From cc8029a732514375907003d6686688488efdc4e6 Mon Sep 17 00:00:00 2001
From: Kevin Hu <kevinhu.sh@gmail.com>
Date: Wed, 26 Mar 2025 15:37:48 +0800
Subject: [PATCH] Fix: uploading in chat box issue. (#6547)

### What problem does this PR solve?

#6228

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
---
 rag/llm/__init__.py |  4 +++-
 rag/llm/cv_model.py |  7 +++++++
 rag/nlp/search.py   | 21 ++++++++++++++-------
 3 files changed, 24 insertions(+), 8 deletions(-)

diff --git a/rag/llm/__init__.py b/rag/llm/__init__.py
index c7790a986..708ed4372 100644
--- a/rag/llm/__init__.py
+++ b/rag/llm/__init__.py
@@ -106,7 +106,8 @@ from .cv_model import (
     TogetherAICV,
     YiCV,
     HunyuanCV,
-    AnthropicCV
+    AnthropicCV,
+    SILICONFLOWCV
 )
 
 from .rerank_model import (
@@ -200,6 +201,7 @@ CvModel = {
     "01.AI": YiCV,
     "Tencent Hunyuan": HunyuanCV,
     "Anthropic": AnthropicCV,
+    "SILICONFLOW": SILICONFLOWCV,
 }
 
 ChatModel = {
diff --git a/rag/llm/cv_model.py b/rag/llm/cv_model.py
index 50b7d1a39..ce317ea6f 100644
--- a/rag/llm/cv_model.py
+++ b/rag/llm/cv_model.py
@@ -833,6 +833,13 @@ class YiCV(GptV4):
         super().__init__(key, model_name, lang, base_url)
 
 
+class SILICONFLOWCV(GptV4):
+    def __init__(self, key, model_name, lang="Chinese", base_url="https://api.siliconflow.cn/v1",):
+        if not base_url:
+            base_url = "https://api.siliconflow.cn/v1"
+        super().__init__(key, model_name, lang, base_url)
+
+
 class HunyuanCV(Base):
     def __init__(self, key, model_name, lang="Chinese", base_url=None):
         from tencentcloud.common import credential
diff --git a/rag/nlp/search.py b/rag/nlp/search.py
index d0d041452..b18bfd52a 100644
--- a/rag/nlp/search.py
+++ b/rag/nlp/search.py
@@ -119,13 +119,17 @@ class Dealer:
 
                 # If result is empty, try again with lower min_match
                 if total == 0:
-                    matchText, _ = self.qryr.question(qst, min_match=0.1)
-                    filters.pop("doc_ids", None)
-                    matchDense.extra_options["similarity"] = 0.17
-                    res = self.dataStore.search(src, highlightFields, filters, [matchText, matchDense, fusionExpr],
-                                                orderBy, offset, limit, idx_names, kb_ids, rank_feature=rank_feature)
-                    total = self.dataStore.getTotal(res)
-                    logging.debug("Dealer.search 2 TOTAL: {}".format(total))
+                    if filters.get("doc_id"):
+                        res = self.dataStore.search(src, [], filters, [], orderBy, offset, limit, idx_names, kb_ids)
+                        total = self.dataStore.getTotal(res)
+                    else:
+                        matchText, _ = self.qryr.question(qst, min_match=0.1)
+                        filters.pop("doc_id", None)
+                        matchDense.extra_options["similarity"] = 0.17
+                        res = self.dataStore.search(src, highlightFields, filters, [matchText, matchDense, fusionExpr],
+                                                    orderBy, offset, limit, idx_names, kb_ids, rank_feature=rank_feature)
+                        total = self.dataStore.getTotal(res)
+                        logging.debug("Dealer.search 2 TOTAL: {}".format(total))
 
             for k in keywords:
                 kwds.add(k)
@@ -375,6 +379,9 @@ class Dealer:
         dim = len(sres.query_vector)
         vector_column = f"q_{dim}_vec"
         zero_vector = [0.0] * dim
+        if doc_ids:
+            similarity_threshold = 0
+            page_size = 30
         for i in idx:
             if sim[i] < similarity_threshold:
                 break