refactor: rename the data privacy module

rename data privace module.

.github/workflows/python-publish.yml

@@ -3,7 +3,7 @@
 
 # This workflow uses actions that are not certified by GitHub.
 # They are provided by a third-party and are governed by
-# separate terms of service, privacy policy, and support
+# separate terms of service, data_privacy policy, and support
 # documentation.
 
 name: Upload Python Package

pilot/model/llm_out/proxy_llm.py

@@ -4,12 +4,12 @@ import time
 
 from pilot.configs.config import Config
 
-from pilot.model.proxy.proxy_llms.chatgpt import chatgpt_generate_stream
-from pilot.model.proxy.proxy_llms.bard import bard_generate_stream
-from pilot.model.proxy.proxy_llms.claude import claude_generate_stream
-from pilot.model.proxy.proxy_llms.wenxin import wenxin_generate_stream
-from pilot.model.proxy.proxy_llms.tongyi import tongyi_generate_stream
-from pilot.model.proxy.proxy_llms.gpt4 import gpt4_generate_stream
+from pilot.model.proxy.llms.chatgpt import chatgpt_generate_stream
+from pilot.model.proxy.llms.bard import bard_generate_stream
+from pilot.model.proxy.llms.claude import claude_generate_stream
+from pilot.model.proxy.llms.wenxin import wenxin_generate_stream
+from pilot.model.proxy.llms.tongyi import tongyi_generate_stream
+from pilot.model.proxy.llms.gpt4 import gpt4_generate_stream
 
 CFG = Config()
 

pilot/model/proxy/llms/__init__.py

@@ -1,5 +1,5 @@
 """
     There are several limitations to privatizing large models: high deployment costs and poor performance.
-    In scenarios where data privacy requirements are relatively low, connecting with commercial large models can enable
+    In scenarios where data data_privacy requirements are relatively low, connecting with commercial large models can enable
     rapid and efficient product implementation with high quality.
 """