get order data

eda/smart_rank/tag3_update_user_portrait_offline.py

@@ -8,8 +8,8 @@ import pymysql
 from pyspark import SparkConf
 from pyspark.sql import SparkSession
 
-from tool import (get_doris_prod, get_redis_client, get_tag3_user_log, get_user_portrait_tag3_with_score, send_email,
-                  write_user_portrait)
+from tool import (get_doris_prod, get_redis_client, get_tag3_user_log, get_tag3_user_order_log, get_user_portrait_tag3_with_score,
+                  send_email, write_user_portrait)
 
 
 # [{'激光': 1.949194898204873}, {'手术': 1.949194898204873}, {'手术': 1.949194898204873}, {'手术': 1.949194898204873, '植发际线': 7.1}]
@@ -350,7 +350,10 @@ if __name__ == "__main__":
     # df[["projects", "business_tags"]]
     # update_tag3_user_portrait(cl_id)
 
-    consume_kafka()
+    cl_id = "867617044159377"
+    print(get_tag3_user_order_log(cl_id))
+
+    # consume_kafka() TODO
     end = datetime.datetime.now()
     print(end - start)
     print("done")

eda/smart_rank/tool.py

@@ -442,6 +442,34 @@ def get_tag3_user_log(cl_id):
         return pd.DataFrame(columns=columns)
 
 
+def 1get_tag3_user_order_log(cl_id):
+    columns = [
+        "log_time", "score_type", "business_tags", "event_cn"
+    ]
+    try:
+        sql = """select log_time, score_type, business_tags, event_cn
+                 from kafka_tag3_log
+                 where cl_id = '{}' and event_cn in ('支付订单', '验证订单')""".format(cl_id)
+        db, cursor = get_doris_prod()
+        cursor.execute(sql)
+        data = list(cursor.fetchall())
+        db.close()
+        cursor.close()
+
+        if data:
+            user_df = pd.DataFrame(data)
+            user_df.columns = columns
+        else:
+            return pd.DataFrame(columns=columns)
+
+        user_df["days_diff_now"] = round((int(time.time()) - user_df["log_time"].astype(float)) / (24 * 60 * 60))
+        user_df["tag_score"] = user_df.apply(lambda x: compute_tag3_score(x), axis=1)
+        user_df["business_tags"] = user_df["business_tags"].fillna("")
+    except Exception as e:
+        print(e)
+        return pd.DataFrame(columns=columns)
+
+
 # CREATE TABLE `user_tag3_portrait` (
 #   `id` int(11) NOT NULL AUTO_INCREMENT,
 #   `date` text NOT NULL,