canonicalization, tracing, mitsein

Author: Porges

Cargo.lock

@@ -14,11 +14,13 @@ name = "html2rdf"
 path = "src/main.rs"
 
 [dependencies]
-clap = { version = "4.5.59", features = ["derive", "wrap_help"] }
+clap = { version = "4.5.59", features = ["cargo", "derive", "wrap_help"] }
 html2rdf = { workspace = true }
+jiff = "0.2.20"
 oxiri = { workspace = true }
-oxrdf = { workspace = true }
+oxrdf = { workspace = true, features = ["rdfc-10"] }
 oxttl = { workspace = true }
 reqwest = { version = "0.13.2", default-features = false, features = ["charset", "http2", "native-tls", "system-proxy"] }
 tokio = { version = "1.49.0", features = ["macros", "rt"] }
+tracing-subscriber = { version = "0.3.22" }
 url = "2.5.8"

html2rdf-cli/Cargo.toml

@@ -1,16 +1,22 @@
-use std::process::ExitCode;
+use std::{io::Write, process::ExitCode};
 
 use clap::Parser;
+use oxrdf::graph::{CanonicalizationAlgorithm, CanonicalizationHashAlgorithm};
 
 #[derive(Parser)]
 #[command(version, about)]
 struct Args {
     #[arg(value_name = "URL")]
     target: url::Url,
+
+    #[arg(long, short = 'c')]
+    canonicalize: bool,
 }
 
 #[tokio::main(flavor = "current_thread")]
 async fn main() -> Result<ExitCode, Box<dyn std::error::Error>> {
+    tracing_subscriber::fmt::init();
+
     let args = Args::parse();
     let client = reqwest::Client::new();
     let base = args.target.to_string();
@@ -28,6 +34,8 @@ async fn main() -> Result<ExitCode, Box<dyn std::error::Error>> {
 
     drop(client);
 
+    let final_url = response.url().clone();
+
     let content = response.text().await?;
     let mut output_graph = oxrdf::Graph::new();
     let mut processor_graph = oxrdf::Graph::new();
@@ -58,7 +66,27 @@ async fn main() -> Result<ExitCode, Box<dyn std::error::Error>> {
             |serializer, (prefix, value)| serializer.with_prefix(prefix, value),
         )?;
 
+        if args.canonicalize {
+            output_graph.canonicalize(CanonicalizationAlgorithm::Rdfc10 {
+                hash_algorithm: CanonicalizationHashAlgorithm::Sha256,
+            });
+        }
+
         let mut locked_out = std::io::stdout().lock();
+        locked_out.write_all(
+            format!(
+                "# generated by html2rdf {} at {:.0} from: {}\n",
+                clap::crate_version!(),
+                jiff::Timestamp::now(),
+                final_url,
+            )
+            .as_bytes(),
+        )?;
+
+        if args.canonicalize {
+            locked_out.write_all(b"# (output has been canonicalized)\n")?;
+        }
+
         let mut writer = serializer.for_writer(&mut locked_out);
         for triple in output_graph.iter() {
             writer.serialize_triple(triple)?;