diff --git a/novel-front/pom.xml b/novel-front/pom.xml index 6b3ad971..10df5b22 100644 --- a/novel-front/pom.xml +++ b/novel-front/pom.xml @@ -10,7 +10,7 @@ xyz.zinglizingli novel-front - 2.4.1.beta + 2.5.0.beta novel-front 小说精品楼-前台web网站 diff --git a/novel-front/src/main/java/xyz/zinglizingli/books/core/config/CrawlBiquwoConfig.java b/novel-front/src/main/java/xyz/zinglizingli/books/core/config/CrawlBiquwoConfig.java new file mode 100644 index 00000000..3f83300f --- /dev/null +++ b/novel-front/src/main/java/xyz/zinglizingli/books/core/config/CrawlBiquwoConfig.java @@ -0,0 +1,27 @@ +package xyz.zinglizingli.books.core.config; + +import lombok.extern.slf4j.Slf4j; +import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty; +import org.springframework.boot.context.properties.ConfigurationProperties; +import org.springframework.context.annotation.Bean; +import org.springframework.context.annotation.Configuration; +import xyz.zinglizingli.books.core.crawl.BaseHtmlCrawlSource; +import xyz.zinglizingli.books.core.crawl.BiquCrawlSource; + +/** + * @author 11797 + */ +@Slf4j +@Configuration +public class CrawlBiquwoConfig { + + + @Bean + @ConfigurationProperties(prefix = "biquwo.crawlsource") // prefix值必须是application.yml中对应属性的前缀 + @ConditionalOnProperty(prefix = "crawl.website",name = "type",havingValue = "6") + public BaseHtmlCrawlSource biquwoCrawlSource() { + return new BiquCrawlSource(); + } + + +} diff --git a/novel-front/src/main/resources/application-crawl.yml b/novel-front/src/main/resources/application-crawl.yml index 3b068fe2..b2b8bb40 100644 --- a/novel-front/src/main/resources/application-crawl.yml +++ b/novel-front/src/main/resources/application-crawl.yml @@ -85,4 +85,21 @@ biquge: pic-pattern: \s*([^/]+)

catalog-url-pattern: 查看完整目录 - catalog-pattern:
\s*([^<]+)\s*
\ No newline at end of file + catalog-pattern:
\s*([^<]+)\s*
+ + +biquwo: + crawlsource: + index-url: http://m.biquwo.net + list-page-url: http://m.biquwo.net/sort{0}/0/{1}.html + book-url-pattern: href="/(dudu/\d+/\d+)/" + score-pattern: (\d+\.\d+)分 + book-name-pattern:

([^/]+)

+ author-pattern: 作者:([^/]+)< + status-pattern: 状态:([^/]+) + cat-pattern: 类别:([^/]+) + update-time-pattern: 更新:(\d+-\d+-\d+\s\d+:\d+:\d+) + pic-pattern: ([^<]+)

+ catalog-url-pattern: 查看完整目录 + catalog-pattern: ([^/]+) \ No newline at end of file diff --git a/novel-front/src/main/resources/application.yml b/novel-front/src/main/resources/application.yml index 9e4df5ad..4fcc8789 100644 --- a/novel-front/src/main/resources/application.yml +++ b/novel-front/src/main/resources/application.yml @@ -86,10 +86,10 @@ books: #爬取小说数量 maxNum: 300000 -#爬取的网站名称类型 1:笔趣岛 ,2:笔趣塔,3:顶点小说 ,4:百书斋 更多网站解析中,敬请期待 +#爬取的网站名称类型 1:笔趣岛 ,2:笔趣塔,3:顶点小说 ,4:百书斋,6: 笔趣窝 更多网站解析中,敬请期待 crawl: website: - type: 3 + type: 6 diff --git a/script/crawlbook/application.yml b/script/crawlbook/application.yml index 957e96ce..d4c37c86 100644 --- a/script/crawlbook/application.yml +++ b/script/crawlbook/application.yml @@ -9,8 +9,8 @@ mybatis: mysql: {charset: utf8mb4} books: {lowestScore: 6.0} crawl: - patten: 3 - website: {type: 3} - soft-novel: '1' - manhua: '1' + patten: '1' + website: {type: 6} + soft-novel: '0' + manhua: '0' logging: {config: 'classpath:logback-boot.xml'} diff --git a/script/crawlbook/crawl-book-1.0-SNAPSHOT.jar b/script/crawlbook/crawl-book-1.0-SNAPSHOT.jar index 73e27ca8..492706ec 100644 Binary files a/script/crawlbook/crawl-book-1.0-SNAPSHOT.jar and b/script/crawlbook/crawl-book-1.0-SNAPSHOT.jar differ