ValueError: Prompt missing required variables: {'tools', 'tool_names'} when using third-party models

[Sui-Xing]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain_community.llms import HuggingFaceHub
from transformers import AutoTokenizer, AutoModel, pipeline
from langchain import HuggingFacePipeline
from langchain import PromptTemplate
import os
model_dir = '/home/jupiter/model/AI-ModelScope/Mistral-7B-Instruct-v0___2'
tokenizer = AutoTokenizer.from_pretrained(model_dir, trust_remote_code=True)
model = AutoModel.from_pretrained(model_dir, trust_remote_code=True)
model.to(device)

pipe = pipeline(
    "text2text-generation",
    model=model,
    tokenizer=tokenizer,
    max_length=512,
    temperature=0.8,
    top_p=1,
    repetition_penalty=1.15
)
llm = HuggingFacePipeline(pipeline=pipe)

# using code from https://python.langchain.com/docs/use_cases/sql/prompting/#few-shot-examples


from langchain_community.agent_toolkits import create_sql_agent
from langchain.agents import AgentType
agent = create_sql_agent(
    llm=llm,
    db=db,
    prompt=full_prompt,
    verbose=True,
    # agent_type='openai-tools',
    # agent_type='zero-shot-react-description',
)

Description

When I use a third-party model instead of an OpenAI model to create_sql_agent, it reports an error "ValueError: Prompt missing required variables: {'tools', 'tool_names'} when using third-party models."

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[15], line 3
      1 from langchain_community.agent_toolkits import create_sql_agent
      2 from langchain.agents import AgentType
----> 3 agent = create_sql_agent(
      4     llm=llm,
      5     db=db,
      6     prompt=full_prompt,
      7     verbose=True,
      8     # agent_type='openai-tools',
      9     # agent_type='zero-shot-react-description',
     10 )

File [~/miniconda3/envs/langchain/lib/python3.10/site-packages/langchain_community/agent_toolkits/sql/base.py:182](http://127.0.0.1:8888/lab/tree/mnt/c/Users/Jupiter/Documents/Work/~/miniconda3/envs/langchain/lib/python3.10/site-packages/langchain_community/agent_toolkits/sql/base.py#line=181), in create_sql_agent(llm, toolkit, agent_type, callback_manager, prefix, suffix, format_instructions, input_variables, top_k, max_iterations, max_execution_time, early_stopping_method, verbose, agent_executor_kwargs, extra_tools, db, prompt, **kwargs)
    172         template = "\n\n".join(
    173             [
    174                 react_prompt.PREFIX,
   (...)
    178             ]
    179         )
    180         prompt = PromptTemplate.from_template(template)
    181     agent = RunnableAgent(
--> 182         runnable=create_react_agent(llm, tools, prompt),
    183         input_keys_arg=["input"],
    184         return_keys_arg=["output"],
    185     )
    187 elif agent_type == AgentType.OPENAI_FUNCTIONS:
    188     if prompt is None:

File [~/miniconda3/envs/langchain/lib/python3.10/site-packages/langchain/agents/react/agent.py:114](http://127.0.0.1:8888/lab/tree/mnt/c/Users/Jupiter/Documents/Work/~/miniconda3/envs/langchain/lib/python3.10/site-packages/langchain/agents/react/agent.py#line=113), in create_react_agent(llm, tools, prompt, output_parser, tools_renderer, stop_sequence)
    110 missing_vars = {"tools", "tool_names", "agent_scratchpad"}.difference(
    111     prompt.input_variables
    112 )
    113 if missing_vars:
--> 114     raise ValueError(f"Prompt missing required variables: {missing_vars}")
    116 prompt = prompt.partial(
    117     tools=tools_renderer(list(tools)),
    118     tool_names=", ".join([t.name for t in tools]),
    119 )
    120 if stop_sequence:

ValueError: Prompt missing required variables: {'tools', 'tool_names'}

System Info

OS: Linux
OS Version: #91-Ubuntu SMP Thu Jul 15 19:09:17 UTC 2021
Python Version: 3.10.14 (main, Mar 21 2024, 16:24:04) [GCC 11.2.0]

Package Information

langchain_core: 0.1.40
langchain: 0.1.14
langchain_community: 0.0.31
langsmith: 0.1.40
langchain_experimental: 0.0.56
langchain_openai: 0.1.1
langchain_text_splitters: 0.0.1

Packages not installed (Not Necessarily a Problem)

The following packages were not found:

langgraph
langserve

[dosubot[bot]]

The error you're encountering is due to the prompt missing required variables: tools, tool_names, and agent_scratchpad. These variables are expected to be present in the prompt's input variables. The relevant section of the code that checks for these variables is:

missing_vars = {"tools", "tool_names", "agent_scratchpad"}.difference(
    prompt.input_variables
)
if missing_vars:
    raise ValueError(f"Prompt missing required variables: {missing_vars}")

This code snippet checks if the tools, tool_names, and agent_scratchpad variables are present in the prompt.input_variables. If any of these are missing, it raises the ValueError.

To fix this issue, you need to ensure that the prompt you are using is correctly defined with all the required input variables. Here's an example of how the prompt template should include these variables:

from langchain_core.prompts import PromptTemplate

template = '''Answer the following questions as best you can. You have access to the following tools:

{tools}

Use the following format:

Question: the input question you must answer
Thought: you should always think about what to do
Action: the action to take, should be one of [{tool_names}]
Action Input: the input to the action
Observation: the result of the action
... (this Thought/Action/Action Input/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question

Begin!

Question: {input}
Thought:{agent_scratchpad}'''

prompt = PromptTemplate.from_template(template)

Ensure that the prompt you are using is similar to the example above, with placeholders for tools, tool_names, and agent_scratchpad. This should resolve the ValueError you are encountering.

You can find more details about this in the LangChain repository.

Regarding the create_sql_agent function, it is designed to facilitate the creation of SQL agents within the LangChain framework. The inclusion of create_sql_agent in the __all__ list indicates that it is intended to be a publicly accessible part of the module's API, suggesting its significance in enabling interactions with SQL databases or executing SQL queries through the LangChain's agent-based architecture. However, without the actual implementation details of the create_sql_agent function, it's not possible to provide a more detailed explanation of its exact role or how it operates within the framework. You can find more details about this in the LangChain repository.

The exact implementation of the create_sql_agent function in the langchain_community package can be found in the LangChain repository.

Sources

• libs/langchain/langchain/agents/react/agent.py
• libs/langchain/langchain/agents/agent_toolkits/sql/base.py
• libs/community/langchain_community/agent_toolkits/sql/base.py

About Dosu

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[VaticanCameos99]

Hi, thanks for detailing the specifics. Please can you share how to create and pass the agent_scratchpad?

[why168]

mark