fin facebook crawler

Dockerfile

@@ -4,6 +4,6 @@ COPY . .
 RUN apk add --no-cache git make && \
   go build -o crawler .
 
-FROM geckodriver:latest
+FROM docker.mtfos.xyz/mtfos/geckodriver-alpine:latest
 COPY --from=builder /data/crawler /usr/bin/crawler
 CMD ["/usr/bin/crawler"]

main.go

@@ -4,7 +4,7 @@ import (
 	"fmt"
 	"log"
 
-	"github.com/tebeka/selenium"
+	"git.trj.tw/golang/go-crawler/modules/facebook"
 
 	"git.trj.tw/golang/go-crawler/modules/browser"
 )
@@ -21,22 +21,5 @@ func main() {
 		log.Fatal(err)
 	}
 	defer func() { browser.StopService() }()
-
-	wd, err := browser.NewWD()
-	if err != nil {
-		log.Fatal(err)
-	}
-	defer wd.Quit()
-
-	err = wd.Get("https://tools.trj.tw")
-	if err != nil {
-		log.Fatal(err)
-	}
-
-	el, err := wd.FindElement(selenium.ByCSSSelector, "body")
-	if err != nil {
-		log.Fatal(err)
-	}
-
-	fmt.Println(el.Text())
+	fmt.Println(facebook.GetLastPost("MTFoS-404079100096306"))
 }

modules/facebook/facebook.go

@@ -0,0 +1,174 @@
+package facebook
+
+import (
+	"errors"
+	"fmt"
+	"net/url"
+	"regexp"
+	"sort"
+	"strconv"
+	"strings"
+
+	"git.trj.tw/golang/go-crawler/modules/browser"
+	"github.com/tebeka/selenium"
+)
+
+var fbURL = "https://www.facebook.com"
+
+// PostInfo -
+type PostInfo struct {
+	Text string
+	ID   string
+	Link string
+	Time int64
+}
+
+// ByPostInfo -
+type ByPostInfo []PostInfo
+
+// Len -
+func (p ByPostInfo) Len() int { return len(p) }
+
+// Swap -
+func (p ByPostInfo) Swap(i, j int) { p[i], p[j] = p[j], p[i] }
+
+// Less -
+func (p ByPostInfo) Less(i, j int) bool { return p[i].Time < p[j].Time }
+
+var (
+	idRegexps []*regexp.Regexp
+	qsRegexp  *regexp.Regexp
+)
+
+func init() {
+	idRegexps = make([]*regexp.Regexp, 0, 4)
+	idRegexps = append(idRegexps,
+		regexp.MustCompile("[\\?|&]story_fbid\\=(\\d+)"),
+		regexp.MustCompile("\\/posts\\/(\\d+)"),
+		regexp.MustCompile("\\/photos\\/.+?\\/(\\d+)"),
+		regexp.MustCompile("\\/videos\\/(\\d+)"))
+
+	qsRegexp = regexp.MustCompile("id")
+}
+
+// GetLastPost -
+func GetLastPost(page string) (info *PostInfo, err error) {
+	if len(page) == 0 {
+		return nil, errors.New("page id is empty")
+	}
+
+	pageURL, err := url.Parse(fbURL)
+	if err != nil {
+		return nil, err
+	}
+	pageURL, err = pageURL.Parse(fmt.Sprintf("/%s/posts", page))
+
+	wd, err := browser.NewWD()
+	if err != nil {
+		return nil, err
+	}
+	defer wd.Quit()
+
+	err = wd.Get(pageURL.String())
+	if err != nil {
+		return nil, err
+	}
+
+	elements, err := wd.FindElements(selenium.ByCSSSelector, "div.userContentWrapper")
+	if err != nil {
+		return nil, err
+	}
+	_ = elements
+	posts := make([]PostInfo, 0, len(elements))
+
+	for _, el := range elements {
+		post := PostInfo{}
+
+		// get timestamp
+		timeEl, err := el.FindElement(selenium.ByTagName, "abbr")
+		if err != nil {
+			continue
+		}
+		timeStr, err := timeEl.GetAttribute("data-utime")
+		if err != nil {
+			continue
+		}
+		timestamp, err := strconv.Atoi(timeStr)
+		if err != nil {
+			continue
+		}
+		post.Time = int64(timestamp)
+
+		// get link element
+		linkEl, err := timeEl.FindElement(selenium.ByXPATH, "..")
+		if err != nil {
+			continue
+		}
+		post.Link, err = linkEl.GetAttribute("href")
+		if err != nil {
+			continue
+		}
+
+		// get post content
+		postEl, err := el.FindElement(selenium.ByCSSSelector, "div.userContent")
+		if err != nil {
+			continue
+		}
+		post.Text, err = postEl.Text()
+		if err != nil {
+			continue
+		}
+
+		// get post id
+		postChile, err := postEl.FindElement(selenium.ByXPATH, "./*")
+		if err != nil {
+			continue
+		}
+		id, err := postChile.GetAttribute("id")
+		if err != nil || len(id) == 0 {
+			for _, regex := range idRegexps {
+				if regex.Match([]byte(post.Link)) {
+					strs := regex.FindAllStringSubmatch(post.Link, -1)
+					if len(strs) > 0 && len(strs[0]) > 1 {
+						id = strs[0][1]
+						break
+					}
+				}
+			}
+		}
+
+		posts = append(posts, post)
+
+	}
+
+	sort.Slice(posts, func(i, j int) bool {
+		return posts[i].Time > posts[j].Time
+	})
+
+	info = &posts[0]
+
+	// remove qs
+	urls := strings.Split(info.Link, "?")
+	if len(urls) > 1 {
+		qsStr := strings.Split(urls[1], "&")
+		if len(qsStr) > 0 {
+			qs := make([]string, 0)
+			for _, val := range qsStr {
+				items := strings.SplitN(val, "=", 1)
+				if len(items) > 0 {
+					if qsRegexp.MatchString(items[0]) {
+						qs = append(qs, val)
+					}
+				}
+			}
+
+			if len(qs) > 0 {
+				info.Link = urls[0] + "?" + strings.Join(qs, "&")
+			}
+		}
+	}
+
+	info.Link = fbURL + info.Link
+
+	return info, nil
+}