few shot key error

[ko-devHong]

I set up the samples correctly, but I get an error.

This is my code

from langchain.prompts import PromptTemplate
import chromadb
from langchain_chroma import Chroma
from langchain.prompts.example_selector import SemanticSimilarityExampleSelector
from ...langchain_module.embedding import get_bedrock_embedding
from ..answer import DEFAULT_FALLBACK_ANSWER


prompt_examples = [
    {
        "title": "aaaa",
        "contents": """
           aaaa
        """,
        "images": "image_url_13",
        "format": DEFAULT_FALLBACK_ANSWER
    },
    {
        "title": "Ibbbbby",
        "contents": """
            Ibbbbby
        """,
        "images": "image_url_14",
        "format": DEFAULT_FALLBACK_ANSWER
    },
    {
        "title": "cccccc",
        "contents": """
           cccccc
        """,
        "images": "image_url_15",
        "format": DEFAULT_FALLBACK_ANSWER
    }
]

persistent_client = chromadb.PersistentClient(path="chroma/_prompt")
embedding = get_bedrock_embedding()
chroma = Chroma(
    client=persistent_client,
    collection_name="collecton_name",
    embedding_function=embedding,
)

example_prompt = PromptTemplate(
    input_variables=["title", "contents", "images", "format"],
    template="Title: {title}\nContents: {contents}\nImages: {images}\nFormat: {format}"
)


examples_selector = SemanticSimilarityExampleSelector.from_examples(
    prompt_examples,
    embedding,
    chroma,
    k=3,
)

PROMPT_PREFIX = """
# Community Post
The community posts are as follows:

## ${title}
## Contents
{contents}
## Images
{images}
"""

prompt_few_shot = FewShotPromptTemplate(
    example_selector=examples_selector,
    example_prompt=example_prompt,  
    prefix=PROMPT_PREFIX, 
    suffix="",  
    input_variables=["title","contents","images"] 
)

def get_result_chain(self) -> LLMChain:
        output_parser = PydanticOutputParser(pydantic_object=RESULT)
        prompt = prompt_few_shot.partial(format=output_parser.get_format_instructions())
        fixing_parser = OutputFixingParser.from_llm(
            parser=output_parser, 
            llm=self._llm,
            max_retries=3
        )

        return prompt | self._llm | fixing_parser

result_chain = self.get_result_chain()
result: RESULT = result_chain.invoke({"contents": elaborated_contents, "title": self.title, "images": images })

'result_chain.invoke' is get error

{"timestamp": "2024-10-25T08:21:00.081675Z", "levelname": "ERROR", "message": "", "path": "/v1/ask", "method": "POST", "reason": "'title'", "exc_info": "Traceback (most recent call last):\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/streams/memory.py\", line 105, in receive\n    return self.receive_nowait()\n           ^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/streams/memory.py\", line 100, in receive_nowait\n    raise WouldBlock\nanyio.WouldBlock\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/streams/memory.py\", line 118, in receive\n    return receiver.item\n           ^^^^^^^^^^^^^\nAttributeError: 'MemoryObjectItemReceiver' object has no attribute 'item'\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/base.py\", line 159, in call_next\n    message = await recv_stream.receive()\n              ^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/streams/memory.py\", line 120, in receive\n    raise EndOfStream\nanyio.EndOfStream\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"/src/middleware/router_logging.py\", line 93, in _execute_request\n    response: Response = await call_next(request)\n                         ^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/base.py\", line 165, in call_next\n    raise app_exc\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/base.py\", line 151, in coro\n    await self.app(scope, receive_or_disconnect, send_no_error)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/cors.py\", line 93, in __call__\n    await self.simple_response(scope, receive, send, request_headers=headers)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/cors.py\", line 148, in simple_response\n    await self.app(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/middleware/exceptions.py\", line 65, in __call__\n    await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/_exception_handler.py\", line 64, in wrapped_app\n    raise exc\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/_exception_handler.py\", line 53, in wrapped_app\n    await app(scope, receive, sender)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/routing.py\", line 756, in __call__\n    await self.middleware_stack(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/routing.py\", line 776, in app\n    await route.handle(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/routing.py\", line 297, in handle\n    await self.app(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/routing.py\", line 77, in app\n    await wrap_app_handling_exceptions(app, request)(scope, receive, send)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/_exception_handler.py\", line 64, in wrapped_app\n    raise exc\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/_exception_handler.py\", line 53, in wrapped_app\n    await app(scope, receive, sender)\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/routing.py\", line 72, in app\n    response = await func(request)\n               ^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/fastapi/routing.py\", line 278, in app\n    raw_response = await run_endpoint_function(\n                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/fastapi/routing.py\", line 193, in run_endpoint_function\n    return await run_in_threadpool(dependant.call, **values)\n           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/starlette/concurrency.py\", line 42, in run_in_threadpool\n    return await anyio.to_thread.run_sync(func, *args)\n           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/to_thread.py\", line 56, in run_sync\n    return await get_async_backend().run_sync_in_worker_thread(\n           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/_backends/_asyncio.py\", line 2177, in run_sync_in_worker_thread\n    return await future\n           ^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/anyio/_backends/_asyncio.py\", line 859, in run\n    result = context.run(func, *args)\n             ^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/src/auth/decorator.py\", line 93, in wrapper\n    return func(*args, **kwargs)\n           ^^^^^^^^^^^^^^^^^^^^^\n  File \"/src//router.py\", line 15, in ask_question\n    doctor.invoke()\n  File \"/src//.py\", line 41, in invoke\n    result: RESULT = result_chain.invoke({\"contents\": elaborated_contents, \"title\": \"none\", \"images\": images })\n                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/runnables/base.py\", line 2877, in invoke\n    input = context.run(step.invoke, input, config, **kwargs)\n            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/prompts/base.py\", line 187, in invoke\n    return self._call_with_config(\n           ^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/runnables/base.py\", line 1786, in _call_with_config\n    context.run(\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/runnables/config.py\", line 398, in call_func_with_variable_args\n    return func(input, **kwargs)  # type: ignore[call-arg]\n           ^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/prompts/base.py\", line 162, in _format_prompt_with_error_handling\n    return self.format_prompt(**_inner_input)\n           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/prompts/string.py\", line 277, in format_prompt\n    return StringPromptValue(text=self.format(**kwargs))\n                                  ^^^^^^^^^^^^^^^^^^^^^\n  File \"/Library/Caches/pypoetry/virtualenvs/bedrock-anmY417P-py3.12/lib/python3.12/site-packages/langchain_core/prompts/few_shot.py\", line 180, in format\n    {k: e[k] for k in self.example_prompt.input_variables} for e in examples\n        ~^^^\nKeyError: 'title'", "taskName": "Task-11", "ddReqId": "49ed5f0b292f4b97a2f62ac2558bff47"}

[dosubot[bot]]

Hello, @ko-devHong! I'm here to help you with any issues you're facing. If you need assistance with bugs, questions, or want to become a contributor, feel free to ask.

It looks like you're encountering a KeyError related to the 'title' key in your FewShotPromptTemplate. This error typically occurs when the input data does not match the expected input variables defined in your prompt template.

Here are a few things you can check and adjust:

1. Input Variables Consistency: Ensure that the input variables in your PromptTemplate and FewShotPromptTemplate match the keys in the data you are passing. In your example_prompt, you have defined input_variables=["title", "contents", "images", "format"], but in your prompt_few_shot, you have input_variables=["title", "contents", "images"]. Make sure that the data you pass to result_chain.invoke() includes all these keys.

2. Data Passed to invoke: When you call result_chain.invoke({"contents": elaborated_contents, "title": self.title, "images": images }), ensure that self.title and images are correctly defined and not None. If any of these are missing or None, it could lead to a KeyError.

3. Example Data: Verify that all examples in community_suggest_solution_prompt_examples contain the 'title' key. If any example is missing this key, it could cause the error when the SemanticSimilarityExampleSelector tries to access it.

4. Partial Method: When using the partial method on prompt_few_shot, ensure that the format key is correctly handled. Since you are using partial(format=output_parser.get_format_instructions()), make sure that get_format_instructions() returns a valid format string that can be used in your template.

By ensuring that all keys are consistently defined and that the data passed to the invoke method matches the expected structure, you should be able to resolve the KeyError. If the issue persists, double-check the data being passed and the structure of your templates.

To continue talking to Dosu, mention @dosu.

[ko-devHong]

@dosu When I checked, an error occurs when a form that is not in the experiment comes in. I don't want an error to occur, but I want it to be flexible on its own

[dosubot[bot]]

To handle cases where the input data might not match the expected input variables without causing an error, you can implement a few strategies:

1. Default Values: Modify your code to provide default values for missing keys. You can do this by checking for the presence of keys and assigning defaults if they are missing.

2. Custom Error Handling: Wrap the code that formats the prompt in a try-except block to catch KeyError exceptions and handle them gracefully, perhaps by logging a warning and using default values.

3. Dynamic Input Handling: Before invoking the prompt, dynamically adjust the input dictionary to ensure all required keys are present. You can create a function that checks for missing keys and adds them with default values.

Here's a simple example of how you might implement default values:

def prepare_input_data(input_data, default_values):
    for key, default in default_values.items():
        input_data.setdefault(key, default)
    return input_data

# Define default values for missing keys
default_values = {
    "title": "Default Title",
    "contents": "Default Contents",
    "images": "Default Image URL",
    "format": "Default Format"
}

# Prepare input data with defaults
input_data = prepare_input_data(
    {"contents": elaborated_contents, "title": self.title, "images": images},
    default_values
)

# Invoke the result chain with prepared input data
result: RESULT = result_chain.invoke(input_data)

This approach ensures that your code can handle unexpected input forms by filling in any missing data with predefined defaults, thus preventing errors and allowing for more flexible input handling.

To continue talking to Dosu, mention @dosu.

[ko-devHong]

@dosu How do I restrict the "SemanticSimilarityExampleSelector.from_examples" function from selecting anything other than "input" and "output" in the executions variable when running?

[dosubot[bot]]

To restrict the SemanticSimilarityExampleSelector.from_examples function to only select "input" and "output" in the executions variable, you can use the example_keys parameter. Set example_keys to ["input", "output"] when calling the from_examples method. Here's how you can do it:

selector = SemanticSimilarityExampleSelector.from_examples(
    examples=your_examples_list,
    embeddings=your_embeddings,
    vectorstore_cls=your_vectorstore_class,
    k=4,
    input_keys=None,  # or specify if you have specific input keys
    example_keys=["input", "output"],  # Restrict to "input" and "output"
    vectorstore_kwargs=None
)

This configuration ensures that only the "input" and "output" keys are considered when selecting examples ^[1].

To continue talking to Dosu, mention @dosu.