Fix: Do not strip spaces from streaming chunks to prevent words from merging

gemini_web2api.py

@@ -575,7 +575,7 @@ def clean_gemini_text(text: str) -> str:
         r'https?://googleusercontent\.com/(?:image_generation_content|image_retrieval|image_collection)[/\w]*\d*',
         '', text
     )
-    return text.strip()
+    return text
 
 
 def _scan_complete_wrb_frames(buf: str) -> list: