ChatOllama .invoke method giving 'NoneType' object is not iterable #28287
-
Checked other resources
Commit to Help
Example Codefrom langchain.prompts import ChatPromptTemplate
from langchain_ollama import ChatOllama
# Initialize the language model
llm = ChatOllama(model="llama3.1", temperature=0)
# Define the prompt template
prompt = ChatPromptTemplate(
input_variables=["input", "response"],
messages=[
(
"system",
"You are a helpful assistant that extracts the address from the user's input and provides it in a structured JSON format with keys 'address_line_1', 'address_line_2', and 'address_line_3'. If the address is incomplete, leave empty strings in the respective lines.",
),
("human", "{input}"),
]
)
# The input query with the address
input_query = "Hello, I am living at 123, Elm Street, Downtown, Springfield, USA, ZIP - 12345."
response = {
'address_line_1': '123, Elm Street, Downtown', # Generic address part
'address_line_2': 'Springfield, USA, ZIP - 12345', # Generic address part
'address_line_3': '', # Empty if not present
'status_code': 200 # Status code if relevant
}
# Setup the chain with the prompt and model
chain = prompt | llm
try:
# Invoke the chain with the input and response
output = chain.invoke({
"input": input_query,
"response": response,
})
# Print the raw model response for debugging
print("Raw output:", output)
# Check if the output contains the expected structure
if output is not None:
# If output is not None, check if it contains the message key
if "message" in output:
print("Output contains message:", output["message"])
else:
print("Output does not contain 'message' key:", output)
else:
print("Output is None. Model response might be malformed or empty.")
except Exception as e:
print(f"Error occurred: {str(e)}")DescriptionI am trying to use ollama from langchain, i am newbie and was trying things out, when i started to do this .invoke to get output i am getting `TypeError: 'NoneType' object is not iterable`` like this, ollama is running and working but getting this error here, i am not sure how to proceed even this small code ` messages = [ ai_msg = llm.invoke(messages) doesnt work. the expected output is `The translation of "I love programming" from English to French is: "J'adore programmer."` System Info
|
Beta Was this translation helpful? Give feedback.
Replies: 6 comments 4 replies
-
|
@pradeep0806 try upgrading the packages to the latest version by running from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate
prompt = ChatPromptTemplate.from_template("You are a helpful assistant that extracts the address from the user's input and provides it in a structured JSON format with keys 'address_line_1', 'address_line_2', and 'address_line_3'. If the address is incomplete, leave empty strings in the respective lines.")
chain = prompt | llm | StrOutputParser() |
Beta Was this translation helpful? Give feedback.
-
|
nope still the same answer my langchain packages are up to date langchain 0.3.7 python 3.12.4 `from langchain.prompts import ChatPromptTemplate llm = ChatOllama(model="llama3.1", temperature=0) prompt = ChatPromptTemplate.from_template("You are a helpful assistant that extracts the address from the user's input and provides it in a structured JSON format with keys 'address_line_1', 'address_line_2', and 'address_line_3'. If the address is incomplete, leave empty strings in the respective lines.") input_query = "Hello, I am living at 123, Elm Street, Downtown, Springfield, USA, ZIP - 12345." chain = prompt | llm | StrOutputParser() try: except Exception as e: |
Beta Was this translation helpful? Give feedback.
-
|
am i calling the chain wrongly or what? function calling using tools are working but not this...not sure what to do |
Beta Was this translation helpful? Give feedback.
-
|
#28281 (comment) |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I also encountered this problem, here is my minimal reproducible example:
pyproject.toml |
Beta Was this translation helpful? Give feedback.
-
|
I fixed the same issue by by adding Here is my version info: |
Beta Was this translation helpful? Give feedback.
-
|
By the way, I just realized the issue only occurs on my Windows machine. It works fine on my Mac, which is running the same langchain-ollama-0.2.0 version and the same Ollama version, 0.4.3. |
Beta Was this translation helpful? Give feedback.
-
|
I encountered the same issue and resolved it similarly: def _get_tool_calls_from_response(
response: Mapping[str, Any],
) -> List[ToolCall]:
"""Get tool calls from ollama response."""
tool_calls = []
if "message" in response:
if "tool_calls" in response["message"]:
try:
for tc in response["message"]["tool_calls"]:
tool_calls.append(
tool_call(
id=str(uuid4()),
name=tc["function"]["name"],
args=tc["function"]["arguments"],
)
)
except TypeError:
pass # response["message"]["tool_calls"] is not iterable
return tool_calls |
Beta Was this translation helpful? Give feedback.
-
|
folks the issue is resolved when i upgraded the ollama package from 0.4.0 to 0.4.1. thanks for all the replies, never really thought I would get response. |
Beta Was this translation helpful? Give feedback.
-
|
It seems to be a version issue, try it out: |
Beta Was this translation helpful? Give feedback.
folks the issue is resolved when i upgraded the ollama package from 0.4.0 to 0.4.1. thanks for all the replies, never really thought I would get response.