qq_bot/handler/rss/parse_test.go

package rss

import (
	"fmt"
	"net/http"
	"net/http/httptest"
	"os"
	"testing"

	"git.lxtend.com/lixiangwuxian/qqbot/config"
	. "github.com/bytedance/mockey"
	. "github.com/smartystreets/goconvey/convey"
)

func init() {
	Mock((*config.Config).LoadConfig).Return(nil).Build()
}

// 模拟RSS XML数据
const mockRSSXML = `<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
	<channel>
		<title>测试RSS源</title>
		<link>https://example.com</link>
		<description>这是一个测试RSS源</description>
		<item>
			<title>测试文章1</title>
			<link>https://example.com/article1</link>
			<description>这是第一篇测试文章的描述</description>
			<guid>https://example.com/article1</guid>
			<author>测试作者</author>
			<category>技术</category>
			<pubDate>Mon, 01 Jan 2024 00:00:00 GMT</pubDate>
		</item>
		<item>
			<title>测试文章2</title>
			<link>https://example.com/article2</link>
			<description>这是第二篇测试文章的描述</description>
			<guid>https://example.com/article2</guid>
			<author>测试作者2</author>
			<category>生活</category>
			<pubDate>Tue, 02 Jan 2024 00:00:00 GMT</pubDate>
		</item>
	</channel>
</rss>`

func TestParseRssFeed(t *testing.T) {
	Convey("TestParseRssFeed", t, func() {
		Convey("测试正常RSS解析", func() {
			// 创建模拟HTTP服务器
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				if r.Method == "HEAD" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.Header().Set("Content-Length", fmt.Sprintf("%d", len(mockRSSXML)))
					w.WriteHeader(http.StatusOK)
					return
				}
				if r.Method == "GET" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.WriteHeader(http.StatusOK)
					w.Write([]byte(mockRSSXML))
					return
				}
			}))
			defer server.Close()

			// 测试解析功能
			items, err := ParseRssFeed(server.URL)
			So(err, ShouldBeNil)
			So(len(items), ShouldEqual, 2)

			// 验证第一个条目
			// So(items[0].Title, ShouldEqual, "测试文章1")
			// So(items[0].Link, ShouldEqual, "https://example.com/article1")
			// So(items[0].Description, ShouldEqual, "这是第一篇测试文章的描述")
			// So(items[0].GUID, ShouldEqual, "https://example.com/article1")
			// So(items[0].Author, ShouldEqual, "测试作者")
			// So(items[0].Category, ShouldEqual, "技术")
			// So(items[0].Hash, ShouldNotBeEmpty)

			// // 验证第二个条目
			// So(items[1].Title, ShouldEqual, "测试文章2")
			// So(items[1].Link, ShouldEqual, "https://example.com/article2")
			// So(items[1].Description, ShouldEqual, "这是第二篇测试文章的描述")
			// So(items[1].GUID, ShouldEqual, "https://example.com/article2")
			// So(items[1].Author, ShouldEqual, "测试作者2")
			// So(items[1].Category, ShouldEqual, "生活")
			// So(items[1].Hash, ShouldNotBeEmpty)

			// 验证哈希值不同
			So(items[0].Hash, ShouldNotEqual, items[1].Hash)

			fmt.Printf("成功解析RSS源，共%d个条目\n", len(items))
			for i, item := range items {
				fmt.Printf("条目%d: %s - %s\n", i+1, item.Title, item.Link)
			}
		})

		Convey("测试RSS源大小限制", func() {
			// 创建超大内容的模拟服务器
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				if r.Method == "HEAD" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.Header().Set("Content-Length", "20971520") // 20MB
					w.WriteHeader(http.StatusOK)
					return
				}
			}))
			defer server.Close()

			_, err := ParseRssFeed(server.URL)
			So(err, ShouldNotBeNil)
			So(err.Error(), ShouldContainSubstring, "超出限制")
		})

		Convey("测试无效RSS源状态码", func() {
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				w.WriteHeader(http.StatusNotFound)
			}))
			defer server.Close()

			_, err := ParseRssFeed(server.URL)
			So(err, ShouldNotBeNil)
			So(err.Error(), ShouldContainSubstring, "RSS源无效: 404")
		})

		Convey("测试无效XML格式", func() {
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				if r.Method == "HEAD" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.Header().Set("Content-Length", "100")
					w.WriteHeader(http.StatusOK)
					return
				}
				if r.Method == "GET" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.WriteHeader(http.StatusOK)
					w.Write([]byte("这不是有效的XML"))
					return
				}
			}))
			defer server.Close()

			_, err := ParseRssFeed(server.URL)
			So(err, ShouldNotBeNil)
			So(err.Error(), ShouldContainSubstring, "解析RSS数据失败")
		})
	})
}

func TestCheckRssFeed(t *testing.T) {
	Convey("TestCheckRssFeed", t, func() {
		Convey("测试有效RSS源检查", func() {
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				w.Header().Set("Content-Type", "application/rss+xml")
				w.WriteHeader(http.StatusOK)
			}))
			defer server.Close()

			err := CheckRssFeed(server.URL)
			So(err, ShouldBeNil)
		})

		Convey("测试无效Content-Type", func() {
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				w.Header().Set("Content-Type", "text/html")
				w.WriteHeader(http.StatusOK)
			}))
			defer server.Close()

			err := CheckRssFeed(server.URL)
			So(err, ShouldNotBeNil)
			So(err.Error(), ShouldContainSubstring, "RSS源无效: text/html")
		})

		Convey("测试无效状态码", func() {
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				w.WriteHeader(http.StatusInternalServerError)
			}))
			defer server.Close()

			err := CheckRssFeed(server.URL)
			So(err, ShouldNotBeNil)
			So(err.Error(), ShouldContainSubstring, "RSS源无效: 500")
		})
	})
}

func TestRssItemHash(t *testing.T) {
	Convey("TestRssItemHash", t, func() {
		Convey("测试相同内容生成相同哈希", func() {
			content1 := "测试标题测试链接测试描述"
			content2 := "测试标题测试链接测试描述"

			// 模拟哈希生成过程
			hash1 := fmt.Sprintf("%x", []byte(content1))
			hash2 := fmt.Sprintf("%x", []byte(content2))

			So(hash1, ShouldEqual, hash2)
		})

		Convey("测试不同内容生成不同哈希", func() {
			content1 := "测试标题1测试链接1测试描述1"
			content2 := "测试标题2测试链接2测试描述2"

			hash1 := fmt.Sprintf("%x", []byte(content1))
			hash2 := fmt.Sprintf("%x", []byte(content2))

			So(hash1, ShouldNotEqual, hash2)
		})
	})
}

func TestParseRealRSSFile(t *testing.T) {
	Convey("TestParseRealRSSFile", t, func() {
		Convey("测试解析真实的RSS文件", func() {
			// 创建一个本地文件服务器
			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
				if r.Method == "HEAD" {
					w.Header().Set("Content-Type", "application/rss+xml")
					w.Header().Set("Content-Length", "100000") // 设置一个合理的大小
					w.WriteHeader(http.StatusOK)
					return
				}
				if r.Method == "GET" {
					// 读取test.xml文件
					content, err := os.ReadFile("test.xml")
					if err != nil {
						w.WriteHeader(http.StatusInternalServerError)
						return
					}
					w.Header().Set("Content-Type", "application/rss+xml")
					w.WriteHeader(http.StatusOK)
					w.Write(content)
					return
				}
			}))
			defer server.Close()

			// 测试解析功能
			items, err := ParseRssFeed(server.URL)
			So(err, ShouldBeNil)
			So(len(items), ShouldBeGreaterThan, 0)

			fmt.Printf("成功解析真实RSS源，共%d个条目\n", len(items))
			for i, item := range items {
				if i < 3 { // 只显示前3个条目
					fmt.Printf("条目%d: %s - %s\n", i+1, item.Title, item.Link)
				}
			}
		})
	})
}