From b16ebb95729976fad13c18c16c0a74ccf41f13ec Mon Sep 17 00:00:00 2001
From: Elias Kohout <elias@kohout.de>
Date: Tue, 7 Jan 2025 11:41:17 +0100
Subject: [PATCH] add zeit.de collector

---
 cmd/crawler/collectors/spiegel.go |   8 +-
 cmd/crawler/collectors/zeit.go    | 124 ++++++++++++++++++++++++++++--
 2 files changed, 120 insertions(+), 12 deletions(-)

diff --git a/cmd/crawler/collectors/spiegel.go b/cmd/crawler/collectors/spiegel.go
index 3f1dcbe..08dbd9b 100644
--- a/cmd/crawler/collectors/spiegel.go
+++ b/cmd/crawler/collectors/spiegel.go
@@ -13,17 +13,17 @@ import (
 	"github.com/gocolly/colly/v2"
 )
 
-func (c *Collector) SpiegelCollect() {
+func (c *Collector) CollectSpiegel() {
 	collycollector := colly.NewCollector(
 		colly.AllowedDomains("www.spiegel.de", "spiegel.de"),
 		colly.CacheDir("./persistence/spiegel_cache"),
 		colly.MaxDepth(5),
 	)
 
-	// cache
+	// store articles
 	collycollector.OnResponse(func(r *colly.Response) {
 		url := r.Request.URL.String()
-		err := c.SpiegelExtract(url, r.Body)
+		err := c.ExtractSpiegel(url, r.Body)
 		if err == nil {
 			log.Println("added article", url)
 		} else {
@@ -53,7 +53,7 @@ func (c *Collector) SpiegelCollect() {
 	}
 }
 
-func (c *Collector) SpiegelExtract(url string, body []byte) error {
+func (c *Collector) ExtractSpiegel(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+`)
diff --git a/cmd/crawler/collectors/zeit.go b/cmd/crawler/collectors/zeit.go
index 621d9dd..606b988 100644
--- a/cmd/crawler/collectors/zeit.go
+++ b/cmd/crawler/collectors/zeit.go
@@ -1,33 +1,51 @@
 package collectors
 
 import (
+	"crowsnest/internal/model"
+	"errors"
 	"fmt"
+	"log"
+	"regexp"
+	"strings"
 	"time"
 
+	"github.com/PuerkitoBio/goquery"
 	"github.com/gocolly/colly/v2"
 )
 
 // Gets every page of the archive of zeit.de and stores the responses into the
 // database.
-func (c *Collector) Zeit() {
+func (c *Collector) CollectZeit() {
 	collycollector := colly.NewCollector(
 		colly.AllowedDomains("www.zeit.de", "zeit.de"),
 		colly.CacheDir("./persistence/zeit_cache"),
-		colly.MaxDepth(2),
+		colly.MaxDepth(5),
 	)
 
+	// store articles
+	collycollector.OnResponse(func(r *colly.Response) {
+		url := r.Request.URL.String()
+		err := c.ExtractZeit(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) {
-		e.Request.Visit(e.Attr("href"))
-	})
+		url := e.Attr("href")
 
-	// cache
-	collycollector.OnScraped(func(r *colly.Response) {
-		c.Responses.Insert(r.Request.URL.String(), r.Body)
+		if !strings.HasPrefix(url, "http") {
+			return
+		}
+		e.Request.Visit(url)
 	})
 
 	// go through archive
-	startDate := time.Date(1946, time.January, 1, 0, 0, 0, 0, time.UTC)
+	startDate := time.Date(2025, time.January, 1, 0, 0, 0, 0, time.UTC)
+	//startDate := time.Date(1946, time.January, 1, 0, 0, 0, 0, time.UTC)
 	currentDate := time.Now()
 
 	for date := startDate; date.Before(currentDate) || date.Equal(currentDate); date = date.AddDate(0, 0, 7) {
@@ -37,3 +55,93 @@ func (c *Collector) Zeit() {
 		collycollector.Visit(url)
 	}
 }
+
+func (c *Collector) ExtractZeit(url string, body []byte) error {
+	url_pattern := regexp.MustCompile(`^https://(www\.)?zeit.de.*`)
+	whitespace := regexp.MustCompile(`\s+`)
+
+	var exists bool
+	var pagetype, title, content, datestr 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")
+	}
+
+	// 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")
+	}
+
+	// check for paywall
+	tag = doc.Find("meta[property='article:content_tier']")
+	pagetype, exists = tag.Attr("content")
+	if !exists || pagetype != "free" {
+		return errors.New("unable to extract article due to paywal")
+	}
+
+	// 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")
+	}
+
+	if strings.Contains(content, "Das Liveblog") {
+		return errors.New("unable to extract article, no support for liveblog")
+	}
+
+	// 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 content
+	tag = doc.Find("main > article > div.article-body p.article__item")
+
+	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,
+	}
+
+	err = c.Articles.Insert(&article)
+	return err
+}