PaiEasChatEndpoint._call_eas should return bytes type instead of str type

[marinelay]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

from langchain_community.chat_models import PaiEasChatEndpoint
from langchain_core.prompts import ChatPromptTemplate

prompt = ChatPromptTemplate.from_template("tell me a short joke about {topic}")
prompt_value = prompt.invoke("test_message")

eas_chat_endpoint = PaiEasChatEndpoint(
    eas_service_url="your_service_url",
    eas_service_token="your_service_token"
)
eas_chat_endpoint._call(prompt_value.messages)

It is just example code that has a potential error, so I'll explain why it's a possible type error in the description.

Error Message and Stack Trace (if applicable)

No response

Description

langchain/libs/community/langchain_community/chat_models/pai_eas_endpoint.py

Lines 193 to 230 in b66a4f4

	def _call(
	self,
	messages: List[BaseMessage],
	stop: Optional[List[str]] = None,
	run_manager: Optional[CallbackManagerForLLMRun] = None,
	**kwargs: Any,
	) -> str:
	params = self._invocation_params(stop, **kwargs)
	request_payload = self.format_request_payload(messages, **params)
	response_payload = self._call_eas(request_payload)
	generated_text = self._format_response_payload(response_payload, params["stop"])
	if run_manager:
	run_manager.on_llm_new_token(generated_text)
	return generated_text
	def _call_eas(self, query_body: dict) -> Any:
	"""Generate text from the eas service."""
	headers = {
	"Content-Type": "application/json",
	"Accept": "application/json",
	"Authorization": f"{self.eas_service_token}",
	}
	# make request
	response = requests.post(
	self.eas_service_url, headers=headers, json=query_body, timeout=self.timeout
	)
	if response.status_code != 200:
	raise Exception(
	f"Request failed with status code {response.status_code}"
	f" and message {response.text}"
	)
	return response.text

In the method PaiEasChatEndpoint._call at line 203, self._call_eas(request_payload) returns str type because it returns response.text at line 230 (https://requests.readthedocs.io/en/latest/api/#requests.Response.text).
Then, first argument of _format_response_payload can be str type at line 204, resulting in type mismatch.

langchain/libs/community/langchain_community/chat_models/pai_eas_endpoint.py

Lines 167 to 179 in b66a4f4

	def _format_response_payload(
	self, output: bytes, stop_sequences: Optional[List[str]]
	) -> str:
	"""Formats response"""
	try:
	text = json.loads(output)["response"]
	if stop_sequences:
	text = enforce_stop_tokens(text, stop_sequences)
	return text
	except Exception as e:
	if isinstance(e, json.decoder.JSONDecodeError):
	return output.decode("utf-8")
	raise e

Moreover, if JSONDecodeError occurs in this method, then attribute error occurs at line 178 (AttributeError: 'str' object has no attribute 'decode') because output variable is str type.
I think PaiEasChatEndpoint._call_eas should be fixed to return bytes type.
If I'm mistaken, I'd appreciate it if you could let me know.

System Info

System Information

OS: Linux
OS Version: #190-Ubuntu SMP Fri Feb 2 23:24:22 UTC 2024
Python Version: 3.9.18 (main, Feb 13 2024, 14:37:08)
[GCC 9.4.0]

Package Information

langchain_core: 0.1.42
langchain: 0.1.16
langchain_community: 0.0.32
langsmith: 0.1.47
langchain_text_splitters: 0.0.1

Packages not installed (Not Necessarily a Problem)

The following packages were not found:

langgraph
langserve