Support Xinference (#320)

### What problem does this PR solve?

Issue link:#299

### Type of change

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

rag/llm/embedding_model.py

@@ -170,3 +170,20 @@ class OllamaEmbed(Base):
         res = self.client.embeddings(prompt=text,
                                             model=self.model_name)
         return np.array(res["embedding"]), 128
+
+
+class XinferenceEmbed(Base):
+    def __init__(self, key, model_name="", base_url=""):
+        self.client = OpenAI(api_key="xxx", base_url=base_url)
+        self.model_name = model_name
+
+    def encode(self, texts: list, batch_size=32):
+        res = self.client.embeddings.create(input=texts,
+                                            model=self.model_name)
+        return np.array([d.embedding for d in res.data]
+                        ), res.usage.total_tokens
+
+    def encode_queries(self, text):
+        res = self.client.embeddings.create(input=[text],
+                                            model=self.model_name)
+        return np.array(res.data[0].embedding), res.usage.total_tokens