pass word with tag

suggest-task/src/main/java/com/secoo/so/suggest/task/SuggestTask.java

 package com.secoo.so.suggest.task;
 
 import com.alibaba.fastjson.JSON;
+import com.alibaba.fastjson.JSONObject;
 import com.secoo.so.suggest.config.ConfigUtil;
 import com.secoo.so.suggest.db.DwDataSource;
 import com.secoo.so.suggest.db.ErpDataSource;
@@ -279,7 +280,7 @@ public class SuggestTask {
     }
 
     private static void mergeKeywordTag(ConcurrentHashMap<String, EsSuggestKeywordInfo> esSuggestKeywordMap) {
-        int maxCount = 100;
+        int maxCount = 10000;
         List<String> keyList = new ArrayList<>(esSuggestKeywordMap.keySet());
         Collections.sort(keyList);  // 按照字符排序，确保扩展词都在本词后面
         for(int i = 0;i< keyList.size(); i++) {
@@ -300,14 +301,30 @@ public class SuggestTask {
                         break;
                     }
                     String fulWord = keyList.get(j);
+                    EsSuggestKeywordInfo tmpSuggest = esSuggestKeywordMap.get(fulWord);
+                    if (StringUtils.isNotBlank(tmpSuggest.getSuggestTags())) {
+                        continue;
+                    }
                     int rightLen = StringUtils.getByteLength(fulWord) - wordLen;
-                    if (fulWord.startsWith(word) && rightLen > 3 && rightLen <= 12) {
-                        String subWord = fulWord.substring(length, length + 1);
-                        if (isEN && StringUtils.isEnStr(subWord)){
-                            continue;
+                    if (fulWord.startsWith(word)) {
+                        if (rightLen > 3 && rightLen <= 14) {
+                            String subWord = fulWord.substring(length, length + 1);
+                            boolean isTShirt = false;
+                            String rightWord = fulWord.substring(length, fulWord.length()).trim();
+                            int realLen = StringUtils.getByteLength(rightWord);
+                            if (rightWord.length() >= 2 && rightWord.toLowerCase().startsWith("t恤")) {
+                                isTShirt = true;
+                            }
+                            if (isEN && (StringUtils.isEnStr(subWord) && !isTShirt)) {
+                                continue;
+                            }
+
+                            if (realLen <= 3 || realLen > 14) {
+                                continue;
+                            }
+                            suggestList.add(esSuggestKeywordMap.get(fulWord));
+                            keyCount++;
                         }
-                        suggestList.add(esSuggestKeywordMap.get(fulWord));
-                        keyCount++;
                     } else {
                         break;
                     }
@@ -339,6 +356,9 @@ public class SuggestTask {
                         sb.append(subWord);
                     }
                     if (sb.length() > 0) {
+                        if ("lv女包".equals(word)) {
+                            log.info("debugLog keyword tag:" + sb.toString());
+                        }
                         suggestInfo.setSuggestTags(sb.toString());
                     }
                 }