remove caching in database for performance reasons

cmd/crawler/collectors/collector.go

@@ -4,4 +4,5 @@ import "crowsnest/internal/model/database"
 
 type Collector struct {
 	Responses *database.ResponseModel
+	Articles  *database.ArticleModel
 }

cmd/crawler/collectors/spiegel.go

@@ -2,47 +2,42 @@ package collectors
 
 import (
 	"crowsnest/internal/model"
+	"errors"
 	"fmt"
 	"log"
+	"regexp"
 	"strings"
 	"time"
 
+	"github.com/PuerkitoBio/goquery"
 	"github.com/gocolly/colly/v2"
 )
 
-func (c *Collector) Spiegel() {
+func (c *Collector) SpiegelCollect() {
 	collycollector := colly.NewCollector(
 		colly.AllowedDomains("www.spiegel.de", "spiegel.de"),
 		colly.CacheDir("./persistence/spiegel_cache"),
-		colly.MaxDepth(2),
+		colly.MaxDepth(5),
 	)
 
 	// cache
-	collycollector.OnRequest(func(r *colly.Request) {
-		url := r.URL.String()
-		exists, err := c.Responses.UrlExists(url)
-		if err == nil && !exists {
-			c.Responses.Insert(url, nil)
-			log.Println("request", url)
-		} else {
-			r.Abort()
-		}
-	})
 	collycollector.OnResponse(func(r *colly.Response) {
 		url := r.Request.URL.String()
-		c.Responses.Update(&model.Response{Url: url, Content: r.Body, FetchDate: time.Now(), Processed: false})
-		log.Println("response cached", url)
+		err := c.SpiegelExtract(url, r.Body)
+		if err == nil {
+			log.Println("added article", url)
+		} else {
+			log.Println("failed to add article:", err, "("+url+")")
+		}
 	})
 
 	// cascade
 	collycollector.OnHTML("a[href]", func(e *colly.HTMLElement) {
 		url := e.Attr("href")
-		log.Println("found", url)
 
 		if !strings.HasPrefix(url, "http") {
 			return
 		}
-		log.Println("visiting", url)
 		e.Request.Visit(url)
 	})
 
@@ -57,3 +52,96 @@ func (c *Collector) Spiegel() {
 		collycollector.Visit(url)
 	}
 }
+
+func (c *Collector) SpiegelExtract(url string, body []byte) error {
+	paywall_pattern := regexp.MustCompile(`"paywall":{"attributes":{"is_active":true`)
+	url_pattern := regexp.MustCompile(`^https://(www\.)?spiegel.de.*`)
+	whitespace := regexp.MustCompile(`\s+`)
+
+	var exists bool
+	var pagetype, title, content, datestr, author string
+	var tag *goquery.Selection
+	var date time.Time
+
+	// check url url pattern
+	if !url_pattern.Match([]byte(url)) {
+		return errors.New("invalid url pattern")
+	}
+
+	// check for paywall
+	if paywall_pattern.Match(body) {
+		return errors.New("unable to extract article due to paywal")
+	}
+
+	// construct goquery doc
+	doc, err := goquery.NewDocumentFromReader(strings.NewReader(string(body)))
+	if err != nil {
+		return err
+	}
+
+	// check for article type
+	tag = doc.Find("meta[property='og:type']")
+	pagetype, exists = tag.Attr("content")
+	if !exists || pagetype != "article" {
+		return errors.New("unable to extract article, not of type article")
+	}
+
+	// get title
+	tag = doc.Find("meta[property='og:title']")
+	title, exists = tag.Attr("content")
+	if !exists {
+		return errors.New("unable to extract article, no title tag")
+	}
+
+	// prepend description to content of article
+	tag = doc.Find("meta[name='description']")
+	content, exists = tag.Attr("content")
+	content += " "
+	if !exists {
+		return errors.New("unable to extract article, no description tag")
+	}
+
+	// get publishing date
+	tag = doc.Find("meta[name='date']")
+	datestr, exists = tag.Attr("content")
+	if !exists {
+		return errors.New("unable to extract article, no date tag")
+	}
+
+	date, err = time.Parse("2006-01-02T15:04:05-07:00", datestr)
+	if err != nil {
+		return err
+	}
+
+	// get author
+	tag = doc.Find("meta[name='author']")
+	author, exists = tag.Attr("content")
+	if !exists {
+		return errors.New("unable to extract article, no author tag")
+	}
+
+	// get content
+	tag = doc.Find("main[id='Inhalt'] div > p")
+
+	tag.Each(func(index int, p *goquery.Selection) {
+		content += " " + p.Text()
+	})
+
+	// clean up content string
+	content = string(whitespace.ReplaceAll([]byte(content), []byte(" ")))
+	content = strings.ReplaceAll(content, "»", "\"")
+	content = strings.ReplaceAll(content, "«", "\"")
+
+	// insert new article
+	article := model.Article{
+		SourceUrl:   url,
+		PublishDate: date,
+		FetchDate:   time.Now(),
+		Title:       title,
+		Content:     content,
+		Author:      author,
+	}
+
+	err = c.Articles.Insert(&article)
+	return err
+}