first commit

2026-02-06 23:20:46 +08:00 · 2021-11-28 19:11:49 +08:00
commit 3d2ffb4ba9
5 changed files with 231 additions and 0 deletions
--- a/go.mod
+++ b/go.mod
@@ -0,0 +1,8 @@
+module github.com/fengxxc/wechatmp2markdown
+
+go 1.16
+
+require (
+	github.com/PuerkitoBio/goquery v1.8.0
+	golang.org/x/net v0.0.0-20211123203042-d83791d6bcd9 // indirect
+)
--- a/go.sum
+++ b/go.sum
@@ -0,0 +1,12 @@
+github.com/PuerkitoBio/goquery v1.8.0 h1:PJTF7AmFCFKk1N6V6jmKfrNH9tV5pNE6lZMkG0gta/U=
+github.com/PuerkitoBio/goquery v1.8.0/go.mod h1:ypIiRMtY7COPGk+I/YbZLbxsxn9g5ejnI2HSMtkjZvI=
+github.com/andybalholm/cascadia v1.3.1 h1:nhxRkql1kdYCc8Snf7D5/D3spOX+dBgjA6u8x004T2c=
+github.com/andybalholm/cascadia v1.3.1/go.mod h1:R4bJ1UQfqADjvDa4P6HZHLh/3OxWWEqc0Sk8XGwHqvA=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/net v0.0.0-20211123203042-d83791d6bcd9 h1:0qxwC5n+ttVOINCBeRHO0nq9X7uy8SDsPoi5OaCdIEI=
+golang.org/x/net v0.0.0-20211123203042-d83791d6bcd9/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
+golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
--- a/main.go
+++ b/main.go
@@ -0,0 +1,7 @@
+package main
+
+import "github.com/fengxxc/wechatmp2markdown/parse"
+
+func main() {
+	parse.Test()
+}
--- a/parse/model.go
+++ b/parse/model.go
@@ -0,0 +1,33 @@
+package parse
+
+type Article struct {
+	title   string
+	meta    string
+	tags    string
+	content []Block
+}
+
+type Block struct {
+	tokens []Token
+}
+
+type Token struct {
+	ttype TokenType
+	text  string
+	attrs map[string]string
+}
+
+type TokenType int32
+
+const (
+	TITLE       TokenType = iota // 标题
+	LINK                         // 链接
+	NORMAL_TEXT                  // 文字
+	STRONG_TEXT                  // 强调文字
+	ITALIC_TEXT                  // 斜体文字
+	IMAGE                        // 图片
+	TABLE                        // 表格
+	CODE_INLINE                  // 代码 内联
+	CODE_BLOCK                   // 代码 块
+	CITE                         // 引用
+)
--- a/parse/parse.go
+++ b/parse/parse.go
@@ -0,0 +1,171 @@
+package parse
+
+import (
+	"fmt"
+	"io"
+	"log"
+	"regexp"
+	"strings"
+
+	"github.com/PuerkitoBio/goquery"
+)
+
+func parseSection(s *goquery.Selection) Block {
+	// fmt.Printf("s.Length() = %d\n", s.Length())
+	// fmt.Printf("s.Size() = %d\n", s.Size())
+	// var tokens = make([]Token, s.Size())
+	var tokens []Token
+	s.Children().Each(func(i int, s *goquery.Selection) {
+		var t Token
+		attr := make(map[string]string)
+		if s.Is("span") {
+			t = Token{NORMAL_TEXT, s.Text(), nil}
+		} else if s.Is("a") {
+			attr["href"], _ = s.Attr("href")
+			t = Token{LINK, removeBrAndBlank(s.Text()), attr}
+		} else if s.Is("img") {
+			attr["src"], _ = s.Attr("src")
+			t = Token{IMAGE, "", attr}
+		} else {
+			t = Token{NORMAL_TEXT, s.Text(), nil}
+			// TODO
+		}
+		// fmt.Printf("i = %d\n", i)
+		// fmt.Printf("%+v\n", t)
+		// tokens[i] = t
+		tokens = append(tokens, t)
+	})
+	return Block{tokens}
+}
+
+func ParseFromReader(r io.Reader) Article {
+	var article Article
+	doc, err := goquery.NewDocumentFromReader(r)
+	if err != nil {
+		log.Fatal(err)
+	}
+	var mainContent *goquery.Selection = doc.Find("#img-content")
+
+	// 标题
+	title := mainContent.Find("#activity-name").Text()
+	fmt.Println(title)
+	article.title = title
+
+	// meta 细节待完善
+	meta := mainContent.Find("#meta_content").Text()
+	meta = removeBrAndBlank(meta)
+	fmt.Println(meta)
+	article.meta = meta
+
+	// tags 细节待完善
+	tags := mainContent.Find("#js_tags").Text()
+	tags = removeBrAndBlank(tags)
+	fmt.Println(tags)
+	article.tags = tags
+
+	// content
+	// section[style="line-height: 1.5em;"]>span,a	=> 一般段落（含文本和超链接）
+	// p[style="line-height: 1.5em;"]				=> 项目列表（有序/无序）
+	// section[style=".*text-align:center"]>img		=> 居中段落（图片）
+	content := mainContent.Find("#js_content")
+	var sections []Block
+	content.Find("section,p").Each(func(i int, s *goquery.Selection) {
+		fmt.Println(s.Text())
+		// fmt.Println(s.Attr("style"))
+		var block Block
+		if s.Is("p") {
+			block = parseSection(s)
+		} else if s.Is("section") {
+			block = parseSection(s)
+		} else {
+			// TODO
+		}
+		// sections[i] = block
+		sections = append(sections, block)
+	})
+	article.content = sections
+
+	return article
+}
+
+func ParseFromHTMLString(s string) Article {
+	return ParseFromReader(strings.NewReader(s))
+}
+
+func removeBrAndBlank(s string) string {
+	regstr := "\\s{2,}"
+	reg, _ := regexp.Compile(regstr)
+	sb := make([]byte, len(s))
+	copy(sb, s)
+	spc_index := reg.FindStringIndex(string(sb)) //在字符串中搜索
+	for len(spc_index) > 0 {                     //找到适配项
+		sb = append(sb[:spc_index[0]+1], sb[spc_index[1]:]...) //删除多余空格
+		spc_index = reg.FindStringIndex(string(sb))            //继续在字符串中搜索
+	}
+	return strings.Replace(string(sb), "\n", " ", -1)
+}
+
+var testHTML = `
+<div>
+    <div id="img-content" class="rich_media_wrp">
+        <h1 class="rich_media_title" id="activity-name">终于有人喷我了！</h1>
+        <div id="meta_content" class="rich_media_meta_list">
+            <span id="copyright_logo" class="wx_tap_link js_wx_tap_highlight rich_media_meta icon_appmsg_tag appmsg_title_tag weui-wa-hotarea" wah-hotarea="click">原创</span>
+            <span class="rich_media_meta rich_media_meta_text">
+                <span role="link" id="js_author_name" class="wx_tap_link js_wx_tap_highlight weui-wa-hotarea" datarewardsn="" datatimestamp="" datacanreward="0" wah-hotarea="click">闪客sun</span>
+            </span>
+            <span class="rich_media_meta rich_media_meta_nickname" id="profileBt" wah-hotarea="click">
+                <a href="javascript:void(0);" class="wx_tap_link js_wx_tap_highlight weui-wa-hotarea" id="js_name">
+                    低并发编程 </a>
+                <div id="js_profile_qrcode" aria-hidden="true" class="profile_container" style="display:none;" wah-hotarea="click">
+                    <div class="profile_inner">
+                        <strong class="profile_nickname">低并发编程</strong>
+                        <img class="profile_avatar" id="js_profile_qrcode_img" src="" alt="">
+                        <p class="profile_meta">
+                            <label class="profile_meta_label">微信号</label>
+                            <span class="profile_meta_value">dibingfa</span>
+                        </p>
+                        <p class="profile_meta">
+                            <label class="profile_meta_label">功能介绍</label>
+                            <span class="profile_meta_value">战略上藐视技术，战术上重视技术</span>
+                        </p>
+                    </div>
+                    <span class="profile_arrow_wrp" id="js_profile_arrow_wrp">
+                        <i class="profile_arrow arrow_out"></i>
+                        <i class="profile_arrow arrow_in"></i>
+                    </span>
+                </div>
+            </span>
+            <em id="publish_time" class="rich_media_meta rich_media_meta_text">2021-11-26</em>
+        </div>
+		<div id="js_tags" class="article-tag__list single-tag__wrp js_single js_wx_tap_highlight wx_tap_card" data-len="1" role="link" aria-labelledby="js_article-tag-card__left" aria-describedby="js_article-tag-card__right" wah-hotarea="click">
+			<span aria-hidden="true" id="js_article-tag-card__left" class="article-tag-card__left">
+				<span class="article-tag-card__title">收录于话题</span>
+				<span class="article-tag__item-wrp no-active js_tag" data-url="https://mp.weixin.qq.com/mp/appmsgalbum?__biz=Mzk0MjE3NDE0Ng==&amp;action=getalbum&amp;album_id=1645521656368381958#wechat_redirect" data-tag_id="" data-album_id="1645521656368381958" data-tag_source="4">
+					<span class="article-tag__item">#随便聊聊</span>
+				</span>
+			</span>
+			<span aria-hidden="true" id="js_article-tag-card__right" class="article-tag-card__right">49个<span class="weui-hidden_abs">内容</span></span>
+		</div>
+		<div class="rich_media_content " id="js_content" style="visibility: visible;">
+			<p data-mpa-powered-by="yiban.io"><span style="font-size: 16px;letter-spacing: 0.5px;background-color: transparent;caret-color: var(--weui-BRAND);">写公众号一年了，一直盼着能有人喷喷我，今天终于被我碰到了！</span><br></p>
+			<section style="line-height: 1.5em;"><span style="letter-spacing: 0.5px;font-size: 16px;">‍‍</span></section>
+			<section style="line-height: 1.5em;"><span style="letter-spacing: 0.5px;font-size: 16px;">还是我的一位读者发现的，在推特上，于是分享给了我。<br></span></section>
+			<section style="line-height: 1.5em;text-align: center;"><img class="rich_pages wxw-img" data-galleryid="" data-ratio="1.0826306913996628" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_png/GLeh42uInXRdNibLb2hf6QnMnWgic4Nm0KhCmicJibxESMoGfbuMrXbQB7lrYFSJPlBeGaJyciaavIBN8NLwESxia7cA/640?wx_fmt=png" data-type="png" data-w="593" style="box-shadow: rgb(210, 210, 210) 0em 0em 0.5em 0px; font-size: 17px; width: 346px !important; height: auto !important; visibility: visible !important;" _width="346px" src="https://mmbiz.qpic.cn/mmbiz_png/GLeh42uInXRdNibLb2hf6QnMnWgic4Nm0KhCmicJibxESMoGfbuMrXbQB7lrYFSJPlBeGaJyciaavIBN8NLwESxia7cA/640?wx_fmt=png&amp;tp=webp&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1" crossorigin="anonymous" alt="图片" data-fail="0"></section>
+			<section style="line-height: 1.5em;">
+				<span style="letter-spacing: 0.5px;font-size: 16px;">这是喷我最近的一个新系列，</span>
+				<a target="_blank" href="https://mp.weixin.qq.com/mp/appmsgalbum?__biz=Mzk0MjE3NDE0Ng==&amp;action=getalbum&amp;album_id=2123743679373688834#wechat_redirect" textvalue="你管这破玩意叫操作系统源码" linktype="text" imgurl="" imgdata="null" tab="innerlink" data-linktype="2" style="letter-spacing: 0.5px;font-size: 16px;" wah-hotarea="click">
+					<span style="letter-spacing: 0.5px;font-size: 16px;">你管这破玩意叫操作系统源码</span>
+				</a>
+				<span style="letter-spacing: 0.5px;font-size: 16px;">，正愁找不到借口推广一波呢，这不就给我来素材了。<br></span>
+			</section>
+		</div>
+    </div>
+</div>
+`
+
+func Test() {
+	res := ParseFromHTMLString(testHTML)
+	fmt.Println("---------------------------------------")
+	fmt.Printf("%+v\n", res)
+}