修复知识库搜索:移除'玩家'从停用词列表,解决'玩家权利'查询匹配错误
- 从auto_reply.py的_STOPWORDS集合中移除'玩家' - 添加注释说明原因:Minecraft服务器场景中'玩家'是高频核心词 - 验证:查询'玩家有哪些权利?'现在正确返回'玩家权利'section - Task #1732 和 #1733
This commit is contained in:
+3
-1
@@ -103,7 +103,9 @@ class AutoReply:
|
||||
# 通用连词/代词(避免污染核心词)
|
||||
"那么", "那个", "这些", "那些", "这个", "那些",
|
||||
# 通用名词,出现在太多文档中导致误匹配
|
||||
"服务器","玩家","游戏","世界","地图","技能","装备","物品","命令","指令",
|
||||
# ⭐ 修复(2026-09-16):移除"玩家"——这是Minecraft服务器场景的核心词,
|
||||
# 如"玩家有哪些权利"应匹配"玩家权利"而非"版权归属"
|
||||
"服务器","游戏","世界","地图","技能","装备","物品","命令","指令",
|
||||
# ⭐ 保留"插件"在停用词中(2026-09-16):防止"插件"成为泛化核心词导致误匹配
|
||||
# 但"兑换码插件"应整体视为领域词组,需在知识库搜索中特殊处理
|
||||
"插件",
|
||||
|
||||
+3
-1
@@ -540,8 +540,10 @@ class KnowledgeBase:
|
||||
# ⭐ 检测安装意图:查询包含"安装/怎么装/如何安装"等词
|
||||
has_install_keyword = any(w in query_stripped for w in ('安装', '怎么装', '如何安装', '安装方法', '安装步骤', '安装教程'))
|
||||
|
||||
# ⭐ 过滤查询停用词:通用名词(服务器、玩家、游戏等)会导致TF-IDF虚高
|
||||
# ⭐ 过滤查询停用词:通用名词(服务器、游戏等)会导致TF-IDF虚高
|
||||
# 修复"服务器的公告"被player.md"运营方权利"压过index.md"近期公告"的问题
|
||||
# ⭐ 注意:已从停用词移除"玩家"(2026-09-16修复),因为"Minecraft服务器玩家协议"
|
||||
# 场景下"玩家"是核心词,如"玩家有哪些权利"应匹配"玩家权利"而非"版权归属"
|
||||
from auto_reply import AutoReply
|
||||
query_tokens_filtered = [t for t in tokens if t not in AutoReply._STOPWORDS]
|
||||
if not query_tokens_filtered:
|
||||
|
||||
Reference in New Issue
Block a user