fix: 修复正文解析失败后重复采集的问题，加入错误次数统计

src/main/kotlin/com/lanyuanxiaoyao/digtal/market/Entity.kt

@@ -1,11 +1,11 @@
 package com.lanyuanxiaoyao.digtal.market
 
+import com.blinkfox.fenix.jpa.FenixJpaRepository
 import com.blinkfox.fenix.specification.FenixJpaSpecificationExecutor
 import jakarta.persistence.Entity
 import jakarta.persistence.Id
 import jakarta.persistence.Lob
 import java.util.Date
-import org.springframework.data.jpa.repository.JpaRepository
 import org.springframework.data.jpa.repository.Modifying
 import org.springframework.data.jpa.repository.Query
 import org.springframework.data.repository.query.Param
@@ -27,10 +27,15 @@ class Article(
     var tags: String?,
     var createTime: Date?,
     var pushed: Boolean?,
-)
+    var error: Int?,
+) {
+    override fun toString(): String {
+        return "Article(id='$id', code='$code', url='$url', title=$title, author=$author, category=$category, text=$text, html=$html, subtitle=$subtitle, description=$description, tags=$tags, createTime=$createTime, pushed=$pushed, error=$error)"
+    }
+}
 
 @Repository
-interface ArticleRepository : JpaRepository<Article, String>, FenixJpaSpecificationExecutor<Article> {
+interface ArticleRepository : FenixJpaRepository<Article, String>, FenixJpaSpecificationExecutor<Article> {
     fun findAllByDescriptionIsNullAndTextIsNotNull(): List<Article>
     fun findAllByHtmlIsNotNull(): List<Article>
 
@@ -47,11 +52,14 @@ interface ArticleRepository : JpaRepository<Article, String>, FenixJpaSpecificat
     @Query("update Article article set article.pushed = :pushed where article.id = :id")
     fun updatePushedById(@Param("id") id: String, @Param("pushed") pushed: Boolean)
 
-    @Query("select article.id from Article article where article.description is not null and article.text is not null and article.text <> ''")
+    @Query("select article.id from Article article where (article.error is not null and article.error > 3) or (article.description is not null and article.text is not null and article.text <> '')")
     fun findAllId(): List<String>
 
     @Query("select new com.lanyuanxiaoyao.digtal.market.CountGroupByString(article.code, count(article.code)) from Article article group by article.code")
     fun countGroupByCode(): List<CountGroupByString>
+
+    @Query("select article.error from Article article where article.id = :id")
+    fun findErrorById(@Param("id") id: String): Int?
 }
 
 data class CountGroupByString(

src/main/kotlin/com/lanyuanxiaoyao/digtal/market/Helper.kt

@@ -148,6 +148,7 @@ fun parseArticle(management: Management, code: String, url: String, hash: String
             tags = null,
             createTime = createTime,
             pushed = false,
+            error = null,
         )
     } ?: return null
 }

src/main/kotlin/com/lanyuanxiaoyao/digtal/market/runner/NewsRunner.kt

@@ -87,6 +87,11 @@ class NewsRunner : Runner {
                                             article.subtitle = triple?.subtitle
                                             article.description = triple?.description
                                             article.tags = triple?.tags
+
+                                            if (triple == null) {
+                                                val error = articleRepository.findErrorById(link.hash)
+                                                article.error = error?.plus(1) ?: 1
+                                            }
                                         }
 
                                         article.category = keyword