feat: support xinference rerank model (#1466)

### What problem does this PR solve?

support xinference rerank model
#1455 

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

rag/llm/__init__.py

@@ -68,4 +68,5 @@ RerankModel = {
     "BAAI": DefaultRerank,
     "Jina": JinaRerank,
     "Youdao": YoudaoRerank,
+    "Xinference": XInferenceRerank
 }

rag/llm/rerank_model.py

@@ -136,4 +136,22 @@ class YoudaoRerank(DefaultRerank):
             else: res.extend(scores)
         return np.array(res), token_count
     
+class XInferenceRerank(Base):
+    def __init__(self,model_name="",base_url=""):
+        self.model_name=model_name
+        self.base_url=base_url
+        self.headers = {
+            "Content-Type": "application/json",
+            "accept": "application/json"
+        }
 
+    def similarity(self, query: str, texts: list):
+        data = {
+            "model":self.model_name,
+            "query":query,
+            "return_documents": "true",
+            "return_len": "true",
+            "documents":texts
+        }
+        res = requests.post(self.base_url, headers=self.headers, json=data).json()
+        return np.array([d["relevance_score"] for d in res["results"]]),res["tokens"]["input_tokens"]+res["tokens"]["output_tokens"]