changes

Reviewed-on: #6

Crawler.toml

@@ -3,9 +3,10 @@ surreal_url = "localhost:8000"
 surreal_username = "root"
 surreal_password = "root"
 surreal_ns = "test"
-surreal_db = "v1.19.5"
+surreal_db = "v1.21.1"
 
 # Crawler config
-crawl_filter = "en.wikipedia.org" 
+crawl_filter = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI" 
-start_url = "https://en.wikipedia.org"
+start_url = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI"
-budget = 100
+budget = 10000
+batch_size = 50

docker/prometheus.yaml

@@ -7,7 +7,7 @@ scrape_configs:
     static_configs:
     # change this your machine's ip, localhost won't work
     # because localhost refers to the docker container.
-      - targets: ['172.20.239.48:2500']
+      - targets: ['192.168.1.200:2500']
         #- targets: ['192.168.8.209:2500']
   - job_name: loki
     static_configs:

src/db.rs

@@ -20,12 +20,18 @@ pub struct Website {
     pub site: Url,
     /// Wether or not this link has been crawled yet
     pub crawled: bool,
+    /// 200, 404, etc
+    pub status_code: u16,
 }
 
 // manual impl to make tracing look nicer
 impl Debug for Website {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("Website").field("site", &self.site).finish()
+        f.debug_struct("Website")
+            .field("host", &self.site.host())
+            .field("path", &self.site.path())
+            .field("status_code", &self.status_code)
+            .finish()
     }
 }
 
@@ -38,15 +44,11 @@ impl Website {
         };
         Self {
             crawled,
-            site
+            site,
+            status_code: 0,
         }
     }
 
-    pub fn set_crawled(&mut self) {
-        trace!("Set crawled to true");
-        self.crawled = true
-    }
-
     // Insert ever item in the vec into surreal, crawled state will be preserved as TRUE
     // if already in the database as such or incoming data is TRUE.
     #[instrument(skip(db))]
@@ -54,11 +56,13 @@ impl Website {
         counter!(STORE).increment(1);
         let mut things = Vec::with_capacity(all.len());
 
+        // FIXME failes *sometimes* because "Resource Busy"
         match db
             .query(
                 "INSERT INTO website $array
                     ON DUPLICATE KEY UPDATE
                         accessed_at = time::now(),
+                        status_code = $input.status_code,
                         crawled = crawled OR $input.crawled
                     RETURN VALUE id;
                  ",

src/filesystem.rs

@@ -1,30 +1,37 @@
-use std::{ffi::OsStr, io::ErrorKind, path::PathBuf};
+use std::{io::ErrorKind, path::PathBuf};
 
+use reqwest::header::HeaderValue;
 use tokio::fs;
-use tracing::{error, trace};
+use tracing::{error, trace, warn};
 use url::Url;
 
-pub fn as_path(url: &Url) -> PathBuf {
+pub fn as_path(url: &Url, content_type: &HeaderValue) -> PathBuf {
     // extract data from url to save it accurately
-    let url_path = PathBuf::from("./downloaded/".to_string() + url.domain().unwrap_or("UnknownDomain") + url.path());
+    let mut url_path = PathBuf::from("./downloaded/".to_string() + url.domain().unwrap_or("UnknownDomain") + url.path());
 
-    // if it's a file
+    if let Ok(header) = content_type.to_str() {
-    let (basepath, filename) = if url_path.extension().filter(valid_file_extension).is_some() {
+        // text/html; charset=UTF-8; option=value
-        // get everything up till the file
+        let ttype = if let Some((t, _)) = header.split_once(';') {
-        let basepath = url_path.ancestors().skip(1).take(1).collect::<PathBuf>();
+            t
-        // get the file name
+        } else {
-        let filename = url_path.file_name().expect("This should exist").to_string_lossy();
+            header
-        trace!("Save path: {:?} and base path: {:?}", &url_path, &basepath);
+        };
-        (basepath, filename.to_string())
+
+        if let Some((ttype, subtype)) = ttype.split_once('/') {
+            trace!("Found Content-Type to be: {ttype}/{subtype} for {}", url.to_string());
+            // If the Content-Type header is "*/html" (most likely "text/html") and the path's
+            // extension is anything but html:
+            if subtype=="html" && !url_path.extension().is_some_and(|f| f=="html" || f=="htm" ) {
+                // time to slap a index.html to the end of that path there!
+                url_path = url_path.join("index.html");
+            }
+        }
     } else {
-        (url_path.clone(), "index.html".into())
+        warn!("Header: {:?} couldn't be parsed into a string!", content_type);
-    };
+    }
+    trace!("Final path for {} is: {:?}", url, url_path);
 
-    let mut path = PathBuf::new();
+    url_path
-    path = path.join(basepath);
-    path = path.join(filename);
-
-    path
 }
 
 pub async fn init(filename: &PathBuf) -> Option<fs::File> {
@@ -50,29 +57,10 @@ pub async fn init(filename: &PathBuf) -> Option<fs::File> {
                     error!("Couldn't get file's parents: {:?}", &filename);
                 }
             } else {
-                error!("File creation: {err} {:?}", filename);
+                error!("File open error: {err} {:?}", filename);
             }
             // we don't care about other errors, we can't/shouldn't fix them
             None
         }
     }
 }
-
-fn valid_file_extension(take: &&OsStr) -> bool {
-    let los = take.to_string_lossy();
-    let all = los.split('.');
-    match all.last() {
-        Some(s) => {
-            // FIXME it's worth noting that the dumb tlds like .zip are in here,
-            // which could cause problems
-            let all_domains = include_str!("tlds-alpha-by-domain.txt");
-
-            // check if it is a domain
-            match all_domains.lines().map(str::to_lowercase).find(|x| x==s.to_lowercase().as_str()) {
-                Some(_) => false,
-                None => true
-            }
-        },
-        None => false,
-    }
-}

src/main.rs

@@ -1,5 +1,4 @@
 #![feature(ip_from)]
-#![feature(async_closure)]
 #![warn(clippy::expect_used)]
 #![deny(clippy::unwrap_used)]
 
@@ -19,8 +18,8 @@ use metrics::{counter, gauge};
 use metrics_exporter_prometheus::PrometheusBuilder;
 use serde::Deserialize;
 use surrealdb::{engine::remote::ws::Client, Surreal};
-use tokio::{io::AsyncWriteExt, task::JoinSet};
+use tokio::{io::{AsyncWriteExt, BufWriter}, task::JoinSet};
-use tracing::{debug, error, info, instrument, level_filters::LevelFilter, trace, trace_span};
+use tracing::{debug, debug_span, error, info, instrument, level_filters::LevelFilter, trace, trace_span, warn};
 use tracing_subscriber::{fmt, layer::SubscriberExt, EnvFilter, Layer, Registry};
 
 mod db;
@@ -173,24 +172,45 @@ async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Clien
 
     // Send the http request (get)
     if let Ok(response) = request_builder.send().await {
-        // Get body from response
+        let headers = response.headers();
+        let code = response.status();
 
-        let path = filesystem::as_path(&site.site);
+        #[allow(non_snake_case)]
+        let CT = headers.get("Content-Type");
+        let ct = headers.get("content-type");
+
+        let ct = match (CT,ct) {
+            (None, None) => {
+                warn!("Server did not respond with Content-Type header. Url: {} Headers: ({:?})", site.site.to_string(), headers);
+                return
+            },
+            (None, Some(a)) => a,
+            (Some(a), None) => a,
+            (Some(a), Some(_)) => a,
+        };
+
+        // create filepath (handles / -> /index.html)
+        let path = filesystem::as_path(&site.site, ct);
 
         // make sure that the file is good to go
-        if let Some(mut file) = filesystem::init(&path).await {
+        if let Some(file) = filesystem::init(&path).await {
-            let should_parse = path.to_string_lossy().ends_with(".html");
+            // Get body from response
-            let mut buf: Vec<u8> = Vec::new();
-
             // stream the response onto the disk
             let mut stream = response.bytes_stream();
+
+            let should_parse = path.to_string_lossy().ends_with(".html");
+            let mut writer = BufWriter::new(file);
+            let mut buf: Vec<u8> = Vec::new();
+
+            // Write file to disk
+            info!("Writing at: {:?}", path);
             while let Some(data) = stream.next().await {
                 match data {
                     Ok(data) => {
-                        debug!("Writing at: {:?}", path);
+                        let _ = writer.write_all(&data).await;
-                        let _ = file.write_all(&data).await;
                         // If we are going to parse this file later, we will save it
                         // into memory as well as the disk.
+                        // We do this because the data here might be incomplete
                         if should_parse {
                             data.iter().for_each(|f| buf.push(*f));
                         }
@@ -200,8 +220,14 @@ async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Clien
                     },
                 }
             }
+            let _ = writer.flush();
 
+
+            // (If needed) Parse the file
             if should_parse {
+                let span = debug_span!("Should Parse");
+                let enter = span.enter();
+
                 // Parse document and get relationships
                 let sites = parser::parse(&site, &buf).await;
                 // De-duplicate this list
@@ -215,6 +241,8 @@ async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Clien
                 trace!("Saved {diff} from being entered into the db by de-duping");
                 // Store all the other sites so that we can link to them.
                 let _ = Website::store_all(de_dupe_sites, &db).await;
+
+                drop(enter);
             }
 
             // METRICS
@@ -222,11 +250,14 @@ async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Clien
             counter!(GET_METRIC).increment(1);
 
             // update self in db
-            site.set_crawled();
+            site.crawled = true;
-            Website::store_all(vec![site], &db).await;
+            site.status_code = code.as_u16();
+            Website::store_all(vec![site.clone()], &db).await;
         } else {
             error!("File failed to cooperate: {:?}", path);
         }
+
+        trace!("Done processing: {}", &site.site);
     } else {
         error!("Failed to get: {}", &site.site);
     }
@@ -244,7 +275,7 @@ async fn get_uncrawled_links(
         count = config.batch_size;
     }
 
-    debug!("Getting uncrawled links");
+    debug!("Getting {} uncrawled links", count);
 
     let mut response = db
         .query("SELECT * FROM website WHERE crawled = false AND site ~ type::string($format) LIMIT $count;")

src/parser.rs

@@ -39,7 +39,7 @@ impl TokenSink for Website {
                                     if let Some(mut parsed) = url {
                                         parsed.set_query(None);
                                         parsed.set_fragment(None);
-                                        debug!("Final cleaned URL: `{}`", parsed.to_string());
+                                        trace!("Final cleaned URL: `{}`", parsed.to_string());
                                         let web = Website::new(&parsed.to_string(), false);
                                         links.push(web);
                                     }
@@ -127,16 +127,15 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
                         }
                         let url = origin.clone() + link;
 
-                        trace!("Built `{url}` from `{origin} + `{}`", link.to_string());
-
                         if let Ok(url) = Url::parse(&url) {
-                            trace!("Saved relative url `{}` AS: `{}`", link, url);
+                            trace!("Built `{url}` from `{origin} + `{}`", link.to_string());
                             Some(url)
                         } else {
                             error!(
-                                "Failed to reconstruct a url from relative url: `{}` on site: `{}`",
+                                "Failed to reconstruct a url from relative url: `{}` on site: `{}`. Failed url was: {}",
                                 link,
-                                parent.to_string()
+                                parent.to_string(),
+                                url
                             );
                             None
                         }