more tests, add configurable well known prefix

nbittich · Jan 27, 2024 · 768a5fc · 768a5fc
1 parent 38178e1
commit 768a5fc
Show file tree

Hide file tree

Showing 15 changed files with 277 additions and 110 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -28,7 +28,7 @@ test-case = "3.3.1"
 env_logger = "0.11.1"
 tortank = "0.20.0"
 wasm-bindgen = "0.2.90"
-graph-rdfa-processor = "0.1.2"
+graph-rdfa-processor = { path = "./lib-rdfa" }
 
 [profile.release]
 opt-level = 'z'   # Optimize for size.

diff --git a/README.md b/README.md
@@ -1,23 +1,58 @@
-# graph-rdfa-processor
+# RDFa processor library (WIP)
 
-### RDFa processor library (WIP)
+Rust and wasm library to extract rdf statements (in n-triples format) from an html file
+based on rdfa annotations.
 
-- covers [RDFa 1.1 Primer - Third Edition](https://www.w3.org/TR/rdfa-primer/).
-- used [RDFa/Play](https://rdfa.info/play/) for comparing.
-- WIP:
+## Examples
+
+```rust
+use graph_rdfa_processor::RdfaGraph;
+    let html = r#"
+    <div prefix="foaf: http://xmlns.com/foaf/0.1/" about="http://www.example.org/#somebody" rel="foaf:knows">
+        <p about="http://danbri.org/foaf.rdf#danbri" typeof="foaf:Person" property="foaf:name">Dan Brickley</p>
+	  </div>
+    "#;
+    let base = "http://example.com"
+    let well_known_prefix = Some("http://example.org/.well_known");
+
+    RdfaGraph::parse_str(html, base, well_known_prefix).unwrap()
+
+```
+
+```js
+ <script type="module">
+    import init, {html_to_rdfa} from "./rdfa-wasm/pkg/rdfa_wasm.js";
+
+    async function run() {
+      await init();
+      let html =`
+        <div prefix="foaf: http://xmlns.com/foaf/0.1/" about="http://www.example.org/#somebody" rel="foaf:knows">
+            <p about="http://danbri.org/foaf.rdf#danbri" typeof="foaf:Person" property="foaf:name">Dan Brickley</p>
+	      </div>
+      `;
+      let base = "http://example.com"
+      let well_known_prefix = "http://example.org/.well_known";
+      let res = html_to_rdfa(html, base, well_known_prefix);
 
+    }
+    run();
+  </script>
+
+```
+
+- covers:
+
+  - [RDFa 1.1 Primer - Third Edition](https://www.w3.org/TR/rdfa-primer/)
   - [RDFa Core](https://www.w3.org/TR/rdfa-core/)
   - [Earl-Reports](https://rdfa.info/earl-reports/#RDFa-rdfa1.1-tests-for-html5)
 
+- used [RDFa/Play](https://rdfa.info/play/) for comparing.
 - [Demo](https://nbittich.github.io/graph-rdfa-processor/)
 
-  ### WIP
-
-  The plan is to make as much tests as possible, and fix the bugs along the way.
+### WIP
 
-  Once enough tests have been implemented, a full refactoring will be needed before
-  doing anything else.
+The plan is to make as much tests as possible, and fix the bugs along the way.
 
-  ### KNOWN ISSUES
+### KNOWN ISSUES
 
-  - White space not preserved. See test 0329.
+- White space not preserved. See test 0329.
diff --git a/index.html b/index.html
@@ -17,28 +17,64 @@
       form.classList.remove("d-none");
 
       const text_area = document.querySelector("#html");
-      text_area.value = `<!DOCTYPE html>
-<html prefix="foaf: http://xmlns.com/foaf/0.1/">
-  <head>
-		<title>Test 0083</title>    
-  </head>
-  <body>
-	<div about="http://www.example.org/#somebody" rel="foaf:knows">
-	    <p property="foaf:name">Ivan Herman</p>
-		<p rel="foaf:mailbox" resource="mailto:ivan@w3.org">mailto:ivan@w3.org</p>
-		<p about="http://danbri.org/foaf.rdf#danbri" typeof="foaf:Person" property="foaf:name">Dan Brickley</p>
-	</div>
-  </body>
-</html>
-`;
+      text_area.value = `
+            <!DOCTYPE html>
+            <html prefix="foaf: http://xmlns.com/foaf/0.1/">
+              <head>
+              <title>Test 0083</title>
+              </head>
+              <body>
+             <div about="http://www.example.org/#somebody" rel="foaf:knows">
+                 <p property="foaf:name">Ivan Herman</p>
+              <p rel="foaf:mailbox" resource="mailto:ivan@w3.org">mailto:ivan@w3.org</p>
+              <p about="http://danbri.org/foaf.rdf#danbri" typeof="foaf:Person" property="foaf:name">Dan Brickley</p>
+             </div>
+              </body>
+            </html>
+        `;
 
       form.addEventListener("submit", (e) => {
         e.preventDefault();
         const data = new FormData(e.target);
-        let res = html_to_rdfa(data.get("html"), data.get("base"));
+        let res = html_to_rdfa(
+          data.get("html") || "",
+          data.get("base") || "",
+          data.get("wellKnownPrefix") || "",
+        );
         const out = document.querySelector("pre");
         out.innerText = res;
       });
+      const issueLink = document.querySelector("#issueLink");
+      issueLink.onclick = (e) => {
+        e.preventDefault();
+        const a = document.createElement("a");
+        let params = new URLSearchParams();
+        const data = new FormData(form);
+        params.append("title", "RDFa processing bug");
+
+        params.append(
+          "body",
+          `
+### Base: 
+  
+  \`${data.get("base") || ""}\`
+
+### Well Known Prefix:
+
+  \`${data.get("base") || ""}\`
+
+### Html:
+  
+   \`\`\`html
+   ${data.get("html") || ""}
+   \`\`\`
+
+          `,
+        );
+        a.href = `https://github.com/nbittich/graph-rdfa-processor/issues/new?${params.toString()}`;
+        a.target = "_blank";
+        a.click();
+      };
     }
     run();
   </script>
@@ -48,11 +84,11 @@
   <h1>RDFa Playground</h1>
   <p>
     This is a poc of an rdfa library written in rust. If you find any bug,
-    please open an issue.
+    please <a href="#issueLink" id="issueLink">open an issue.</a>
   </p>
   <hr />
-  <div class="row">
-    <div class="col-4 ms-1">
+  <div class="d-flex justify-content-center">
+    <div class="col-3 me-1 border p-2">
       <form action="" id="rdfaForm">
         <div class="mb-2 align-items-center">
           <div class="col-auto">
@@ -63,6 +99,20 @@ <h1>RDFa Playground</h1>
               value="http://example.com" />
           </div>
         </div>
+        <div class="mb-2 align-items-center">
+          <div class="col-auto">
+            <label for="base" class="col-form-label">Well Known prefix</label>
+          </div>
+          <div class="col-auto">
+            <input required type="url" id="wellKnownPrefix" name="wellKnownPrefix" class="form-control"
+              aria-describedby="well known prefix" value="http://example.com/.well-known" />
+            <div id="passwordHelpBlock" class="form-text">
+              Well known prefix for
+              <a target="_blank" href="https://www.w3.org/2011/rdf-wg/wiki/Skolemisation">Skolemisation</a>
+            </div>
+          </div>
+        </div>
+
         <div class="mb-2 align-items-center">
           <div class="col-auto">
             <label for="html" class="col-form-label">Html</label>
@@ -72,10 +122,12 @@ <h1>RDFa Playground</h1>
               aria-describedby="Html content"></textarea>
           </div>
         </div>
-        <button type="submit" class="btn btn-outline-danger">Submit</button>
+        <div class="d-grid gap-2">
+          <button type="submit" class="btn btn-outline-danger">Submit</button>
+        </div>
       </form>
     </div>
-    <div class="col-7 pt-5 bg-light">
+    <div class="col-8 pt-5 bg-light border">
       <pre id="out"></pre>
     </div>
   </div>

diff --git a/lib-rdfa/examples/earl_html5/example0035.html b/lib-rdfa/examples/earl_html5/example0035.html
@@ -0,0 +1,15 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/MarkUp/DTD/html401-rdfa11-1.dtd">
+<html version="HTML+RDFa 1.1" prefix="foaf: http://xmlns.com/foaf/0.1/">
+  <head>
+		<title>Test 0035</title>    
+  </head>
+  <body>
+  	<div>
+      <img 	about="http://sw-app.org/mic.xhtml#i"
+  	   		rel="foaf:img"
+  	   		src="http://sw-app.org/img/mic_2007_01.jpg" 
+  	   		href="http://sw-app.org/img/mic_2006_03.jpg" 
+  	   		alt="A photo depicting Michael" />  	
+  	</div>
+  </body>
+</html>
diff --git a/lib-rdfa/examples/earl_html5/example0035.ttl b/lib-rdfa/examples/earl_html5/example0035.ttl
@@ -0,0 +1 @@
+<http://sw-app.org/mic.xhtml#i> <http://xmlns.com/foaf/0.1/img> <http://sw-app.org/img/mic_2007_01.jpg>.
diff --git a/lib-rdfa/examples/earl_html5/example0037.html b/lib-rdfa/examples/earl_html5/example0037.html
@@ -0,0 +1,18 @@
+
+
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/MarkUp/DTD/html401-rdfa11-1.dtd">
+<html version="HTML+RDFa 1.1" prefix="foaf: http://xmlns.com/foaf/0.1/">
+  <head>
+		<title>Test 0037</title>    
+  </head>
+  <body>
+  	<div>
+      <img 	about="http://sw-app.org/mic.xhtml#i"
+  	   		rel="foaf:img"
+  	   		src="http://sw-app.org/img/mic_2007_01.jpg" 
+  	   		href="http://sw-app.org/img/mic_2006_03.jpg"
+  	   		resource="http://sw-app.org/mic.xhtml#photo" 
+  	   		alt="A photo depicting Michael" />  	
+  	</div>
+  </body>
+</html>
diff --git a/lib-rdfa/examples/earl_html5/example0037.ttl b/lib-rdfa/examples/earl_html5/example0037.ttl
@@ -0,0 +1,5 @@
+
+
+<http://sw-app.org/mic.xhtml#i> <http://xmlns.com/foaf/0.1/img> <http://sw-app.org/mic.xhtml#photo>.
+
+
diff --git a/lib-rdfa/examples/earl_html5/example0039.html b/lib-rdfa/examples/earl_html5/example0039.html
@@ -0,0 +1,15 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/MarkUp/DTD/html401-rdfa11-1.dtd">
+<html version="HTML+RDFa 1.1" prefix="foaf: http://xmlns.com/foaf/0.1/">
+  <head>
+		<title>Test 0039</title>    
+  </head>
+  <body>
+  	<div>
+      <img 	about="http://sw-app.org/mic.xhtml#i"
+  	   		rev="foaf:depicts"
+  	   		src="http://sw-app.org/img/mic_2007_01.jpg" 
+  	   		href="http://sw-app.org/img/mic_2006_03.jpg" 
+  	   		alt="A photo depicting Michael" />  	
+  	</div>
+  </body>
+</html>
diff --git a/lib-rdfa/examples/earl_html5/example0039.ttl b/lib-rdfa/examples/earl_html5/example0039.ttl
@@ -0,0 +1 @@
+<http://sw-app.org/img/mic_2007_01.jpg> <http://xmlns.com/foaf/0.1/depicts> <http://sw-app.org/mic.xhtml#i>.
diff --git a/lib-rdfa/src/lib.rs b/lib-rdfa/src/lib.rs
@@ -27,19 +27,28 @@ impl<'a> RdfaGraph<'a> {
         initial_context: Context<'a>,
     ) -> Result<RdfaGraph<'a>, Box<dyn Error>> {
         let mut triples = vec![];
+        let well_known_prefix = initial_context.well_known_prefix.clone();
         traverse_element(input, None, initial_context, &mut triples, &mut vec![])?;
 
         triples = copy_pattern(triples)?;
 
-        Ok(RdfaGraph(triples.into_iter().collect()))
+        Ok(RdfaGraph {
+            statements: triples.into_iter().collect(),
+            well_known_prefix,
+        })
     }
     // temporary thing
-    pub fn parse_str(html: &'a str, base: &'a str) -> Result<String, Box<dyn Error>> {
+    pub fn parse_str(
+        html: &'a str,
+        base: &'a str,
+        well_known_prefix: Option<&'a str>,
+    ) -> Result<String, Box<dyn Error>> {
         let document = scraper::Html::parse_document(html);
         let root = document.root_element();
 
         let root_ctx = Context {
             base,
+            well_known_prefix,
             ..Default::default()
         };
         RdfaGraph::parse(&root, root_ctx).map(|g| g.to_string())