From 85d5d1bc363601fd826f97e3812ad324853fc61a Mon Sep 17 00:00:00 2001
From: Eduard Urbach <e.urbach@gmail.com>
Date: Thu, 8 Mar 2018 21:00:51 +0100
Subject: [PATCH] Improved crawler

---
 jobs/crawler/crawler.go | 46 ++++++++++++++++++++++++++++++++++++-----
 1 file changed, 41 insertions(+), 5 deletions(-)

diff --git a/jobs/crawler/crawler.go b/jobs/crawler/crawler.go
index 1696283e..12d6a90c 100644
--- a/jobs/crawler/crawler.go
+++ b/jobs/crawler/crawler.go
@@ -2,6 +2,7 @@ package main
 
 import (
 	"fmt"
+	"os"
 	"time"
 
 	"github.com/animenotifier/arn"
@@ -10,13 +11,17 @@ import (
 	"github.com/aerogo/crawler"
 )
 
-const userAgent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/64.0.3282.166 Safari/537.36"
+const (
+	// The maximum age of files we accept until we force a refresh.
+	maxAge    = 30 * 24 * time.Hour
+	userAgent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/64.0.3282.166 Safari/537.36"
+)
 
 func main() {
 	defer arn.Node.Close()
 
-	malCrawler := crawler.New(userAgent, 1*time.Second, 20000)
-	count := 0
+	// Filter anime with MAL ID
+	animes := []*arn.Anime{}
 
 	for anime := range arn.StreamAnime() {
 		malID := anime.GetMapping("myanimelist/anime")
@@ -25,9 +30,40 @@ func main() {
 			continue
 		}
 
+		animes = append(animes, anime)
+	}
+
+	color.Yellow("Found %d anime", len(animes))
+
+	// Create crawler
+	malCrawler := crawler.New(
+		map[string]string{
+			"User-Agent": userAgent,
+		},
+		1500*time.Millisecond,
+		len(animes),
+	)
+
+	// Sort so that we download the most important ones first
+	arn.SortAnimeByQuality(animes, "")
+
+	// Queue up URLs
+	count := 0
+
+	for _, anime := range animes {
+		malID := anime.GetMapping("myanimelist/anime")
+		url := "https://myanimelist.net/anime/" + malID
+		filePath := fmt.Sprintf("mal/anime-%s.html", malID)
+		fileInfo, err := os.Stat(filePath)
+
+		if err == nil && time.Since(fileInfo.ModTime()) <= maxAge {
+			// fmt.Println(color.YellowString(url), "skip")
+			continue
+		}
+
 		malCrawler.Queue(&crawler.Task{
-			URL:         "https://myanimelist.net/anime/" + malID,
-			Destination: fmt.Sprintf("mal/anime-%s.html", malID),
+			URL:         url,
+			Destination: filePath,
 		})
 
 		count++