back to spider-rs

Cargo.toml

@@ -8,9 +8,10 @@ html5ever = "0.29.0"
 # minio = "0.1.0"
 minio = {git="https://github.com/minio/minio-rs.git", rev = "c28f576"}
 reqwest = "0.12.9"
-serde = { version = "1.0.214", features = ["derive"] }
-surrealdb = "2.0.4"
-tokio = { version="1.41.0", features = ["full"] }
-tracing = "0.1.40"
-tracing-subscriber = { version = "0.3.18", features = ["env-filter"] }
+serde = { version = "1.0", features = ["derive"] }
+spider = { version = "2.21", features = ["sync"] }
+surrealdb = "2.0"
+tokio = { version="1.41", features = ["full"] }
+tracing = "0.1"
+tracing-subscriber = { version = "0.3", features = ["env-filter"] }
 url = { version = "2.5.3", features = ["serde"] }

src/db.rs

@@ -22,31 +22,20 @@ pub struct Website {
 
 impl Website {
     /// Creates a blank site (assumes that url param is site's root)
-    pub fn new(url: &str, crawled: bool) -> Self {
-        let site = match Url::parse(url) {
-            Ok(a) => a,
-            Err(_) => todo!(),
-        };
+    pub fn new(url: Url, crawled: bool) -> Self {
         Self {
             id: None,
             crawled,
-            site,
+            site: url,
         }
     }
 
-    pub fn set_crawled(&mut self) {
-        trace!("Set crawled to true");
-        self.crawled = true
-    }
-
     #[instrument(skip_all)]
     pub async fn links_to(&self, other: Vec<Thing>, db: &Surreal<Client>) {
         let len = other.len();
         if len == 0 {return}
 
         let from = self.site.to_string();
-        // let to = other.site.to_string();
-        trace!("Linking {from} to {} other pages.", other.len());
         let msg = format!("Linked {len} pages");
         let timer = Timer::start(&msg);
         // prevent the timer from being dropped instantly.
@@ -65,7 +54,6 @@ impl Website {
                     let _: Vec<usize> = vec;
                     if let Some(num) = vec.get(0) {
                         if *num == len {
-                            trace!("Link OK");
                             return;
                         } else {
                             warn!("Didn't link all the records. {num}/{len}");

src/main.rs

@@ -3,14 +3,17 @@ extern crate html5ever;
 use std::time::Instant;
 
 use db::{connect, Website};
+use parser::parse;
 use s3::S3;
 use surrealdb::{engine::remote::ws::Client, Surreal};
-use tracing::{debug, info, instrument, trace, trace_span};
+use tokio::sync::broadcast::Receiver;
+use tracing::{debug, info, trace, trace_span};
 use tracing_subscriber::EnvFilter;
+use url::Url;
 
 mod db;
-mod s3;
 mod parser;
+mod s3;
 
 struct Config<'a> {
     surreal_ns: &'a str,
@@ -33,94 +36,82 @@ async fn main() {
         .without_time()
         .init();
     debug!("Starting...");
-    
+
     let config = Config {
-        surreal_ns: "test",
-        surreal_db: "v1.7",
         surreal_url: "localhost:8000",
         surreal_username: "root",
         surreal_password: "root",
+        surreal_ns: "test",
+        surreal_db: "v1.9",
+        s3_bucket: "v1.9",
         s3_url: "http://localhost:9000",
-        s3_bucket: "v1.7",
-        s3_access_key: "8tUJn7e1paMFZQr0PKIT",
-        s3_secret_key: "uSMvYxNOeCejCUgXVqgTfYlUEcmiZY0xcZ91M9E0",
+        s3_access_key: "0zv7GbLQsw4ZI8TclMps",
+        s3_secret_key: "5dB7QkGFw7fYbUJ5LpHk2GbWR7Bl710HlRz4NbzB",
     };
 
     // Would probably take these in as parameters from a cli
-    let starting_url = "https://oliveratkinson.net/";
-    let budget = 15;
-    let mut crawled = 0;
+    // let starting_url = "https://oliveratkinson.net/";
+    let starting_url = "https://en.wikipedia.org/wiki/Main_Page";
 
-    let s3 = S3::connect(&config).await.expect("Failed to connect to minio, aborting.");
-    let db = connect(&config).await.expect("Failed to connect to surreal, aborting.");
-    
-    let reqwest = reqwest::Client::builder()
-        // .use_rustls_tls()
+    let s3 = S3::connect(&config)
+        .await
+        .expect("Failed to connect to minio, aborting.");
+    let db = connect(&config)
+        .await
+        .expect("Failed to connect to surreal, aborting.");
+
+    let mut site = spider::website::Website::new(&starting_url)
+        .with_limit(4)
+        .with_depth(0)
         .build()
         .unwrap();
 
-    // Kick off the whole machine - This Website object doesn't matter, it's just to allow for
-    // get() to work.
-    let span = trace_span!("Pre-Loop");
-    let pre_loop_span = span.enter();
-    // Download the site
-    let mut site = Website::new(&starting_url, false);
-    get(&mut site, &db, &reqwest, &s3, &mut crawled).await;
+    let mut rx: Receiver<spider::page::Page> = site.subscribe(0).unwrap();
 
-    drop(pre_loop_span);
+    let subscriber = tokio::spawn(async move {
+        let span = trace_span!("Sub");
+        let span = span.enter();
+        while let Ok(res) = rx.recv().await {
+            // Get body
+            let data = res.get_html();
+            let url = Url::parse(res.get_url()).unwrap();
 
-    let span = trace_span!("Loop");
-    let span = span.enter();
-    while crawled < budget {
-        let get_num = if budget - crawled < 100 { budget - crawled } else { 100 };
+            trace!("Got '{}'", url.to_string());
+            // Store document
+            s3.store(&data, &url).await;
 
-        let uncrawled = get_uncrawled_links(&db, get_num).await;
-        if uncrawled.len() == 0 {
-            info!("Had more budget but finished crawling everything.");
-            return;
+            // Parse document and store relationships
+            let mut page = Website::new(url, true);
+            page.store(&db).await;
+
+            // Relate this page to all the pages it links to
+            let span = trace_span!("Linking");
+            let span = span.enter();
+            let found_links = parse( &page, data).await;
+            let mut stored_links = Vec::new();
+            for mut link in found_links {
+                if let Some(id) = link.store(&db).await {
+                    stored_links.push(id);
+                }
+            }
+            page.links_to(stored_links, &db).await;
+            drop(span);
         }
-        debug!("Crawling {} pages...", uncrawled.len());
+        drop(span);
+    });
 
-        let span = trace_span!("Crawling");
-        let _ = span.enter();
+    let timer = Timer::start("Crawled");
 
-        for mut site in uncrawled {
-            get(&mut site, &db, &reqwest, &s3, &mut crawled).await;
+    site.crawl().await;
+    site.unsubscribe();
 
-            let percent = format!("{:.2}%", (crawled as f32 / budget as f32) * 100f32);
-            info!("Crawled {crawled} out of {budget} pages. ({percent})");
-        }
-    }
-    drop(span);
+    drop(timer);
+
+    subscriber.await.unwrap();
 
     info!("Done");
 }
 
-#[instrument(skip_all)]
-/// Downloads and crawls and stores a webpage.
-async fn get(
-    site: &mut Website,
-    db: &Surreal<Client>,
-    reqwest: &reqwest::Client,
-    s3: &S3,
-    count: &mut usize,
-) {
-    trace!("Get: {}", site.to_string());
-    let timer = Timer::start("Got page");
-    if let Ok(response) = reqwest.get(site.to_string()).send().await {
-        timer.stop();
-
-        // Get body
-        let data = response.text().await.unwrap();
-        // Store document
-        s3.store(&data, &site.site).await; 
-        // Parse document and store relationships
-        parser::parse(db, site, data).await;
-        *count += 1;
-    }
-    trace!("Failed to get: {}", site.to_string());
-}
-
 /// Returns uncrawled links
 async fn get_uncrawled_links(db: &Surreal<Client>, mut count: usize) -> Vec<Website> {
     if count > 100 {

src/parser.rs

@@ -5,8 +5,7 @@ use html5ever::tokenizer::{BufferQueue, TokenizerResult};
 use html5ever::tokenizer::{StartTag, TagToken};
 use html5ever::tokenizer::{Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts};
 use html5ever::{local_name, tendril::*};
-use surrealdb::engine::remote::ws::Client;
-use surrealdb::Surreal;
+use tracing::{instrument, trace};
 
 use crate::db::Website;
 
@@ -67,11 +66,10 @@ impl TokenSink for LinkParser<'_> {
     }
 }
 
-pub async fn parse(db: &Surreal<Client>, site: &mut Website, data: String) {
-    
-    site.set_crawled();
-    site.store(db).await;
-    
+
+#[instrument(skip_all)]
+pub async fn parse(site: &Website, data: String) -> Vec<Website> {
+
     let sink = LinkParser { site };
     let chunk = Tendril::from_str(&data).unwrap();
     let mut input = BufferQueue::default();
@@ -79,17 +77,16 @@ pub async fn parse(db: &Surreal<Client>, site: &mut Website, data: String) {
 
     let token = Tokenizer::new(sink.clone(), TokenizerOpts::default());
     
-    let mut links_to = Vec::new();
+    // let mut links_to = Vec::new();
+    let mut res = Vec::new();
     while !input.is_empty() {
-        if let TokenizerResult::Script(s) = token.feed(&mut input) {
-            for mut web in s {
-                if let Some(id) = web.store(db).await {
-                    links_to.push(id);
-                }
-            }
+        if let TokenizerResult::Script(mut s) = token.feed(&mut input) {
+            res.append(&mut s);
         }
     }
-    sink.site.links_to(links_to, db).await;
+    trace!("Found {} links.", res.len());
+
     assert!(input.is_empty());
     token.end();
+    res
 }