Merge branch 'master' of https://github.com/FloatTech/ZeroBot-Plugin

plugin/moyu_calendar/calendar.go

@@ -12,15 +12,20 @@ import (
 	"strings"
 	"time"
 
-	"github.com/FloatTech/zbputils/binary"
 	control "github.com/FloatTech/zbputils/control"
+	"github.com/FloatTech/zbputils/control/order"
 	"github.com/FloatTech/zbputils/process"
 	zero "github.com/wdvxdr1123/ZeroBot"
 	"github.com/wdvxdr1123/ZeroBot/message"
 
-	"github.com/FloatTech/zbputils/control/order"
+	xpath "github.com/antchfx/htmlquery"
 )
 
+var ua = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.182 Safari/537.36"
+var weixin = regexp.MustCompile(`url \+= '(.+)';`)
+
+var client = &http.Client{}
+
 func init() {
 	control.Register("moyucalendar", order.AcquirePrio(), &control.Options{
 		DisableOnDefault: true,
@@ -29,20 +34,41 @@ func init() {
 			"- /禁用 moyucalendar",
 	}).OnFullMatch("摸鱼人日历").SetBlock(true).
 		Handle(func(ctx *zero.Ctx) {
-			image, err := crew()
+			title := fmt.Sprintf("摸鱼人日历 %d月%d日", time.Now().Month(), time.Now().Day())
+			sg, cookies, err := sougou(title, "摸鱼人日历", ua)
 			if err != nil {
 				ctx.SendChain(message.Text("ERROR: ", err))
+				return
+			}
+			wx, err := redirect(sg, cookies, ua)
+			if err != nil {
+				ctx.SendChain(message.Text("ERROR: ", err))
+				return
+			}
+			image, err := calendar(wx, ua)
+			if err != nil {
+				ctx.SendChain(message.Text("ERROR: ", err))
+				return
 			}
 			ctx.SendChain(message.Image(image))
 		})
 
-	// 定时任务每天8点执行一次
+	// 定时任务每天8点30分执行一次
 	_, err := process.CronTab.AddFunc("30 8 * * *", func() {
 		m, ok := control.Lookup("moyucalendar")
 		if !ok {
 			return
 		}
-		image, err := crew()
+		title := fmt.Sprintf("摸鱼人日历 %d月%d日", time.Now().Month(), time.Now().Day())
+		sg, cookies, err := sougou(title, "摸鱼人日历", ua)
+		if err != nil {
+			return
+		}
+		wx, err := redirect(sg, cookies, ua)
+		if err != nil {
+			return
+		}
+		image, err := calendar(wx, ua)
 		if err != nil {
 			return
 		}
@@ -62,23 +88,73 @@ func init() {
 	}
 }
 
-var newest = regexp.MustCompile(`href="(/link.+?)" id="sogou_vr_11002601_title_0" uigs="article_title_0"`)
+func sougou(title, publisher, ua string) (string, []*http.Cookie, error) {
-var weixin = regexp.MustCompile(`url \+= '(.+)';`)
-var calendar = regexp.MustCompile(`data-src="(.{0,300})" data-type="png" data-w="540"`)
-
-func crew() (string, error) {
-	client := &http.Client{}
 	u, _ := url.Parse("https://weixin.sogou.com/weixin")
 	u.RawQuery = url.Values{
 		"type":   []string{"2"},
 		"s_from": []string{"input"},
-		"query":  []string{fmt.Sprintf("摸鱼人日历 %d月%d日", time.Now().Month(), time.Now().Day())},
+		"query":  []string{title},
 	}.Encode()
 	req, err := http.NewRequest("GET", u.String(), nil)
+	if err != nil {
+		return "", nil, err
+	}
+	req.Header.Set("User-Agent", ua)
+	resp, err := client.Do(req)
+	if err != nil {
+		return "", nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return "", nil, errors.New("status not ok")
+	}
+	// 解析XPATH
+	doc, err := xpath.Parse(resp.Body)
+	if err != nil {
+		return "", nil, err
+	}
+	// 取出每个返回的结果
+	list := xpath.Find(doc, `//*[@class="news-list"]/li/div[@class="txt-box"]`)
+	if len(list) == 0 {
+		return "", nil, errors.New("sougou result not found")
+	}
+	var match string
+	for i := range list {
+		account := xpath.FindOne(list[i], `//div[@class="s-p"]/a[@class="account"]`)
+		if account == nil {
+			continue
+		}
+		if xpath.InnerText(account) != publisher {
+			continue
+		}
+		target := xpath.FindOne(list[i], `//h3/a[@target="_blank"]`)
+		if target == nil {
+			continue
+		}
+		match = xpath.SelectAttr(target, "href")
+		break
+	}
+	if match == "" {
+		return "", nil, errors.New("sougou result not found")
+	}
+	return "https://weixin.sogou.com" + match, resp.Cookies(), nil
+}
+
+func redirect(link string, cookies []*http.Cookie, ua string) (string, error) {
+	req, err := http.NewRequest("GET", link, nil)
 	if err != nil {
 		return "", err
 	}
-	req.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.182 Safari/537.36")
+	req.Header.Set("User-Agent", ua)
+	var c = make([]string, 0, 4)
+	for _, cookie := range cookies {
+		if cookie.Name != "ABTEST" && cookie.Name != "SNUID" &&
+			cookie.Name != "IPLOC" && cookie.Name != "SUID" {
+			continue
+		}
+		c = append(c, cookie.Name+"="+cookie.Value)
+	}
+	req.Header.Set("Cookie", strings.Join(c, "; "))
 	resp, err := client.Do(req)
 	if err != nil {
 		return "", err
@@ -87,38 +163,8 @@ func crew() (string, error) {
 	if resp.StatusCode != http.StatusOK {
 		return "", errors.New("status not ok")
 	}
-	b, err := io.ReadAll(resp.Body)
+	br := bufio.NewReader(resp.Body)
-	if err != nil {
+	var u = make([]string, 0)
-		return "", err
-	}
-	match := newest.FindStringSubmatch(binary.BytesToString(b))
-	if len(match) < 2 {
-		return "", errors.New("newest not found")
-	}
-	var link = "https://weixin.sogou.com" + match[1]
-	reqa, err := http.NewRequest("GET", link, nil)
-	if err != nil {
-		return "", err
-	}
-	reqa.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.182 Safari/537.36")
-	var cookies = make([]string, 0, 4)
-	for _, cookie := range resp.Cookies() {
-		if cookie.Name != "ABTEST" && cookie.Name != "SNUID" && cookie.Name != "IPLOC" && cookie.Name != "SUID" {
-			continue
-		}
-		cookies = append(cookies, cookie.Name+"="+cookie.Value)
-	}
-	reqa.Header.Set("Cookie", strings.Join(cookies, "; "))
-	respa, err := client.Do(reqa)
-	if err != nil {
-		return "", err
-	}
-	defer respa.Body.Close()
-	if respa.StatusCode != http.StatusOK {
-		return "", errors.New("status not ok")
-	}
-	br := bufio.NewReader(respa.Body)
-	var weixinurl = make([]string, 0)
 	for {
 		b, _, err := br.ReadLine()
 		if err == io.EOF {
@@ -127,39 +173,54 @@ func crew() (string, error) {
 		if err != nil {
 			return "", err
 		}
-		matcha := weixin.FindStringSubmatch(binary.BytesToString(b))
+		matcha := weixin.FindStringSubmatch(string(b))
 		if len(matcha) < 2 {
 			continue
 		}
-		weixinurl = append(weixinurl, strings.ReplaceAll(matcha[1], "@", ""))
+		u = append(u, strings.ReplaceAll(matcha[1], "@", ""))
 	}
-	if len(weixinurl) == 0 {
+	if len(u) == 0 {
 		return "", errors.New("weixin url not found")
 	}
-	reqw, err := http.NewRequest("GET", strings.Join(weixinurl, ""), nil)
+	return strings.Join(u, ""), nil
-	reqa.Header.Set("User-Agent", "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.182 Safari/537.36")
+}
+
+func calendar(link, ua string) (string, error) {
+	req, err := http.NewRequest("GET", link, nil)
+	req.Header.Set("User-Agent", ua)
 	if err != nil {
 		return "", err
 	}
-	respw, err := client.Do(reqw)
+	resp, err := client.Do(req)
 	if err != nil {
 		return "", err
 	}
-	defer respw.Body.Close()
+	defer resp.Body.Close()
-	if respw.StatusCode != http.StatusOK {
+	if resp.StatusCode != http.StatusOK {
 		return "", errors.New("status not ok")
 	}
-	bw, _ := io.ReadAll(respw.Body)
+	doc, err := xpath.Parse(resp.Body)
-	today, err := regexp.Compile(time.Now().Format("2006-01-02"))
 	if err != nil {
 		return "", err
 	}
-	if !today.Match(bw) {
+	html := xpath.OutputHTML(doc, false)
-		return "", errors.New("today not found")
+	if !strings.Contains(html, time.Now().Format("2006-01-02")) {
+		return "", errors.New("calendar not today")
 	}
-	matchw := calendar.FindStringSubmatch(binary.BytesToString(bw))
+	images := xpath.Find(doc, `//*[@id="js_content"]/p/img`)
-	if len(matchw) < 2 {
+	if images == nil {
 		return "", errors.New("calendar not found")
 	}
-	return matchw[1], nil
+	var image string
+	for i := range images {
+		if xpath.SelectAttr(images[i], "data-w") != "540" {
+			continue
+		}
+		image = xpath.SelectAttr(images[i], "data-src")
+		break
+	}
+	if image == "" {
+		return "", errors.New("image not found")
+	}
+	return image, nil
 }