<'a> {
             visited_urls: HashSet::new(),
             pending_urls: HashSet::new(),
             parallelism,
+            span: debug_span!("Traverse Repository", %root),
         };
         this.visited_urls.insert(root.clone());
         this.stream
@@ -89,6 +92,7 @@ impl<'a> Stream for Traverse<'a> {
     type Item = crate::Result<RdfSource<Dataset>>;
 
     fn poll_next(mut self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<Option<Self::Item>> {
+        let _enter = self.span.clone().entered();
         let result = match ready!(self.stream.poll_next_unpin(cx)) {
             Some(Ok(document)) => {
                 self.visited_urls.insert(document.origin().clone());

ldp/src/vocab.rs

@@ -5,4 +5,23 @@ pub mod ldp {
         NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#contains");
     pub const RESOURCE: NamedNodeRef<'_> =
         NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#Resource");
+    pub const NON_RDF_SOURCE: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#NonRDFSource");
+    pub const RDF_SOURCE: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#RDFSource");
+    pub const CONTAINER: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#Container");
+    pub const BASIC_CONTAINER: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#BasicContainer");
+    pub const DIRECT_CONTAINER: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#DirectContainer");
+    pub const INDIRECT_CONTAINER: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#IndirectContainer");
+
+    pub const PREFER_CONTAINMENT: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#PreferContainment");
+    pub const PREFER_MEMBERSHIP: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#PreferMembership");
+    pub const PREFER_MINIMAL_CONTAINER: NamedNodeRef<'_> =
+        NamedNodeRef::new_unchecked("http://www.w3.org/ns/ldp#PreferMinimalContainer");
 }