fix: Restore embedding-mode parameter to all examples

- All examples now have --embedding-mode parameter (unified interface benefit)
- Default is 'sentence-transformers' (consistent with original behavior)
- Users can now use OpenAI or MLX embeddings with any data source
- Maintains functional equivalence with original scripts

examples/wechat_rag.py

@@ -25,8 +25,7 @@ class WeChatRAG(BaseRAGExample):
         super().__init__(
             name="WeChat History",
             description="Process and query WeChat chat history with LEANN",
-            default_index_name="wechat_history_magic_test_11Debug_new",  # Match original default
-            include_embedding_mode=False,  # Original wechat_history_reader_leann.py doesn't have embedding_mode
+            default_index_name="wechat_history_magic_test_11Debug_new",
         )
 
     def _add_specific_arguments(self, parser):