Crawler.toml

@@ -3,8 +3,9 @@ surreal_url = "localhost:8000"
 surreal_username = "root"
 surreal_password = "root"
 surreal_ns = "test"
-surreal_db = "v1.19.2"
+surreal_db = "v1.19.5"
 
 # Crawler config
-crawl_filter = "en.wikipedia.com" 
-budget = 1000
+crawl_filter = "en.wikipedia.org" 
+start_url = "https://en.wikipedia.org"
+budget = 100

README.md

@@ -4,11 +4,12 @@ Crawls sites saving all the found links to a surrealdb database. It then proceed
 
 ### TODO
 
-- [ ] Domain filtering - prevent the crawler from going on alternate versions of wikipedia.
+- [x] Domain filtering - prevent the crawler from going on alternate versions of wikipedia.
 - [ ] Conditionally save content - based on filename or file contents
 - [x] GUI / TUI ? - Graphana
 - [x] Better asynchronous getting of the sites. Currently it all happens serially.
-- [ ] Allow for storing asynchronously
+- [x] Allow for storing asynchronously - dropping the "links to" logic fixes this need
+- [x] Control crawler via config file (no recompliation needed)
 
 3/17/25: Took >1hr to crawl 100 pages
 

src/filesystem.rs

@@ -59,6 +59,13 @@ fn valid_file_extension(take: &&OsStr) -> bool {
                 "pdf" => true,
                 "json" => true,
                 "xml" => true,
+
+                // IGNORE
+                // TODO Should this be a list of all domains?
+                "org" => false,
+                "com" => false,
+                "net" => false,
+
                 _ => {
                     warn!("Might be forgetting a file extension: {s}");
                     false

src/main.rs

@@ -33,11 +33,14 @@ struct Config {
     surreal_password: String,
 
     crawl_filter: String,
+    start_url: String,
     budget: usize,
 }
 
 #[tokio::main]
 async fn main() {
+    println!("Logs and metrics are provided to the Grafana dashboard");
+
     let writer = std::fs::OpenOptions::new()
         .append(true)
         .create(true)
@@ -70,8 +73,7 @@ async fn main() {
         .expect("failed to install recorder/exporter");
 
     info!("Starting...");
-    // Would probably take these in as parameters from a cli
-    let starting_url = "https://en.wikipedia.org/";
+
     // When getting uncrawled pages, name must contain this variable. "" will effectively get ignored.
     // let crawl_filter = "en.wikipedia.org/";
     // let budget = 50;
@@ -82,6 +84,7 @@ async fn main() {
     let _ = file.read_to_string(&mut buf);
 
     let config: Config = toml::from_str(&buf).expect("Failed to parse Crawler.toml");
+    let starting_url = &config.start_url;
 
     let db = connect(&config)
         .await
@@ -138,6 +141,15 @@ async fn main() {
     }
     drop(span);
 
+    if let Ok(mut ok) = db.query("count(select id from website where crawled = true)").await {
+        let res = ok.take::<Option<usize>>(0);
+        if let Ok(i) = res {
+            if let Some(n) = i {
+                info!("Total crawled pages now equals {n}");
+            }
+        }
+    }
+
     info!("Done");
 }