patch: remove usage of llm, chat model __call__ (#20788)

- `llm(prompt)` -> `llm.invoke(prompt)`
- `llm(prompt=prompt` -> `llm.invoke(prompt)` (same with `messages=`)
- `llm(prompt, callbacks=callbacks)` -> `llm.invoke(prompt,
config={"callbacks": callbacks})`
- `llm(prompt, **kwargs)` -> `llm.invoke(prompt, **kwargs)`

libs/community/langchain_community/llms/chatglm3.py

@@ -106,7 +106,7 @@ class ChatGLM3(LLM):
         Example:
             .. code-block:: python
 
-                response = chatglm_llm("Who are you?")
+                response = chatglm_llm.invoke("Who are you?")
         """
         import httpx