don't immediately run fetcher a second time on startup

.gitignore

@@ -1,2 +1,3 @@
 /target
-data.pdf
+/data.pdf
+/result

Cargo.lock

@@ -288,6 +288,17 @@ version = "0.10.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "619743e34b5ba4e9703bba34deac3427c72507c7159f5fd030aea8cac0cfe341"
 
+[[package]]
+name = "atty"
+version = "0.2.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d9b39be18770d11421cdb1b9947a45dd3f37e93092cbf377614828a319d5fee8"
+dependencies = [
+ "hermit-abi",
+ "libc",
+ "winapi",
+]
+
 [[package]]
 name = "autocfg"
 version = "1.1.0"
@@ -425,6 +436,17 @@ dependencies = [
  "generic-array",
 ]
 
+[[package]]
+name = "colored"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b3616f750b84d8f0de8a58bda93e08e2a81ad3f523089b05f1dffecab48c6cbd"
+dependencies = [
+ "atty",
+ "lazy_static",
+ "winapi",
+]
+
 [[package]]
 name = "convert_case"
 version = "0.4.0"
@@ -1389,10 +1411,12 @@ dependencies = [
  "futures",
  "json",
  "lazy_static",
+ "log",
  "pdf",
  "regex",
  "reqwest",
  "scraper",
+ "simple_logger",
  "thiserror",
  "time 0.3.9",
  "tokio",
@@ -1764,6 +1788,19 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "simple_logger"
+version = "2.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c75a9723083573ace81ad0cdfc50b858aa3c366c48636edb4109d73122a0c0ea"
+dependencies = [
+ "atty",
+ "colored",
+ "log",
+ "time 0.3.9",
+ "winapi",
+]
+
 [[package]]
 name = "siphasher"
 version = "0.3.10"

Cargo.toml

@@ -14,10 +14,12 @@ bytes = "1.1"
 futures = "0.3"
 json = "0.12.4"
 lazy_static = "1.4"
+log = "0.4.16"
 pdf = "0.7.2"
 regex = "1.5.5"
 reqwest = { version = "0.11", features = ["rustls-tls", "stream"], default-features = false }
 scraper = "0.12"
+simple_logger = { version = "2.1.0", features = ["stderr"] }
 thiserror = "1"
 time = { version = "0.3.9", features = ["formatting", "macros", "parsing"] }
 tokio = { version = "1", features = ["full"] }

src/fetch.rs

@@ -2,6 +2,7 @@ use std::time::{Duration, Instant, SystemTime};
 
 use futures::{sink::SinkExt, TryStreamExt};
 use lazy_static::lazy_static;
+use log::info;
 use reqwest::Url;
 use scraper::Selector;
 use time::PrimitiveDateTime;
@@ -50,6 +51,8 @@ impl PdfFetcher {
     }
 
     pub async fn fetch(&mut self) -> Result<pdf::file::File<Vec<u8>>, Error> {
+        info!("Fetching data PDF");
+
         let cache_modtime = match tokio::fs::File::open(CACHED_PDF_PATH).await {
             Ok(file) => Some(
                 file.metadata()
@@ -65,6 +68,7 @@ impl PdfFetcher {
         if let Some(instant) = self.last_checked {
             if now - instant < *MIN_CHECK_INTERVAL {
                 return if cache_modtime.is_some() {
+                    info!("Already checked origin recently, not rechecking");
                     self.cached_pdf()
                 } else {
                     Err(Error::TooSoon)
@@ -111,6 +115,8 @@ impl PdfFetcher {
             .map_or(true, |(cache, origin)| origin > cache);
 
         if outdated {
+            info!("Cached PDF is outdated, downloading latest version");
+
             let mut pdf_stream = self
                 .client
                 .get(pdf_url)
@@ -124,6 +130,8 @@ impl PdfFetcher {
             sink.send_all(&mut pdf_stream).await?;
             <dyn futures::Sink<bytes::Bytes, Error = std::io::Error> + Unpin>::close(&mut sink)
                 .await?;
+        } else {
+            info!("Cached PDF is already up to date");
         }
         self.last_checked = Some(now);
         self.cached_pdf()

src/main.rs

@@ -1,6 +1,9 @@
 use std::{sync::Arc, time::Duration};
 
-use actix_web::{get, http::header::ContentType, web, App, HttpResponse, HttpServer, Responder};
+use actix_web::{
+    get, http::header::ContentType, middleware::Logger, web, App, HttpResponse, HttpServer,
+    Responder,
+};
 use lazy_static::lazy_static;
 use tokio::sync::RwLock;
 
@@ -47,10 +50,10 @@ async fn start_updater() -> Result<web::Data<AppState>, Error> {
     std::thread::spawn(move || {
         actix_web::rt::System::new().block_on(async {
             loop {
+                actix_web::rt::time::sleep(*UPDATE_INTERVAL).await;
                 if let Err(e) = try_update(&state_copy, &mut fetcher).await {
                     eprintln!("Error updating data: {:#?}", e);
                 }
-                actix_web::rt::time::sleep(*UPDATE_INTERVAL).await;
             }
         });
     });
@@ -60,11 +63,14 @@ async fn start_updater() -> Result<web::Data<AppState>, Error> {
 
 #[actix_web::main]
 async fn main() -> std::io::Result<()> {
+    simple_logger::init_with_level(log::Level::Info).unwrap();
+
     let state = start_updater().await.expect("Failed to initialize state");
 
     HttpServer::new(move || {
         App::new()
             .app_data(state.clone())
+            .wrap(Logger::default())
             .service(csv)
             .service(json)
             .service(actix_files::Files::new("/", "./static/").index_file("index.html"))