From 5a2c542ce22829c356e46b0b922d53db4ba34c5e Mon Sep 17 00:00:00 2001
From: Kevin Hu <kevinhu.sh@gmail.com>
Date: Tue, 3 Sep 2024 14:30:07 +0800
Subject: [PATCH] make term similarity robust (#2212)

### What problem does this PR solve?


### Type of change

- [x] Performance Improvement
---
 rag/nlp/query.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/rag/nlp/query.py b/rag/nlp/query.py
index a8d2d0b59..81a4bc304 100644
--- a/rag/nlp/query.py
+++ b/rag/nlp/query.py
@@ -106,10 +106,9 @@ class EsQueryer:
                         m) for m in sm]
                 sm = [EsQueryer.subSpecialChar(m) for m in sm if len(m) > 1]
                 sm = [m for m in sm if len(m) > 1]
-                if len(sm) < 2:
-                    sm = []
 
                 keywords.append(re.sub(r"[ \\\"']+", "", tk))
+                keywords.extend(sm)
                 if len(keywords) >= 12: break
 
                 tk_syns = self.syn.lookup(tk)