long-article-recommend/long-article-recommend-service/src/main/resources/mapper/crawler/CrawlerBaseMapper.xml

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN" "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
<mapper namespace="com.tzld.longarticle.recommend.server.mapper.crawler.CrawlerBaseMapper">

    <insert id="batchInsertAccountCorrelation">
        INSERT INTO account_correlation (date_str, gh_id, account_name, rel_gh_id, rel_account_name, status, correlation)
        VALUES
        <foreach collection="list" item="item" separator=",">
            (#{item.dateStr}, #{item.ghId}, #{item.accountName}, #{item.relGhId}, #{item.relAccountName}, #{item.status}, #{item.correlation})
        </foreach>
    </insert>
    <select id="countGetOffVideos" resultType="java.lang.Integer">
        select count(1) from get_off_videos where publish_time > 1727239202
    </select>
    <select id="pageGetOffVideos" resultType="com.tzld.longarticle.recommend.server.model.dto.GetOffVideos">
        select * from get_off_videos order by video_id limit #{offset}, #{pageSize}
    </select>
    <select id="countLongArticlesRootSourceId" resultType="java.lang.Integer">
        select count(1) from long_articles_root_source_id where requestTime > 1727192229
    </select>
    <select id="pageLongArticlesRootSourceId" resultType="com.tzld.longarticle.recommend.server.model.entity.longArticle.LongArticlesRootSourceId">
        select * from long_articles_root_source_id where requestTime > 1727192229 order by rootSourceId limit #{offset}, #{pageSize}
    </select>
    <select id="countArticleMatchVideos" resultType="java.lang.Integer">
        select count(1) from article_match_videos where DATE(update_time) &gt;= '2024-09-20'
    </select>
    <select id="pageArticleMatchVideos"
            resultType="com.tzld.longarticle.recommend.server.model.dto.ArticleMatchVideos">
        select * from article_match_videos order by video_id
    </select>
    <select id="countLongArticlesVideos" resultType="java.lang.Integer">
        select count(1) from long_articles_video where content_status = 2 and success = 0
    </select>
    <select id="pageLongArticlesVideos"
            resultType="com.tzld.longarticle.recommend.server.model.dto.LongArticlesVideoDTO">
        select * from long_articles_video where content_status = 2 and success = 0 and id > #{id} order by id limit #{pageSize}
    </select>
    <select id="getLongArticlesText" resultType="com.tzld.longarticle.recommend.server.model.entity.longArticle.LongArticlesText">
        select content_id, kimi_title, kimi_summary, kimi_keys from long_articles_video where kimi_summary is not null group by content_id
    </select>

    <select id="getLongArticlesVideo"
            resultType="com.tzld.longarticle.recommend.server.model.dto.LongArticlesVideoDTO">
        select * from long_articles_video where trace_id in
        <foreach collection="traceIds" item="item" open="(" close=")" separator=",">
            #{item}
        </foreach>
    </select>

    <insert id="batchInsertAccountAvgInfo">
        INSERT INTO account_avg_info_v3
        (gh_id, position, update_time, account_name, fans, read_avg, like_avg, status, account_type, account_mode,
        account_source, account_status, business_type, read_rate_avg)
        VALUES
        <foreach collection="list" item="item" separator=",">
            (#{item.ghId}, #{item.position}, #{item.updateTime}, #{item.accountName}, #{item.fans}, #{item.readAvg},
            #{item.likeAvg}, #{item.status}, #{item.accountType}, #{item.accountMode}, #{item.accountSource},
            #{item.accountStatus}, #{item.businessType}, #{item.readRateAvg})
        </foreach>
    </insert>

    <delete id="deleteAccountAvgInfoByGhIdAndUpdateTime">
        delete from account_avg_info_v3 where gh_id = #{ghId} and update_time = #{date}
    </delete>

    <update id="updateAccountAvgInfoStatus">
        update account_avg_info_v3 set status = 0 where gh_id = #{ghId} and update_time != #{date}
    </update>

    <select id="getAllByWxSnIn"
            resultType="com.tzld.longarticle.recommend.server.model.entity.crawler.ArticleDetailInfo">
        select wx_sn, recall_dt, first_level, fission_0, fission_0_head, fission_0_recommend, fission_1, fission_1_head,
        fission_1_recommend, fission_2, fission_2_head, fission_2_recommend
        from long_articles_detail_info
        where wx_sn in
        <foreach collection="wxSnList" item="item" separator="," open="(" close=")">
            #{item}
        </foreach>
    </select>

    <select id="getAllByGhIdIn"
            resultType="com.tzld.longarticle.recommend.server.model.entity.crawler.AccountAvgInfo">
        select gh_id, position, update_time, fans, read_avg, read_avg_ci_upper, open_rate_avg
        from account_avg_info_v3
        where gh_id in
        <foreach collection="ghIdList" item="item" separator="," open="(" close=")">
            #{item}
        </foreach>
    </select>

    <select id="getByGhIdInAndPublishTimestampLessThanAndTypeEquals"
            resultType="com.tzld.longarticle.recommend.server.model.entity.crawler.Article">
        select ghId, accountName, ItemIndex, title, ContentUrl, wx_sn, publish_timestamp, show_view_count
        from official_articles_v2
        where ghId in
        <foreach collection="ghIds" item="item" separator="," open="(" close=")">
            #{item}
        </foreach>
        and publish_timestamp &lt; #{publishTimestamp}
        and type = #{type}
    </select>

    <select id="getNeedResetUserGroupGhIds" resultType="java.lang.String">
        select distinct t.gzh_id
        from (select gzh_id, user_group_id, count(1)
              from article_user_group
              GROUP BY gzh_id, user_group_id having count(1) > 10000) t
    </select>

</mapper>