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/bedrock.py

@@ -829,7 +829,7 @@ class Bedrock(LLM, BedrockBase):
         Example:
             .. code-block:: python
 
-                response = llm("Tell me a joke.")
+                response = llm.invoke("Tell me a joke.")
         """
 
         if self.streaming: