Cargo.toml

@@ -4,4 +4,9 @@ version = "0.1.0"
 edition = "2021"
 
 [dependencies]
-spider = { features = [], git="https://github.com/Rushmore75/spider.git", rev="ff91646973ad04ff423010f36206f550e37c4278" }
+html5ever = "0.29.0"
+markup5ever_rcdom = "0.5.0-unofficial"
+reqwest = "0.12.8"
+tokio = { version="1.40.0", features = ["full"] }
+tracing = "0.1.40"
+tracing-subscriber = "0.3.18"

src/main.rs

@@ -1,35 +1,58 @@
-use spider::{hashbrown::HashMap, tokio};
+extern crate markup5ever_rcdom as rcdom;
+extern crate html5ever;
+
+use std::env;
+use html5ever::{parse_document, tendril::TendrilSink, tree_builder::TreeBuilderOpts, ParseOpts};
+use rcdom::RcDom;
+use tracing::{debug, info, trace, warn};
+
 
 #[tokio::main]
 async fn main() {
+    tracing_subscriber::fmt::init();
+    debug!("Starting...");
 
-    let args = std::env::args().collect::<Vec<String>>();
+    let url = "https://oliveratkinson.net";
+    let budget = "10";
+    
+    let response = reqwest::get(url).await.unwrap();
+    let data = response.text().await.unwrap();
 
-    let url = &args[1];
-    let budget = &args[2];
-    let budget = match budget.parse::<u32>() {
-        Ok(x) => x,
-        Err(_) => panic!("Second arg must be a int"),
+
+    let opts = ParseOpts {
+        tree_builder: TreeBuilderOpts {
+            drop_doctype: true,
+            ..Default::default()
+        },
+        ..Default::default()
     };
 
-
-    let mut site = spider::website::Website::new(url)
-        .with_budget(Some(HashMap::from([
-            ("*", budget),
-        ])))
-        .with_tld(true)
-        .with_on_link_find_callback(Some(|from, to| {
-            let from = from.as_ref().to_string();
-            let to = to.as_ref().to_string();
-
-            let from = from.trim();
-            let to= to.trim();
-
-            println!("{from};->;{to}"); 
-        }))
-        .build()
+    let dom = parse_document(RcDom::default(), opts)
+        .from_utf8()
+        .read_from(&mut data.as_bytes())
         .unwrap();
 
-    site.crawl().await;
+    let a = &dom.document;
+    warn!("Walking...");
+
+    walk(a);
 }
 
+fn walk(node: &rcdom::Handle) {
+    match &node.data {
+        rcdom::NodeData::Document => (),
+        rcdom::NodeData::Doctype { name, public_id, system_id } => debug!("doctype"),
+        rcdom::NodeData::Text { contents } => {},
+        rcdom::NodeData::Comment { contents } => debug!("comment"),
+        rcdom::NodeData::Element { name, attrs, template_contents, mathml_annotation_xml_integration_point } => {
+            attrs.borrow().iter().for_each(|attr| {
+                let name = name.local.to_string();
+                let internal = &*attr.value;
+                debug!("element: {name}, attr: {internal}"); 
+            });
+        },
+        rcdom::NodeData::ProcessingInstruction { target, contents } => debug!("ProcessingInstruction"),
+    };
+
+    node.children.borrow().iter().for_each(|n| walk(n));
+}