checkpoint

.gitignore

@@ -6,4 +6,3 @@ flamegraph.svg
 perf.data.old
 /docker/logs/*
 /downloaded
-/Crawler.toml

.vscode/launch.json

@@ -7,15 +7,18 @@
         {
             "type": "lldb",
             "request": "launch",
-            "name": "Debug executable 'internet_mapper'",
+            "name": "Debug executable 'surreal_spider'",
+            "env": {
+                "RUST_LOG": "surreal_spider=trace,reqwest=info",
+            },
             "cargo": {
                 "args": [
                     "build",
-                    "--bin=internet_mapper",
-                    "--package=internet_mapper"
+                    "--bin=surreal_spider",
+                    "--package=surreal_spider"
                 ],
                 "filter": {
-                    "name": "internet_mapper",
+                    "name": "surreal_spider",
                     "kind": "bin"
                 }
             },
@@ -25,16 +28,16 @@
         {
             "type": "lldb",
             "request": "launch",
-            "name": "Debug unit tests in executable 'internet_mapper'",
+            "name": "Debug unit tests in executable 'surreal_spider'",
             "cargo": {
                 "args": [
                     "test",
                     "--no-run",
-                    "--bin=internet_mapper",
-                    "--package=internet_mapper"
+                    "--bin=surreal_spider",
+                    "--package=surreal_spider"
                 ],
                 "filter": {
-                    "name": "internet_mapper",
+                    "name": "surreal_spider",
                     "kind": "bin"
                 }
             },

.vscode/settings.json

@@ -3,6 +3,6 @@
         "creds",
         "reqwest",
         "rustls",
-        "surql",
+        "surql"
     ]
 }

Cargo.toml

@@ -5,17 +5,14 @@ edition = "2021"
 
 [dependencies]
 base64 = "0.22.1"
-futures-util = "0.3.31"
 html5ever = "0.29"
 metrics = "0.24.1"
 metrics-exporter-prometheus = { version = "0.16.2", features=["http-listener"]}
-opentelemetry = "0.30.0"
-opentelemetry-otlp = { version = "0.30.0", features = ["metrics", "trace", "logs", "grpc-tonic"] }
-opentelemetry_sdk = "0.30.0"
-rand = "0.9.1"
-reqwest = { version = "0.12", features = ["gzip", "default", "rustls-tls", "stream"] }
+# minio = "0.1.0"
+minio = {git="https://github.com/minio/minio-rs.git", rev = "c28f576"}
+reqwest = { version = "0.12", features = ["gzip", "default", "rustls-tls"] }
+rusqlite = { version = "0.34.0", features = ["bundled"] }
 serde = { version = "1.0", features = ["derive"] }
-surrealdb = "2.2"
 tokio = { version="1.41.0", features = ["full"] }
 toml = "0.8.20"
 tracing = "0.1"

Crawler.toml

@@ -1,22 +1,10 @@
-# Visability config
-# Alloy (for Tempo)
-tracing_endpoint = "http://localhost:4317"
-# Prometheus
-metrics_endpoint = "http://localhost:9090/api/v1/otlp/v1/metrics"
-# Alloy (for Loki)
-log_file = "./docker/logs/tracing.log"
-
 # Surreal config
 surreal_url = "localhost:8000"
 surreal_username = "root"
 surreal_password = "root"
 surreal_ns = "test"
-surreal_db = "v1.21.1"
+surreal_db = "v1.19.2"
 
 # Crawler config
-crawl_filter = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI/Imagery/2023_NAIP/UTM_County_Mosaics/" 
-# crawl_filter = "https://oliveratkinson.net" 
-start_url = "https://ftpgeoinfo.msl.mt.gov/Data/Spatial/MSDI/Imagery/2023_NAIP/UTM_County_Mosaics/"
-# start_url = "https://oliveratkinson.net"
-budget = 100
-batch_size = 2
+crawl_filter = "en.wikipedia.com" 
+budget = 1000

README.md

@@ -2,43 +2,13 @@
 
 Crawls sites saving all the found links to a surrealdb database. It then proceeds to take batches of 100 uncrawled links untill the crawl budget is reached. It saves the data of each site in a minio database.
 
-## How to use
-
-1. Clone the repo and `cd` into it.
-2. Build the repo with `cargo build -r`
-3. Start the docker conatiners
-	1. cd into the docker folder `cd docker`
-	2. Bring up the docker containers `docker compose up -d`
-4. From the project's root, edit the `Crawler.toml` file to your liking.
-5. Run with `./target/release/internet_mapper`
-
-You can view stats of the project at `http://<your-ip>:3000/dashboards`
-
-```bash
-# Untested script but probably works
-git clone https://git.oliveratkinson.net/Oliver/internet_mapper.git
-cd internet_mapper
-
-cargo build -r
-
-cd docker
-docker compose up -d
-cd ..
-
-$EDITOR Crawler.toml
-
-./target/release/internet_mapper
-
-```
-
 ### TODO
 
-- [x] Domain filtering - prevent the crawler from going on alternate versions of wikipedia.
+- [ ] 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.
-- [x] Allow for storing asynchronously - dropping the "links to" logic fixes this need
-- [x] Control crawler via config file (no recompliation needed)
+- [ ] Allow for storing asynchronously
 
 3/17/25: Took >1hr to crawl 100 pages
 

docker/alloy.conf

@@ -12,25 +12,3 @@ loki.write "local_loki" {
         url = "http://loki:3100/loki/api/v1/push"
     }
 }
-
-otelcol.receiver.otlp "otlp_receiver" {
-  grpc {
-    endpoint = "0.0.0.0:4317"
-  }
-  http {
-    endpoint = "0.0.0.0:4318"
-  }
-
-  output {
-    traces = [otelcol.exporter.otlp.tempo.input,]
-  }
-}
-
-otelcol.exporter.otlp "tempo" {
-  client {
-    endpoint = "tempo:4317"
-    tls {
-			insecure = true
-		}
-  }
-}

docker/compose.yml

@@ -1,6 +1,4 @@
 services:
-
-  # Database
   surreal:
     image: surrealdb/surrealdb:latest-dev
     ports:
@@ -17,18 +15,6 @@ services:
       - root
       - rocksdb:/mydata/database.db
 
-  # Tracing
-  tempo:
-    image: grafana/tempo:latest
-    command: [ "-config.file=/etc/tempo.yaml" ]
-    volumes:
-      - ./tempo.yaml:/etc/tempo.yaml
-      - tempo_storage:/var/tempo
-    ports:
-      - 3200:3200 # self metrics for prometheus
-      - 4317:4317 # otlp grpc - (alloy) 
-
-  # Log scraper
   alloy:
     image: grafana/alloy:latest
     ports:
@@ -38,13 +24,9 @@ services:
       - ./logs/:/tmp/alloy-logs
       - ./alloy.conf:/etc/alloy/config.alloy
       - alloy_storage:/var/lib/alloy
-    command:
-      - run
-      - --server.http.listen-addr=0.0.0.0:12345
-      - --storage.path=/var/lib/alloy/data
-      - /etc/alloy/config.alloy
+    command:  run --server.http.listen-addr=0.0.0.0:12345 --storage.path=/var/lib/alloy/data /etc/alloy/config.alloy
 
-  # Log storage / analysis
+  #logs
   loki:
     image: grafana/loki:latest
     ports:
@@ -53,21 +35,16 @@ services:
     volumes:
       - ./loki.yaml:/etc/loki/local-config.yaml
 
-  # Metrics
+  # Metrics collector
   prometheus:
     image: prom/prometheus:latest
-    ports:
-      - 9090:9090
+    expose:
+      - 9090
     volumes:
       - ./prometheus.yaml:/etc/prometheus/prometheus.yml
       # persist data
-      # - prometheus_storage:/prometheus
-    command:
-      - --enable-feature=native-histograms
-      - --web.enable-remote-write-receiver
-      - --web.enable-lifecycle
-      - --web.enable-otlp-receiver
-      - --config.file=/etc/prometheus/prometheus.yml
+      - prometheus_storage:/prometheus
+    command: --web.enable-lifecycle --config.file=/etc/prometheus/prometheus.yml
 
   # Everything viewer
   grafana:
@@ -89,4 +66,4 @@ volumes:
   grafana_storage:
   alloy_storage:
   surrealdb_storage:
-  tempo_storage:
+  minio_storage:

docker/grafana.yaml

@@ -22,20 +22,3 @@ datasources:
   editable: false
   jsonData:
     httpMethod: GET
-- name: Tempo
-  type: tempo
-  access: proxy
-  orgId: 1
-  url: http://tempo:3200
-  basicAuth: false
-  isDefault: false
-  version: 1
-  editable: true
-  apiVersion: 1
-  uid: tempo
-  jsonData:
-    httpMethod: GET
-    serviceMap:
-      datasourceUid: prometheus
-    streamingEnabled:
-      search: true

docker/prometheus.yaml

@@ -1,15 +1,17 @@
 global:
-  scrape_interval: 60s 
+  scrape_interval: 5s 
   query_log_file: /etc/prometheus/query.log
 
 scrape_configs:
-  # Crawler configs get pushed with OTLP
-#  - job_name: 'loki'
-#    static_configs:
-#      - targets: ['loki:3100']
-#  - job_name: 'prometheus'
-#    static_configs:
-#      - targets: ['localhost:9090']
-  - job_name: 'tempo'
+  - job_name: crawler
     static_configs:
-      - targets: ['tempo:3200']
+    # change this your machine's ip, localhost won't work
+    # because localhost refers to the docker container.
+      - targets: ['172.20.239.48:2500']
+        #- targets: ['192.168.8.209:2500']
+  - job_name: loki
+    static_configs:
+      - targets: ['loki:3100']
+  - job_name: prometheus
+    static_configs:
+      - targets: ['localhost:9090']

docker/tempo.yaml

@@ -1,48 +0,0 @@
-stream_over_http_enabled: true
-server:
-  http_listen_port: 3200
-  log_level: info
-
-query_frontend:
-  search:
-    duration_slo: 5s
-    throughput_bytes_slo: 1.073741824e+09
-    metadata_slo:
-      duration_slo: 5s
-      throughput_bytes_slo: 1.073741824e+09
-  trace_by_id:
-    duration_slo: 5s
-
-distributor:
-  receivers:
-    otlp:
-      protocols:
-        grpc:
-          endpoint: "tempo:4317"
-
-metrics_generator:
-  registry:
-    external_labels:
-      source: tempo
-      cluster: docker-compose
-  storage:
-    path: /var/tempo/generator/wal
-    remote_write:
-      - url: http://prometheus:9090/api/v1/write
-        send_exemplars: true
-  traces_storage:
-    path: /var/tempo/generator/traces
-
-storage:
-  trace:
-    backend: local # backend configuration to use
-    wal:
-      path: /var/tempo/wal # where to store the wal locally
-    local:
-      path: /var/tempo/blocks
-
-overrides:
-  defaults:
-    metrics_generator:
-      processors: [service-graphs, span-metrics, local-blocks] # enables metrics generator
-      generate_native_histograms: both

src/db.rs

@@ -1,13 +1,7 @@
 use metrics::counter;
+use rusqlite::Connection;
+use std::fmt::Debug;
 use serde::{Deserialize, Serialize};
-use std::{fmt::Debug, time::Duration};
-use surrealdb::{
-    engine::remote::ws::{Client, Ws},
-    opt::auth::Root,
-    sql::Thing,
-    Surreal,
-};
-use tokio::time::sleep;
 use tracing::{error, instrument, trace};
 use url::Url;
 
@@ -17,23 +11,16 @@ const STORE: &str = "surql_store_calls";
 
 #[derive(Serialize, Deserialize, Clone, Eq, PartialEq, Hash)]
 pub struct Website {
-    pub id: Option<Thing>,
     /// The url that this data is found at
     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("host", &self.site.host())
-            .field("path", &self.site.path())
-            .field("status_code", &self.status_code)
-            .finish()
+        f.debug_struct("Website").field("site", &self.site).finish()
     }
 }
 
@@ -46,26 +33,32 @@ impl Website {
         };
         Self {
             crawled,
-            site,
-            status_code: 0,
-            id: None,
+            site
         }
     }
 
+    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))]
-    pub async fn store_all(all: Vec<Self>, db: &Surreal<Client>) -> Vec<Thing> {
+    pub async fn store_all(all: Vec<Self>, db: &Connection) {
         counter!(STORE).increment(1);
         let mut things = Vec::with_capacity(all.len());
 
+        rusqlite::ParamsFromIter;
+
+        db.execute("",
+            params![]
+        );
+
         match db
             .query(
                 "INSERT INTO website $array
                     ON DUPLICATE KEY UPDATE
                         accessed_at = time::now(),
-                        status_code = $input.status_code,
-                        processing = false,
                         crawled = crawled OR $input.crawled
                     RETURN VALUE id;
                  ",
@@ -85,72 +78,22 @@ impl Website {
     }
 }
 
-/// Returns uncrawled links
-#[instrument(skip(db, config))]
-pub async fn get_next(db: &Surreal<Client>, config: &Config) -> Option<Website> {
-    let mut res: Option<Website> = None;
-    let mut fails = 0;
-
-    while res == None {
-        let mut response = db
-            .query("fn::get_next($format)")
-            .bind(("format", config.crawl_filter.to_string()))
-            .await
-            .expect("Hard-coded query failed..?");
-
-        res = match response.take(0) {
-            Ok(ok) => ok,
-            Err(_err) => {
-                // basically just CSMA/CA
-                let delay = rand::random_range(10..500);
-                sleep(Duration::from_millis(delay)).await;
-                fails += 1;
-                // Don't get stuck here forever, failing...
-                // (most I've seen is 1)
-                if fails > 5 {
-                    error!("Max attempts to get_next() reached... ({fails})");
-                    return None
-                }
-                None
-            }
-        };
-    }
-
-    res
-}
-
 #[derive(Debug, Serialize)]
-#[allow(dead_code)]
 pub struct Email {
     pub email: String,
     pub on: String,
 }
 
-#[instrument(skip_all, name = "SurrealDB")]
-pub async fn connect(config: &Config) -> surrealdb::Result<Surreal<Client>> {
-    trace!("Establishing connection to surreal...");
-    // Connect to the server
-    let db = Surreal::new::<Ws>(&config.surreal_url).await?;
-
-    trace!("Logging in...");
-    // Signin as a namespace, database, or root user
-    db.signin(Root {
-        username: &config.surreal_username,
-        password: &config.surreal_password,
-    })
-    .await?;
-
-    // Select a specific namespace / database
-    db.use_ns(&config.surreal_ns)
-        .use_db(&config.surreal_db)
-        .await?;
-
-    let setup = include_bytes!("setup.surql");
-    let file = setup.iter().map(|c| *c as char).collect::<String>();
-
-    db.query(file)
-        .await
-        .expect("Failed to setup surreal tables.");
-
-    Ok(db)
+#[derive(Debug, Deserialize)]
+pub struct Record {
+    #[allow(dead_code)]
+    pub id: Thing,
 }
+
+#[instrument(skip_all, name = "sqlite_connect")]
+pub async fn connect(config: &Config) -> Result<Connection, rusqlite::Error> {
+    trace!("Establishing connection to sqlite...");
+    // Connect to the server
+    Connection::open("./squeelite.db")
+}
+

src/filesystem.rs

@@ -1,77 +1,70 @@
-use std::{io::ErrorKind, path::PathBuf};
+use std::{ffi::OsStr, path::PathBuf};
 
-use reqwest::header::HeaderValue;
 use tokio::fs;
-use tracing::{error, trace, warn};
+use tracing::{debug, error, instrument, trace, warn};
 use url::Url;
 
-pub fn as_path(url: &Url, content_type: &HeaderValue) -> PathBuf {
+#[instrument(skip(data))]
+pub async fn store(data: &str, url: &Url) {
     // extract data from url to save it accurately
-    let mut url_path = PathBuf::from("./downloaded/".to_string() + url.domain().unwrap_or("UnknownDomain") + url.path());
+    let url_path = PathBuf::from("./downloaded/".to_string() + url.domain().unwrap_or("UnknownDomain") + url.path());
 
-    if let Ok(header) = content_type.to_str() {
-        // text/html; charset=UTF-8; option=value
-        let ttype = if let Some((t, _)) = header.split_once(';') {
-            t
+    // if it's a file
+    let (basepath, filename) = if url_path.extension().filter(valid_file_extension).is_some() {
+        // get everything up till the file
+        let basepath = url_path.ancestors().skip(1).take(1).collect::<PathBuf>();
+        // get the file name
+        let filename = url_path.file_name().expect("This should exist").to_string_lossy();
+        trace!("Save path: {:?} and base path: {:?}", &url_path, &basepath);
+        (basepath, filename.to_string())
     } else {
-            header
+        (url_path.clone(), "index.html".into())
     };
 
-        if let Some((ttype, subtype)) = ttype.split_once('/') {
-            trace!(url = url.to_string(), main_type = ttype, sub_type = subtype, "Found Content-Type to be: {ttype}/{subtype}");
-            // 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 {
-        warn!("Header: {:?} couldn't be parsed into a string!", content_type);
-    }
-    trace!(url = url.to_string(), path = &*url_path.to_string_lossy(), "Converted URL into path");
+    debug!("Writing at: {:?} {:?}", basepath, filename);
 
-    url_path
-}
-
-pub async fn init(filename: &PathBuf) -> Option<fs::File> {
-    let file = async || tokio::fs::OpenOptions::new()
-        .write(true)
-        .append(true)
-        .create(true)
-        .open(&filename).await;
-
-    match file().await {
-        Ok(ok) => {
-            trace!("Initialized file {}", filename.to_str().unwrap_or("N/A"));
-            Some(ok)
-        },
-        Err(err) => {
-            // the file/folder isn't found
-            if err.kind() == ErrorKind::NotFound {
-                if let Some(parent ) = &filename.parent() {
     // create the folders
-                    if let Err(err) = fs::create_dir_all(&parent).await {
-                        error!("Dir creation: {err} {:?}", filename);
-                    } else if let Ok(ok) = file().await {
-                        return Some(ok);
-                    }
+    if let Err(err) = fs::create_dir_all(&basepath).await {
+        error!("Dir creation: {err} {:?}", basepath);
     } else {
-                    error!("Couldn't get file's parents: {:?}", &filename);
-                }
-            } else if err.kind() == ErrorKind::NotADirectory {
-                // Example:
-                // 1. example.com/user
-                // 2. example.com/user/post
-                // If file 1 exists it will prevent file 2 from existing
-                // FIXME
-
-                error!("One of the parent directories is actually a file...")
-            } else {
-                error!("File open error: {err} {:?}", filename);
-            }
-            // we don't care about other errors, we can't/shouldn't fix them
-            None
+        // FIXME I don't think this handles index.html files well...
+        // TODO this should probably append .html to non-described files
+        // create the file if that was successful
+        if let Err(err) = fs::write(&basepath.join(filename), data).await {
+            error!("File creation: {err} {:?}", url_path);
         }
     }
 }
+
+fn valid_file_extension(take: &&OsStr) -> bool {
+    let los = take.to_string_lossy();
+    let all = los.split('.');
+    match all.last() {
+        Some(s) => {
+            match s.to_lowercase().as_str() {
+                "html" => true,
+                "css" => true,
+                "js" => true,
+                "ts" => true,
+                "otf" => true, // font
+
+                "png" => true,
+                "svg" => true,
+                "jpg" => true,
+                "jpeg" => true,
+                "mp4" => true,
+                "mp3" => true,
+                "webp" => true,
+
+                "pdf" => true,
+                "json" => true,
+                "xml" => true,
+                _ => {
+                    warn!("Might be forgetting a file extension: {s}");
+                    false
+                }
+            }
+        },
+        None => false,
+    }
+}

src/main.rs

@@ -1,80 +1,31 @@
 #![feature(ip_from)]
-#![feature(path_add_extension)]
-#![deny(clippy::unwrap_used)]
 
 extern crate html5ever;
 
 use std::{
-    collections::HashSet, fs::File, io::Read, sync::{Arc, LazyLock}
+    collections::HashSet, fs::File, io::Read, net::{IpAddr, Ipv4Addr}
 };
 
-use futures_util::StreamExt;
-use opentelemetry::{global::{self}, metrics::{Counter, Meter, UpDownCounter}};
-use opentelemetry_otlp::{Protocol, WithExportConfig};
 use db::{connect, Website};
-use opentelemetry_sdk::{metrics::SdkMeterProvider, trace::SdkTracerProvider};
+use metrics::{counter, gauge};
+use metrics_exporter_prometheus::PrometheusBuilder;
 use serde::Deserialize;
 use surrealdb::{engine::remote::ws::Client, Surreal};
-use tokio::{io::{AsyncWriteExt, BufWriter}, sync::RwLock, task::JoinSet};
-use tracing::{debug, error, info, instrument, level_filters::LevelFilter, trace, warn};
+use tokio::task::JoinSet;
+use tracing::{debug, error, info, instrument, level_filters::LevelFilter, trace, trace_span};
 use tracing_subscriber::{fmt, layer::SubscriberExt, EnvFilter, Layer, Registry};
 
-use crate::db::get_next;
-
 mod db;
-mod filesystem;
 mod parser;
+mod filesystem;
 
-static METER: LazyLock<Meter> = LazyLock::new(|| global::meter("Internet_Mapper"));
-static BEING_PROCESSED: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
-    METER
-        .i64_up_down_counter("crawler_pages_being_processed")
-        .build()
-);
-static BEING_PARSED: LazyLock<UpDownCounter<i64>> = LazyLock::new(||
-    METER
-        .i64_up_down_counter("crawler_pages_being_parsed")
-        .build()
-);
-static BEING_STREAMED: LazyLock<UpDownCounter<i64>> = LazyLock::new(|| 
-    METER
-        .i64_up_down_counter("crawler_pages_being_streamed")
-        .build()
-);
-static GET_IN_FLIGHT: LazyLock<UpDownCounter<i64>> = LazyLock::new(|| 
-    METER
-        .i64_up_down_counter("crawler_gets_in_flight")
-        .build()
-);
-static TOTAL_BYTES_DOWN: LazyLock<Counter<u64>> = LazyLock::new(|| 
-    METER
-        .u64_counter("crawler_total_bytes_down")
-        .build()
-);
-static SITES_CRAWLED: LazyLock<Counter<u64>> = LazyLock::new(|| 
-    METER
-        .u64_counter("crawler_total_sites_crawled")
-        .build()
-);
-
-static CONFIG: LazyLock<Config> = LazyLock::new(|| {
-    let mut file = File::open("./Crawler.toml").expect("Failed to read Crawler.toml");
-    let mut buf = String::new();
-    let _ = file.read_to_string(&mut buf);
-
-    let config: Config = toml::from_str(&buf).expect("Failed to parse Crawler.toml");
-    config
-});
-
-// FIXME Traces aren't working on multiple threads, they block
-// static TRACER: LazyLock<BoxedTracer> = LazyLock::new(|| global::tracer("Internet_Mapper"));
+const GET_METRIC: &str = "total_gets";
+const GET_IN_FLIGHT: &str = "gets_in_flight";
+const SITES_CRAWLED: &str = "pages_crawled";
+const BEING_PROCESSED: &str = "pages_being_processed";
 
 #[derive(Deserialize)]
 struct Config {
-    tracing_endpoint: String,
-    metrics_endpoint: String,
-    log_file: String,
-
     surreal_ns: String,
     surreal_db: String,
     surreal_url: String,
@@ -82,230 +33,15 @@ struct Config {
     surreal_password: String,
 
     crawl_filter: String,
-    start_url: String,
     budget: usize,
-    batch_size: usize,
 }
 
 #[tokio::main]
 async fn main() {
-    println!("Logs and metrics are provided to the Grafana dashboard");
-
-    // Start TRACE / LOGGING / METRICS
-    load_logging(&CONFIG); // this seems to be working ok
-    global::set_tracer_provider(load_tracing(&CONFIG));
-    global::set_meter_provider(load_metrics(&CONFIG));
-
-    // When getting uncrawled pages, name must contain this variable. "" will effectively get ignored.
-    // let crawl_filter = "en.wikipedia.org/";
-    // let budget = 50;
-    let crawled = Arc::new(RwLock::new(0));
-
-    let starting_url = &CONFIG.start_url;
-
-    let db = connect(&CONFIG)
-        .await
-        .expect("Failed to connect to surreal, aborting.");
-
-    let reqwest = reqwest::Client::builder()
-        // .use_rustls_tls()
-        .gzip(true)
-        .build()
-        .expect("Failed to build reqwest client.");
-
-    // Kick off the whole machine - This Website object doesn't matter, it's just to allow for
-    // get() to work.
-    // let mut span = TRACER.start("Pre-Loop");
-    let site = Website::new(starting_url, false);
-    process(site, db.clone(), reqwest.clone()).await;
-    // span.end();
-
-    // let mut main_loop_span= TRACER.start("Main-Loop");
-    let mut futures = JoinSet::new();
-    for _ in 0..CONFIG.batch_size {
-        futures.spawn(process_single_thread(&CONFIG, db.clone(), reqwest.clone(), crawled.clone()));
-    }
-    futures.join_all().await;
-    // main_loop_span.end();
-
-    info!("Done");
-}
-
-async fn process_single_thread(config: &Config, db: Surreal<Client>, reqwest: reqwest::Client, crawled: Arc<RwLock<usize>>) {
-    while *(crawled.read().await) < config.budget {
-        let uncrawled = get_next(&db.clone(), &config).await;
-        match uncrawled {
-            Some(site) => {
-                process(site, db.clone(), reqwest.clone()).await;
-                SITES_CRAWLED.add(1, &[]);
-                // Somehow this write doesn't hang on the while's read?
-                let mut c = crawled.write().await;
-                *c += 1;
-            },
-            None => {
-                warn!("fn::get_next() returned None");
-                return;
-            },
-        }
-    }
-}
-
-#[instrument(skip(db, reqwest))]
-/// Downloads and crawls and stores a webpage.
-/// It is acceptable to clone `db`, `reqwest`, and `s3` because they all use `Arc`s internally. - Noted by Oliver
-async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Client) {
-
-    // METRICS
-    debug!(url = &site.site.as_str(), "Process: {}", &site.site);
-    BEING_PROCESSED.add(1, &[]);
-    // let mut process_span = TRACER.start("Process");
-
-    // Build the request
-    let request_builder = reqwest.get(site.site.to_string());
-    
-    // Send the http request (get)
-    GET_IN_FLIGHT.add(1, &[]);
-    if let Ok(response) = request_builder.send().await {
-        GET_IN_FLIGHT.add(-1, &[]);
-
-        let headers = response.headers();
-        let code = response.status();
-        if code != 200 {
-            warn!("{code} for {}", site.site.as_str());
-        }
-
-        #[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);
-        let mut tmp_path= path.clone();
-        if !(tmp_path.add_extension("crawl_temp")) {
-            warn!("Failed to add extension to file");
-            // fallback ig
-            tmp_path = tmp_path.with_extension("crawl_temp");
-        }
-
-        // make sure that the file is good to go
-        if let Some(file) = filesystem::init(&tmp_path).await {
-            // Get body from response
-            // 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
-            trace!("Writing at: {:?}", tmp_path);
-            BEING_STREAMED.add(1, &[]);
-            // let mut stream_span = TRACER.start("Stream");
-            while let Some(data) = stream.next().await {
-                match data {
-                    Ok(data) => {
-                        TOTAL_BYTES_DOWN.add(data.len() as u64, &[]);
-                        let _ = writer.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));
-                        }
-                    },
-                    Err(err) => {
-                        error!("{err}")
-                    },
-                }
-            }
-            let _ = writer.flush().await;
-            // rename the temp file into the real file name
-            if let Err(err) = tokio::fs::rename(&tmp_path, &path).await {
-                error!(
-                    from = &*tmp_path.to_string_lossy(),
-                    to = &*path.to_string_lossy(),
-                    "Error renaming file: {}",
-                    err
-                );
-            }
-
-            // stream_span.end();
-            BEING_STREAMED.add(-1, &[]);
-
-            // (If needed) Parse the file
-            if should_parse {
-                BEING_PARSED.add(1, &[]);
-                // let mut parsing_span = TRACER.start("Parsing");
-
-                // Parse document and get relationships
-                let sites = parser::parse(&site, &buf).await;
-                // De-duplicate this list
-                let prev_len = sites.len();
-                let set = sites.into_iter().fold(HashSet::new(), |mut set, item| {
-                    set.insert(item);
-                    set
-                });
-                let de_dupe_sites: Vec<Website> = set.into_iter().collect();
-                let diff = prev_len - de_dupe_sites.len();
-                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;
-
-                // parsing_span.end();
-                BEING_PARSED.add(-1, &[]);
-            } else {
-                trace!(url = site.site.as_str(), "Parse = False");
-            }
-
-            // update self in db
-            site.crawled = true;
-            site.status_code = code.as_u16();
-            Website::store_all(vec![site.clone()], &db).await;
-        }
-    } else {
-        error!(url = site.site.as_str(), "Failed to get: {}", &site.site);
-    }
-    
-    // process_span.end();
-    BEING_PROCESSED.add(-1, &[]);
-}
-
-fn load_tracing(config: &Config) -> SdkTracerProvider {
-    // Send spans to Alloy (which will send them to Tempo)
-    let otlp_span = opentelemetry_otlp::SpanExporter::builder()
-        .with_tonic()
-        .with_endpoint(config.tracing_endpoint.clone())
-        .build()
-        .unwrap();
-    let tracer_provider = opentelemetry_sdk::trace::SdkTracerProvider::builder()
-        .with_simple_exporter(otlp_span)
-        .build();
-    tracer_provider
-}
-
-fn load_logging(config: &Config) {
-//    let otlp_log = opentelemetry_otlp::LogExporter::builder()
-//        .with_tonic()
-//        .with_endpoint(endpoint)
-//        .build()
-//        .unwrap();
-    // let logger_provider = opentelemetry_sdk::logs::SdkLoggerProvider::builder()
-    //    .with_simple_exporter(otlp_log)
-    //    .build();
     let writer = std::fs::OpenOptions::new()
         .append(true)
         .create(true)
-        .open(config.log_file.clone())
+        .open("./docker/logs/tracing.log")
         .expect("Couldn't make log file!");
 
     let filter = EnvFilter::builder()
@@ -323,20 +59,163 @@ fn load_logging(config: &Config) {
     );
 
     tracing::subscriber::set_global_default(registry).expect("Failed to set default subscriber");
-}
 
-fn load_metrics(config: &Config) -> SdkMeterProvider {
-    //  Send metrics to Prometheus
-    let otlp_metrics = opentelemetry_otlp::MetricExporter::builder()
-        .with_http()
-        .with_protocol(Protocol::HttpBinary)
-        .with_endpoint(config.metrics_endpoint.clone())
+    let builder = PrometheusBuilder::new();
+    builder
+        .with_http_listener(std::net::SocketAddr::new(
+            IpAddr::V4(Ipv4Addr::from_octets([0, 0, 0, 0])),
+            2500,
+        ))
+        .install()
+        .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;
+    let mut crawled = 0;
+
+    let mut file = File::open("./Crawler.toml").expect("Failed to read Crawler.toml");
+    let mut buf = String::new();
+    let _ = file.read_to_string(&mut buf);
+
+    let config: Config = toml::from_str(&buf).expect("Failed to parse Crawler.toml");
+
+    let db = connect(&config)
+        .await
+        .expect("Failed to connect to surreal, aborting.");
+
+    let reqwest = reqwest::Client::builder()
+        // .use_rustls_tls()
+        .gzip(true)
         .build()
-        .unwrap();
-    let metrics_provider = opentelemetry_sdk::metrics::SdkMeterProvider::builder()
-        .with_periodic_exporter(otlp_metrics) // default delay is 60s, turn down to like 15
-        .build();
-    metrics_provider
+        .expect("Failed to build reqwest client.");
+
+    // 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 site = Website::new(starting_url, false);
+    process(site, db.clone(), reqwest.clone()).await;
+
+    drop(pre_loop_span);
+
+    let span = trace_span!("Loop");
+    let span = span.enter();
+    while crawled < config.budget {
+        let get_num = if config.budget - crawled < 100 {
+            config.budget - crawled
+        } else {
+            100
+        };
+
+        let uncrawled = get_uncrawled_links(&db, get_num, config.crawl_filter.clone()).await;
+        if uncrawled.is_empty() {
+            info!("Had more budget but finished crawling everything.");
+            return;
+        }
+
+        {
+            let mut futures = JoinSet::new();
+            for site in uncrawled {
+                gauge!(BEING_PROCESSED).increment(1);
+                futures.spawn(process(site, db.clone(), reqwest.clone()));
+                // let percent = format!("{:.2}%", (crawled as f32 / budget as f32) * 100f32);
+                // info!("Crawled {crawled} out of {budget} pages. ({percent})");
+            }
+
+            let c = counter!(SITES_CRAWLED);
+            // As futures complete runs code in while block
+            while futures.join_next().await.is_some() {
+                c.increment(1);
+                gauge!(BEING_PROCESSED).decrement(1);
+                crawled += 1;
+            }
+        }
+    }
+    drop(span);
+
+    info!("Done");
 }
 
+#[instrument(skip(db, reqwest))]
+/// Downloads and crawls and stores a webpage.
+/// It is acceptable to clone `db`, `reqwest`, and `s3` because they all use `Arc`s internally. - Noted by Oliver
+async fn process(mut site: Website, db: Surreal<Client>, reqwest: reqwest::Client) {
+    
+    // METRICS
+    trace!("Process: {}", &site.site);
+    // Build the request
+    let request_builder = reqwest.get(site.site.to_string());
+
+    // METRICS
+    let g = gauge!(GET_IN_FLIGHT);
+    g.increment(1);
+
+    // Send the http request (get)
+    if let Ok(response) = request_builder.send().await {
+
+        // METRICS
+        g.decrement(1);
+        counter!(GET_METRIC).increment(1);
+
+        // Get body from response
+        let data = response
+            .text()
+            .await
+            .expect("Failed to read http response's body!");
+
+        // Store document
+        filesystem::store(&data, &site.site).await;
+
+        // Parse document and get relationships
+        let sites = parser::parse(&site, &data).await;
+        
+        // update self in db
+        site.set_crawled();
+        Website::store_all(vec![site], &db).await;
+
+        // De-duplicate this list
+        let prev_len = sites.len();
+        let set = sites.into_iter().fold(HashSet::new(), |mut set,item| {
+            set.insert(item);
+            set
+        });
+        let de_dupe_sites: Vec<Website> = set.into_iter().collect();
+        let diff = prev_len - de_dupe_sites.len();
+        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;
+
+    } else {
+        error!("Failed to get: {}", &site.site);
+    }
+}
+
+/// Returns uncrawled links
+#[instrument(skip(db))]
+async fn get_uncrawled_links(
+    db: &Surreal<Client>,
+    mut count: usize,
+    filter: String,
+) -> Vec<Website> {
+    if count > 100 {
+        count = 100
+    }
+    debug!("Getting uncrawled links");
+
+    let mut response = db
+        .query("SELECT * FROM website WHERE crawled = false AND site ~ type::string($format) LIMIT $count;")
+        .bind(("format", filter))
+        .bind(("count", count))
+        .await
+        .expect("Hard-coded query failed..?");
+    response
+        .take(0)
+        .expect("Returned websites couldn't be parsed")
+}
 

src/parser.rs

@@ -1,10 +1,11 @@
 use std::default::Default;
+use std::str::FromStr;
 
 use html5ever::tokenizer::{BufferQueue, TokenizerResult};
 use html5ever::tokenizer::{StartTag, TagToken};
 use html5ever::tokenizer::{Token, TokenSink, TokenSinkResult, Tokenizer, TokenizerOpts};
 use html5ever::{local_name, tendril::*};
-use tracing::{error, instrument, trace, warn};
+use tracing::{debug, error, instrument, trace, warn};
 use url::Url;
 
 use crate::db::Website;
@@ -12,6 +13,7 @@ use crate::db::Website;
 impl TokenSink for Website {
     type Handle = Vec<Website>;
 
+    #[instrument(skip(token, _line_number))]
     fn process_token(&self, token: Token, _line_number: u64) -> TokenSinkResult<Self::Handle> {
         match token {
             TagToken(tag) => {
@@ -32,13 +34,13 @@ impl TokenSink for Website {
                                 let attr_name = attr.name.local.to_string();
                                 if attr_name == "src" || attr_name == "href" || attr_name == "data"
                                 {
-                                    trace!(url = self.site.as_str(),"Found `{}` in html `{}` tag", &attr.value, tag.name);
+                                    trace!("Found `{}` in html `{}` tag", &attr.value, tag.name);
                                     let url = try_get_url(&self.site, &attr.value);
 
                                     if let Some(mut parsed) = url {
                                         parsed.set_query(None);
                                         parsed.set_fragment(None);
-                                        trace!(url = self.site.as_str(), "Final cleaned URL: `{}`", parsed.to_string());
+                                        debug!("Final cleaned URL: `{}`", parsed.to_string());
                                         let web = Website::new(&parsed.to_string(), false);
                                         links.push(web);
                                     }
@@ -59,16 +61,14 @@ impl TokenSink for Website {
     }
 }
 
-#[instrument(skip(data))]
+#[instrument(skip_all)]
 /// Parses the passed site and returns all the sites it links to.
-pub async fn parse(site: &Website, data: &[u8]) -> Vec<Website> {
-    trace!(url = site.site.as_str(), "Parsing {}", site.site.to_string());
+pub async fn parse(site: &Website, data: &str) -> Vec<Website> {
     // prep work
     let mut other_sites: Vec<Website> = Vec::new();
 
     // change data into something that can be tokenized
-    let s: Result<Tendril<fmt::UTF8>, ()> = Tendril::try_from_byte_slice(data);
-    if let Ok(chunk) = s {
+    let chunk = Tendril::from_str(data).expect("Failed to parse string into Tendril!");
     // create buffer of tokens and push our input into it
     let token_buffer = BufferQueue::default();
     token_buffer.push_back(
@@ -86,9 +86,6 @@ pub async fn parse(site: &Website, data: &[u8]) -> Vec<Website> {
     }
     assert!(token_buffer.is_empty());
     tokenizer.end();
-    } else {
-        warn!(url = site.site.as_str(), "Tendril failed to parse on: {}", site.site.to_string());
-    }
 
     other_sites
 }
@@ -99,7 +96,7 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
         Ok(ok) => Some(ok),
         Err(e) => {
             if link.starts_with('#') {
-                trace!(url = parent.as_str(), "Rejecting # url");
+                trace!("Rejecting # url");
                 None
             } else if link.starts_with("//") {
                 // if a url starts with "//" is assumed that it will adopt
@@ -107,34 +104,32 @@ fn try_get_url(parent: &Url, link: &str) -> Option<Url> {
                 // https://stackoverflow.com/questions/9646407/two-forward-slashes-in-a-url-src-href-attribute
                 let scheme = parent.scheme();
 
-                match Url::parse(&format!("{scheme}://{link}")) {
+                match Url::parse(&format!("{scheme}://{}", link)) {
                     Ok(url) => Some(url),
                     Err(err) => {
-                        error!("Failed parsing relative scheme url: {}", err);
+                        error!("Failed parsing realative scheme url: {}", err);
                         None
                     }
                 }
             } else {
-                // # This is some sort of relative url, gonna try patching it up into an absolute
+                // # This is some sort of realative url, gonna try patching it up into an absolute
                 // url
                 match e {
                     url::ParseError::RelativeUrlWithoutBase => {
                         // Is: scheme://host:port
-                        let mut origin = parent.origin().ascii_serialization();
-                        if !origin.ends_with('/') && !link.starts_with('/') {
-                            origin += "/";
-                        }
+                        let origin = parent.origin().ascii_serialization();
                         let url = origin.clone() + link;
 
+                        trace!("Built `{url}` from `{origin} + {}`", link.to_string());
+
                         if let Ok(url) = Url::parse(&url) {
-                            trace!(url = parent.as_str(), "Built `{url}` from `{origin} + `{}`", link.to_string());
+                            trace!("Saved relative url `{}` AS: `{}`", link, url);
                             Some(url)
                         } else {
                             error!(
-                                "Failed to reconstruct a url from relative url: `{}` on site: `{}`. Failed url was: {}",
+                                "Failed to reconstruct a url from relative url: `{}` on site: `{}`",
                                 link,
-                                parent.to_string(),
-                                url
+                                parent.to_string()
                             );
                             None
                         }

src/setup.surql

@@ -4,13 +4,6 @@ DEFINE FIELD IF NOT EXISTS site ON TABLE website TYPE string;
 DEFINE INDEX IF NOT EXISTS idx ON TABLE website COLUMNS site UNIQUE;
 
 DEFINE FIELD IF NOT EXISTS crawled ON TABLE website TYPE bool;
-DEFINE FIELD IF NOT EXISTS processing ON TABLE website TYPE bool DEFAULT false;
 
 DEFINE FIELD IF NOT EXISTS accessed_at ON TABLE website VALUE time::now();
 DEFINE FIELD IF NOT EXISTS first_accessed_at ON TABLE website VALUE time::now();
-
-DEFINE FUNCTION OVERWRITE fn::get_next($filter: string) {
-    LET $site = SELECT * FROM ONLY website WHERE crawled = false AND processing = false AND site ~ type::string($filter) LIMIT 1;
-    UPDATE $site.id SET processing = true;
-    RETURN $site
-};