zhaohaipeng hace 8 meses
padre
commit
2389a3a98b

+ 3 - 1
.gitignore

@@ -35,4 +35,6 @@ apollo-cache-dir
 sentinel
 weblog
 logs
-LOG_PATH_IS_UNDEFINED
+LOG_PATH_IS_UNDEFINED
+
+xgboost

+ 3 - 0
ad-engine-commons/src/main/java/com/tzld/piaoquan/ad/engine/commons/score/ScorerUtils.java

@@ -32,6 +32,7 @@ public final class ScorerUtils {
     public static String LR_ROV_SCORE_20240626 = "ad_score_config_20240626.conf";
     public static String LR_ROV_SCORE_20240813 = "ad_score_config_20240813.conf";
 
+    public static String  XGBOOST_SCORE_CONF = "ad_score_config_xgboost.conf";
     public static void warmUp() {
         log.info("scorer warm up ");
         // ScorerUtils.init(BASE_CONF);
@@ -43,6 +44,8 @@ public final class ScorerUtils {
 
         ScorerUtils.init(LR_ROV_SCORE_20240626);
         ScorerUtils.init(LR_ROV_SCORE_20240813);
+
+        ScorerUtils.init(XGBOOST_SCORE_CONF);
     }
 
     private ScorerUtils() {

+ 494 - 1
ad-engine-commons/src/main/java/com/tzld/piaoquan/ad/engine/commons/score/model/XGBoostModel.java

@@ -44,7 +44,500 @@ public class XGBoostModel extends Model {
             "b2_3h_conver*log(view)",
             "b2_3h_conver*ctcvr",
             "b2_6h_ctr",
-            "b2_6h_ctcvr"
+            "b2_6h_ctcvr",
+            "b2_6h_cvr",
+            "b2_6h_conver",
+            "b2_6h_click",
+            "b2_6h_conver*log(view)",
+            "b2_6h_conver*ctcvr",
+            "b2_12h_ctr",
+            "b2_12h_ctcvr",
+            "b2_12h_cvr",
+            "b2_12h_conver",
+            "b2_12h_click",
+            "b2_12h_conver*log(view)",
+            "b2_12h_conver*ctcvr",
+            "b2_1d_ctr",
+            "b2_1d_ctcvr",
+            "b2_1d_cvr",
+            "b2_1d_conver",
+            "b2_1d_click",
+            "b2_1d_conver*log(view)",
+            "b2_1d_conver*ctcvr",
+            "b2_3d_ctr",
+            "b2_3d_ctcvr",
+            "b2_3d_cvr",
+            "b2_3d_conver",
+            "b2_3d_click",
+            "b2_3d_conver*log(view)",
+            "b2_3d_conver*ctcvr",
+            "b2_7d_ctr",
+            "b2_7d_ctcvr",
+            "b2_7d_cvr",
+            "b2_7d_conver",
+            "b2_7d_click",
+            "b2_7d_conver*log(view)",
+            "b2_7d_conver*ctcvr",
+            "b2_yesterday_ctr",
+            "b2_yesterday_ctcvr",
+            "b2_yesterday_cvr",
+            "b2_yesterday_conver",
+            "b2_yesterday_click",
+            "b2_yesterday_conver*log(view)",
+            "b2_yesterday_conver*ctcvr",
+            "b2_today_ctr",
+            "b2_today_ctcvr",
+            "b2_today_cvr",
+            "b2_today_conver",
+            "b2_today_click",
+            "b2_today_conver*log(view)",
+            "b2_today_conver*ctcvr",
+            "b3_1h_ctr",
+            "b3_1h_ctcvr",
+            "b3_1h_cvr",
+            "b3_1h_conver",
+            "b3_1h_click",
+            "b3_1h_conver*log(view)",
+            "b3_1h_conver*ctcvr",
+            "b3_2h_ctr",
+            "b3_2h_ctcvr",
+            "b3_2h_cvr",
+            "b3_2h_conver",
+            "b3_2h_click",
+            "b3_2h_conver*log(view)",
+            "b3_2h_conver*ctcvr",
+            "b3_3h_ctr",
+            "b3_3h_ctcvr",
+            "b3_3h_cvr",
+            "b3_3h_conver",
+            "b3_3h_click",
+            "b3_3h_conver*log(view)",
+            "b3_3h_conver*ctcvr",
+            "b3_6h_ctr",
+            "b3_6h_ctcvr",
+            "b3_6h_cvr",
+            "b3_6h_conver",
+            "b3_6h_click",
+            "b3_6h_conver*log(view)",
+            "b3_6h_conver*ctcvr",
+            "b3_12h_ctr",
+            "b3_12h_ctcvr",
+            "b3_12h_cvr",
+            "b3_12h_conver",
+            "b3_12h_click",
+            "b3_12h_conver*log(view)",
+            "b3_12h_conver*ctcvr",
+            "b3_1d_ctr",
+            "b3_1d_ctcvr",
+            "b3_1d_cvr",
+            "b3_1d_conver",
+            "b3_1d_click",
+            "b3_1d_conver*log(view)",
+            "b3_1d_conver*ctcvr",
+            "b3_3d_ctr",
+            "b3_3d_ctcvr",
+            "b3_3d_cvr",
+            "b3_3d_conver",
+            "b3_3d_click",
+            "b3_3d_conver*log(view)",
+            "b3_3d_conver*ctcvr",
+            "b3_7d_ctr",
+            "b3_7d_ctcvr",
+            "b3_7d_cvr",
+            "b3_7d_conver",
+            "b3_7d_click",
+            "b3_7d_conver*log(view)",
+            "b3_7d_conver*ctcvr",
+            "b3_yesterday_ctr",
+            "b3_yesterday_ctcvr",
+            "b3_yesterday_cvr",
+            "b3_yesterday_conver",
+            "b3_yesterday_click",
+            "b3_yesterday_conver*log(view)",
+            "b3_yesterday_conver*ctcvr",
+            "b3_today_ctr",
+            "b3_today_ctcvr",
+            "b3_today_cvr",
+            "b3_today_conver",
+            "b3_today_click",
+            "b3_today_conver*log(view)",
+            "b3_today_conver*ctcvr",
+            "b4_1h_ctr",
+            "b4_1h_ctcvr",
+            "b4_1h_cvr",
+            "b4_1h_conver",
+            "b4_1h_click",
+            "b4_1h_conver*log(view)",
+            "b4_1h_conver*ctcvr",
+            "b4_2h_ctr",
+            "b4_2h_ctcvr",
+            "b4_2h_cvr",
+            "b4_2h_conver",
+            "b4_2h_click",
+            "b4_2h_conver*log(view)",
+            "b4_2h_conver*ctcvr",
+            "b4_3h_ctr",
+            "b4_3h_ctcvr",
+            "b4_3h_cvr",
+            "b4_3h_conver",
+            "b4_3h_click",
+            "b4_3h_conver*log(view)",
+            "b4_3h_conver*ctcvr",
+            "b4_6h_ctr",
+            "b4_6h_ctcvr",
+            "b4_6h_cvr",
+            "b4_6h_conver",
+            "b4_6h_click",
+            "b4_6h_conver*log(view)",
+            "b4_6h_conver*ctcvr",
+            "b4_12h_ctr",
+            "b4_12h_ctcvr",
+            "b4_12h_cvr",
+            "b4_12h_conver",
+            "b4_12h_click",
+            "b4_12h_conver*log(view)",
+            "b4_12h_conver*ctcvr",
+            "b4_1d_ctr",
+            "b4_1d_ctcvr",
+            "b4_1d_cvr",
+            "b4_1d_conver",
+            "b4_1d_click",
+            "b4_1d_conver*log(view)",
+            "b4_1d_conver*ctcvr",
+            "b4_3d_ctr",
+            "b4_3d_ctcvr",
+            "b4_3d_cvr",
+            "b4_3d_conver",
+            "b4_3d_click",
+            "b4_3d_conver*log(view)",
+            "b4_3d_conver*ctcvr",
+            "b4_7d_ctr",
+            "b4_7d_ctcvr",
+            "b4_7d_cvr",
+            "b4_7d_conver",
+            "b4_7d_click",
+            "b4_7d_conver*log(view)",
+            "b4_7d_conver*ctcvr",
+            "b4_yesterday_ctr",
+            "b4_yesterday_ctcvr",
+            "b4_yesterday_cvr",
+            "b4_yesterday_conver",
+            "b4_yesterday_click",
+            "b4_yesterday_conver*log(view)",
+            "b4_yesterday_conver*ctcvr",
+            "b4_today_ctr",
+            "b4_today_ctcvr",
+            "b4_today_cvr",
+            "b4_today_conver",
+            "b4_today_click",
+            "b4_today_conver*log(view)",
+            "b4_today_conver*ctcvr",
+            "b5_1h_ctr",
+            "b5_1h_ctcvr",
+            "b5_1h_cvr",
+            "b5_1h_conver",
+            "b5_1h_click",
+            "b5_1h_conver*log(view)",
+            "b5_1h_conver*ctcvr",
+            "b5_2h_ctr",
+            "b5_2h_ctcvr",
+            "b5_2h_cvr",
+            "b5_2h_conver",
+            "b5_2h_click",
+            "b5_2h_conver*log(view)",
+            "b5_2h_conver*ctcvr",
+            "b5_3h_ctr",
+            "b5_3h_ctcvr",
+            "b5_3h_cvr",
+            "b5_3h_conver",
+            "b5_3h_click",
+            "b5_3h_conver*log(view)",
+            "b5_3h_conver*ctcvr",
+            "b5_6h_ctr",
+            "b5_6h_ctcvr",
+            "b5_6h_cvr",
+            "b5_6h_conver",
+            "b5_6h_click",
+            "b5_6h_conver*log(view)",
+            "b5_6h_conver*ctcvr",
+            "b5_12h_ctr",
+            "b5_12h_ctcvr",
+            "b5_12h_cvr",
+            "b5_12h_conver",
+            "b5_12h_click",
+            "b5_12h_conver*log(view)",
+            "b5_12h_conver*ctcvr",
+            "b5_1d_ctr",
+            "b5_1d_ctcvr",
+            "b5_1d_cvr",
+            "b5_1d_conver",
+            "b5_1d_click",
+            "b5_1d_conver*log(view)",
+            "b5_1d_conver*ctcvr",
+            "b5_3d_ctr",
+            "b5_3d_ctcvr",
+            "b5_3d_cvr",
+            "b5_3d_conver",
+            "b5_3d_click",
+            "b5_3d_conver*log(view)",
+            "b5_3d_conver*ctcvr",
+            "b5_7d_ctr",
+            "b5_7d_ctcvr",
+            "b5_7d_cvr",
+            "b5_7d_conver",
+            "b5_7d_click",
+            "b5_7d_conver*log(view)",
+            "b5_7d_conver*ctcvr",
+            "b5_yesterday_ctr",
+            "b5_yesterday_ctcvr",
+            "b5_yesterday_cvr",
+            "b5_yesterday_conver",
+            "b5_yesterday_click",
+            "b5_yesterday_conver*log(view)",
+            "b5_yesterday_conver*ctcvr",
+            "b5_today_ctr",
+            "b5_today_ctcvr",
+            "b5_today_cvr",
+            "b5_today_conver",
+            "b5_today_click",
+            "b5_today_conver*log(view)",
+            "b5_today_conver*ctcvr",
+            "b8_1h_ctr",
+            "b8_1h_ctcvr",
+            "b8_1h_cvr",
+            "b8_1h_conver",
+            "b8_1h_click",
+            "b8_1h_conver*log(view)",
+            "b8_1h_conver*ctcvr",
+            "b8_2h_ctr",
+            "b8_2h_ctcvr",
+            "b8_2h_cvr",
+            "b8_2h_conver",
+            "b8_2h_click",
+            "b8_2h_conver*log(view)",
+            "b8_2h_conver*ctcvr",
+            "b8_3h_ctr",
+            "b8_3h_ctcvr",
+            "b8_3h_cvr",
+            "b8_3h_conver",
+            "b8_3h_click",
+            "b8_3h_conver*log(view)",
+            "b8_3h_conver*ctcvr",
+            "b8_6h_ctr",
+            "b8_6h_ctcvr",
+            "b8_6h_cvr",
+            "b8_6h_conver",
+            "b8_6h_click",
+            "b8_6h_conver*log(view)",
+            "b8_6h_conver*ctcvr",
+            "b8_12h_ctr",
+            "b8_12h_ctcvr",
+            "b8_12h_cvr",
+            "b8_12h_conver",
+            "b8_12h_click",
+            "b8_12h_conver*log(view)",
+            "b8_12h_conver*ctcvr",
+            "b8_1d_ctr",
+            "b8_1d_ctcvr",
+            "b8_1d_cvr",
+            "b8_1d_conver",
+            "b8_1d_click",
+            "b8_1d_conver*log(view)",
+            "b8_1d_conver*ctcvr",
+            "b8_3d_ctr",
+            "b8_3d_ctcvr",
+            "b8_3d_cvr",
+            "b8_3d_conver",
+            "b8_3d_click",
+            "b8_3d_conver*log(view)",
+            "b8_3d_conver*ctcvr",
+            "b8_7d_ctr",
+            "b8_7d_ctcvr",
+            "b8_7d_cvr",
+            "b8_7d_conver",
+            "b8_7d_click",
+            "b8_7d_conver*log(view)",
+            "b8_7d_conver*ctcvr",
+            "b8_yesterday_ctr",
+            "b8_yesterday_ctcvr",
+            "b8_yesterday_cvr",
+            "b8_yesterday_conver",
+            "b8_yesterday_click",
+            "b8_yesterday_conver*log(view)",
+            "b8_yesterday_conver*ctcvr",
+            "b8_today_ctr",
+            "b8_today_ctcvr",
+            "b8_today_cvr",
+            "b8_today_conver",
+            "b8_today_click",
+            "b8_today_conver*log(view)",
+            "b8_today_conver*ctcvr",
+            "b9_1h_ctr",
+            "b9_1h_ctcvr",
+            "b9_1h_cvr",
+            "b9_1h_conver",
+            "b9_1h_click",
+            "b9_1h_conver*log(view)",
+            "b9_1h_conver*ctcvr",
+            "b9_2h_ctr",
+            "b9_2h_ctcvr",
+            "b9_2h_cvr",
+            "b9_2h_conver",
+            "b9_2h_click",
+            "b9_2h_conver*log(view)",
+            "b9_2h_conver*ctcvr",
+            "b9_3h_ctr",
+            "b9_3h_ctcvr",
+            "b9_3h_cvr",
+            "b9_3h_conver",
+            "b9_3h_click",
+            "b9_3h_conver*log(view)",
+            "b9_3h_conver*ctcvr",
+            "b9_6h_ctr",
+            "b9_6h_ctcvr",
+            "b9_6h_cvr",
+            "b9_6h_conver",
+            "b9_6h_click",
+            "b9_6h_conver*log(view)",
+            "b9_6h_conver*ctcvr",
+            "b9_12h_ctr",
+            "b9_12h_ctcvr",
+            "b9_12h_cvr",
+            "b9_12h_conver",
+            "b9_12h_click",
+            "b9_12h_conver*log(view)",
+            "b9_12h_conver*ctcvr",
+            "b9_1d_ctr",
+            "b9_1d_ctcvr",
+            "b9_1d_cvr",
+            "b9_1d_conver",
+            "b9_1d_click",
+            "b9_1d_conver*log(view)",
+            "b9_1d_conver*ctcvr",
+            "b9_3d_ctr",
+            "b9_3d_ctcvr",
+            "b9_3d_cvr",
+            "b9_3d_conver",
+            "b9_3d_click",
+            "b9_3d_conver*log(view)",
+            "b9_3d_conver*ctcvr",
+            "b9_7d_ctr",
+            "b9_7d_ctcvr",
+            "b9_7d_cvr",
+            "b9_7d_conver",
+            "b9_7d_click",
+            "b9_7d_conver*log(view)",
+            "b9_7d_conver*ctcvr",
+            "b9_yesterday_ctr",
+            "b9_yesterday_ctcvr",
+            "b9_yesterday_cvr",
+            "b9_yesterday_conver",
+            "b9_yesterday_click",
+            "b9_yesterday_conver*log(view)",
+            "b9_yesterday_conver*ctcvr",
+            "b9_today_ctr",
+            "b9_today_ctcvr",
+            "b9_today_cvr",
+            "b9_today_conver",
+            "b9_today_click",
+            "b9_today_conver*log(view)",
+            "b9_today_conver*ctcvr",
+            "b6_7d_ctr",
+            "b6_7d_ctcvr",
+            "b6_7d_cvr",
+            "b6_7d_conver",
+            "b6_7d_click",
+            "b6_7d_conver*log(view)",
+            "b6_7d_conver*ctcvr",
+            "b6_14d_ctr",
+            "b6_14d_ctcvr",
+            "b6_14d_cvr",
+            "b6_14d_conver",
+            "b6_14d_click",
+            "b6_14d_conver*log(view)",
+            "b6_14d_conver*ctcvr",
+            "b7_7d_ctr",
+            "b7_7d_ctcvr",
+            "b7_7d_cvr",
+            "b7_7d_conver",
+            "b7_7d_click",
+            "b7_7d_conver*log(view)",
+            "b7_7d_conver*ctcvr",
+            "b7_14d_ctr",
+            "b7_14d_ctcvr",
+            "b7_14d_cvr",
+            "b7_14d_conver",
+            "b7_14d_click",
+            "b7_14d_conver*log(view)",
+            "b7_14d_conver*ctcvr",
+            "viewAll",
+            "clickAll",
+            "converAll",
+            "incomeAll",
+            "ctr_all",
+            "ctcvr_all",
+            "cvr_all",
+            "timediff_view",
+            "timediff_click",
+            "timediff_conver",
+            "actionstatic_view",
+            "actionstatic_click",
+            "actionstatic_conver",
+            "actionstatic_income",
+            "actionstatic_ctr",
+            "actionstatic_ctcvr",
+            "actionstatic_cvr",
+            "e1_tags_3d_matchnum",
+            "e1_tags_3d_maxscore",
+            "e1_tags_3d_avgscore",
+            "e1_tags_7d_matchnum",
+            "e1_tags_7d_maxscore",
+            "e1_tags_7d_avgscore",
+            "e1_tags_14d_matchnum",
+            "e1_tags_14d_maxscore",
+            "e1_tags_14d_avgscore",
+            "e2_tags_3d_matchnum",
+            "e2_tags_3d_maxscore",
+            "e2_tags_3d_avgscore",
+            "e2_tags_7d_matchnum",
+            "e2_tags_7d_maxscore",
+            "e2_tags_7d_avgscore",
+            "e2_tags_14d_matchnum",
+            "e2_tags_14d_maxscore",
+            "e2_tags_14d_avgscore",
+            "d1_feature_3h_ctr",
+            "d1_feature_3h_ctcvr",
+            "d1_feature_3h_cvr",
+            "d1_feature_3h_conver",
+            "d1_feature_6h_ctr",
+            "d1_feature_6h_ctcvr",
+            "d1_feature_6h_cvr",
+            "d1_feature_6h_conver",
+            "d1_feature_12h_ctr",
+            "d1_feature_12h_ctcvr",
+            "d1_feature_12h_cvr",
+            "d1_feature_12h_conver",
+            "d1_feature_1d_ctr",
+            "d1_feature_1d_ctcvr",
+            "d1_feature_1d_cvr",
+            "d1_feature_1d_conver",
+            "d1_feature_3d_ctr",
+            "d1_feature_3d_ctcvr",
+            "d1_feature_3d_cvr",
+            "d1_feature_3d_conver",
+            "d1_feature_7d_ctr",
+            "d1_feature_7d_ctcvr",
+            "d1_feature_7d_cvr",
+            "d1_feature_7d_conver",
+            "vid_rank_ctr_1d",
+            "vid_rank_ctr_3d",
+            "vid_rank_ctr_7d",
+            "vid_rank_ctr_14d",
+            "vid_rank_ctcvr_1d",
+            "vid_rank_ctcvr_3d",
+            "vid_rank_ctcvr_7d",
+            "vid_rank_ctcvr_14d",
+            "ctitle_vtitle_similarity"
     };
 
     @Override

+ 8 - 0
ad-engine-server/src/main/resources/ad_score_config_xgboost.conf

@@ -0,0 +1,8 @@
+scorer-config = {
+  lr-rov-score-config = {
+    scorer-name = "com.tzld.piaoquan.ad.engine.service.score.XGBoostScorer"
+    scorer-priority = 99
+    model-path = "zhangbo/model_xgb_1000.tar.gz"
+  }
+
+}

+ 1 - 1
ad-engine-server/src/main/resources/application-dev.yml

@@ -4,7 +4,7 @@ server:
 eureka:
   client:
     serviceUrl:
-      defaultZone: http://deveureka-internal.piaoquantv.com/eureka/
+      defaultZone: http://127.0.0.1:7000/eureka/
 
 datalog: .
 

+ 1 - 1
ad-engine-service/src/main/java/com/tzld/piaoquan/ad/engine/service/score/RankService680.java

@@ -124,7 +124,7 @@ public class RankService680 {
         }
 
         // 打分排序
-        List<AdRankItem> result = ScorerUtils.getScorerPipeline(ScorerUtils.LR_ROV_SCORE_20240626)
+        List<AdRankItem> result = ScorerUtils.getScorerPipeline(ScorerUtils.XGBOOST_SCORE_CONF)
                 .scoring(sceneFeatureMap, userFeatureMap, adRankItems);
         for (AdRankItem item : result) {
             item.setScore(item.getLrScore() * item.getCpa());