feat: 添加从文件读取功能,添加测试用例文件

This commit is contained in:
fengxxc
2021-11-28 20:23:26 +08:00
parent 3d2ffb4ba9
commit ab6c8488d5
4 changed files with 100 additions and 67 deletions

View File

@@ -1,7 +1,7 @@
package main
import "github.com/fengxxc/wechatmp2markdown/parse"
import "github.com/fengxxc/wechatmp2markdown/test"
func main() {
parse.Test()
test.Test1()
}

View File

@@ -1,9 +1,12 @@
package parse
import (
"bytes"
"fmt"
"io"
"io/ioutil"
"log"
"os"
"regexp"
"strings"
@@ -92,6 +95,19 @@ func ParseFromHTMLString(s string) Article {
return ParseFromReader(strings.NewReader(s))
}
func ParseFromHTMLFile(filepath string) Article {
file, err := os.Open(filepath)
if err != nil {
panic(err)
}
defer file.Close()
content, err2 := ioutil.ReadAll(file)
if err2 != nil {
panic(err)
}
return ParseFromReader(bytes.NewReader(content))
}
func removeBrAndBlank(s string) string {
regstr := "\\s{2,}"
reg, _ := regexp.Compile(regstr)
@@ -104,68 +120,3 @@ func removeBrAndBlank(s string) string {
}
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)
}

13
test/test1.go Normal file
View File

@@ -0,0 +1,13 @@
package test
import (
"fmt"
"github.com/fengxxc/wechatmp2markdown/parse"
)
func Test1() {
res := parse.ParseFromHTMLFile("./test/test1.html")
fmt.Println("-------------------test1.html-------------------")
fmt.Printf("%+v\n", res)
}

69
test/test1.html Normal file
View File

@@ -0,0 +1,69 @@
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>test1</title>
</head>
<body>
<div>
<div id="img-content" class="rich_media_wrp">
<h1 class="rich_media_title" id="activity-name">这里是文章标题</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">这里是作者</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">
这里是公众号名 </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">这里是公众号名</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">这里是微信号</span>
</p>
<p class="profile_meta">
<label class="profile_meta_label">功能介绍</label>
<span class="profile_meta_value">这里是功能介绍</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">2021-01-01</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">#话题1</span>
</span>
</span>
<span aria-hidden="true" id="js_article-tag-card__right" class="article-tag-card__right">69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);">正文第一行</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;">正文第二行。<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;">正文第三行part1文本</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;">正文第三行part2链接</span>
</a>
<span style="letter-spacing: 0.5px;font-size: 16px;">正文第三行part3文本。<br></span>
</section>
</div>
</div>
</div>
</body>
</html>