djangoiislangchainraggpt4all

Getting Access Denied Error When Hitting API (Although, API is working fine when running on LocalHost via Django)

I used GPT4all to create a RAG Pipeline with the help of langchain. When I use API for other purposes, it works fine but when I Hit the RAG pipeline, It gives access denied error.

Error getting: Permission Error
Root Cause: Unable to create model download directory in system profile. Access denied to 'C:\Windows\system32\config\systemprofile.cache'

Environment:
Python Version: 3.12
Operating System: Windows
Key Packages:

Full Error:

{
    "error":"Error in Database Chain: Failed to create model download directory",
    "traceback":"Traceback (most recent call last):

File "C:\Python312\Lib\site-packages\gpt4all\gpt4all.py", line 323, in retrieve_model
    os.makedirs(DEFAULT_MODEL_DIRECTORY, exist_ok=True)

File "", line 215, in makedirs

File "", line 225, in makedirs
    PermissionError: [WinError 5] Access is denied: 
    'C:\\Windows\\system32\\config\\systemprofile\\.cache'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):

File "C:\inetpub\wwwroot\hashmove-ai\HMAI\Controllers\ConversationalAIv2_Controller.py", line 143, in post
    hmgpt_response = query_response(input_query, query_intent)

File "C:\inetpub\wwwroot\hashmove-ai\HMAI_Business\ConversationalAIv2_Business.py", line 281, in query_response
    return response_functions.get(intent_response, irrelevant)(query)

File "C:\inetpub\wwwroot\hashmove-ai\HMAI_Business\ConversationalAIv2_Business.py", line 251, in logistics
    chain = log_chain(load_llm(), vector_db(), memory(), log_prompt())

File "C:\inetpub\wwwroot\hashmove-ai\HMAI_Business\ConversationalAIv2_Business.py", line 90, in load_llm
    loaded_llm = GPT4All(

File "C:\Python312\Lib\site-packages\langchain_core\load\serializable.py", line 125, in init
    super().init(*args, **kwargs)

File "C:\Python312\Lib\site-packages\pydantic\main.py", line 212, in init
    validated_self = self.pydantic_validator.validate_python(data, self_instance=self)

File "C:\Python312\Lib\site-packages\pydantic\_internal\_decorators_v1.py", line 148, in _wrapper1
    return validator(values)

File "C:\Python312\Lib\site-packages\langchain_core\utils\pydantic.py", line 208, in wrapper
    return func(cls, values)

File "C:\Python312\Lib\site-packages\langchain_community\llms\gpt4all.py", line 145, in validate_environment
    values["client"] = GPT4AllModel(

File "C:\Python312\Lib\site-packages\gpt4all\gpt4all.py", line 235, in init
    self.config: ConfigType = self.retrieve_model(model_name, model_path=model_path, allow_download=allow_download, verbose=verbose) 

File "C:\Python312\Lib\site-packages\gpt4all\gpt4all.py", line 325, in retrieve_model
    raise RuntimeError("Failed to create model download directory") from e

RuntimeError: Failed to create model download directory
"}

Tried giving manuall path to the cache:

os.environ["TIKTOKEN_CACHE_DIR"] = "embeddings-cache"

But didn't worked.

Note: I've deployed this on a server using IIS server.

Solution

  • I've deployed this on a server using IIS server

    The error message is clear, it is an identity permission problem. What identity does your application access as? Anonymous identity or something else?

    But based on the error message, it seems that your code related to some system driver folder permission. You can try to set admin permission in Application pool identity.

    1. Open the Application Pools node underneath the machine node. Select the application pool you want to change.
    2. Right click the application pool and select Advanced Settings.
    3. Select the Identity list item and click the ellipsis (the button with the three dots).
    4. Check Custom account and Set....
    5. Enter Enter the local administrator name and password. enter image description here