add support for Gemini (#1465)

### What problem does this PR solve?

#1036

### Type of change

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

Co-authored-by: Zhedong Cen <cenzhedong2@126.com>

rag/llm/cv_model.py

@@ -203,6 +203,29 @@ class XinferenceCV(Base):
         )
         return res.choices[0].message.content.strip(), res.usage.total_tokens
 
+class GeminiCV(Base):
+    def __init__(self, key, model_name="gemini-1.0-pro-vision-latest", lang="Chinese", **kwargs):
+        from google.generativeai import client,GenerativeModel 
+        client.configure(api_key=key)
+        _client = client.get_default_generative_client()
+        self.model_name = model_name
+        self.model = GenerativeModel(model_name=self.model_name)
+        self.model._client = _client
+        self.lang = lang 
+    
+    def describe(self, image, max_tokens=2048):
+        from PIL.Image import open
+        gen_config = {'max_output_tokens':max_tokens}
+        prompt = "请用中文详细描述一下图中的内容，比如时间，地点，人物，事情，人物心情等，如果有数据请提取出数据。" if self.lang.lower() == "chinese" else \
+            "Please describe the content of this picture, like where, when, who, what happen. If it has number data, please extract them out."
+        b64 = self.image2base64(image) 
+        img = open(BytesIO(base64.b64decode(b64))) 
+        input = [prompt,img]
+        res = self.model.generate_content(
+            input,
+            generation_config=gen_config,
+        )
+        return res.text,res.usage_metadata.total_token_count
 
 class LocalCV(Base):
     def __init__(self, key, model_name="glm-4v", lang="Chinese", **kwargs):