Community nodes KNIME AI Extensions GPT4All Connectors bug when used offline

KNIME Extensions
,
1

Hi all!

It’s really awesome to see all those helpful packages and examples popping up that help to try out AI models on your own!

I found a bug in the GPT4All nodes in the KNIME AI Extension package. When you are offline and you select a model to be read from locally, the GPT4All Connectors still try to access gpt4all.io (to fetch /models/models2.json) which is not needed because I am not downloading a model from there, but want to just connect to the one I have stored locally. If I am connected to the internet, the connectors to the local file work just fine and do an awesome job :slight_smile:

WARN  GPT4All Embeddings Connector 3:1979     Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 203, in _new_conn
    sock = connection.create_connection(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\util\connection.py", line 60, in create_connection
    for res in socket.getaddrinfo(host, port, family, socket.SOCK_STREAM):
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\socket.py", line 962, in getaddrinfo
    for res in _socket.getaddrinfo(host, port, family, type, proto, flags):
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
socket.gaierror: [Errno 11001] getaddrinfo failed

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

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 790, in urlopen
    response = self._make_request(
               ^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 491, in _make_request
    raise new_e
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 467, in _make_request
    self._validate_conn(conn)
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 1096, in _validate_conn
    conn.connect()
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 611, in connect
    self.sock = sock = self._new_conn()
                       ^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 210, in _new_conn
    raise NameResolutionError(self.host, self, e) from e
urllib3.exceptions.NameResolutionError: <urllib3.connection.HTTPSConnection object at 0x000001F69988F810>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)

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

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\adapters.py", line 486, in send
    resp = conn.urlopen(
           ^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 844, in urlopen
    retries = retries.increment(
              ^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\util\retry.py", line 515, in increment
    raise MaxRetryError(_pool, url, reason) from reason  # type: ignore[arg-type]
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
urllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='gpt4all.io', port=443): Max retries exceeded with url: /models/models2.json (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0x000001F69988F810>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)"))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\models\gpt4all.py", line 644, in execute
    _Embeddings4All(
  File "pydantic\main.py", line 339, in pydantic.main.BaseModel.__init__
  File "pydantic\main.py", line 1102, in pydantic.main.validate_model
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\models\gpt4all.py", line 476, in validate_environment
    values["client"] = _GPT4All(
                       ^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 97, in __init__
    self.config: ConfigType = self.retrieve_model(model_name, model_path=model_path, allow_download=allow_download, verbose=verbose)
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 149, in retrieve_model
    available_models = GPT4All.list_models()
                       ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 118, in list_models
    resp = requests.get("https://gpt4all.io/models/models2.json")
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\api.py", line 73, in get
    return request("get", url, params=params, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\api.py", line 59, in request
    return session.request(method=method, url=url, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\sessions.py", line 589, in request
    resp = self.send(prep, **send_kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\sessions.py", line 703, in send
    r = adapter.send(request, **kwargs)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\adapters.py", line 519, in send
    raise ConnectionError(e, request=request)
requests.exceptions.ConnectionError: HTTPSConnectionPool(host='gpt4all.io', port=443): Max retries exceeded with url: /models/models2.json (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0x000001F69988F810>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)"))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\plugins\org.knime.python3.nodes_5.2.0.v202311290857\src\main\python\_node_backend_launcher.py", line 718, in execute
    outputs = self._node.execute(exec_context, *inputs)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\models\gpt4all.py", line 650, in execute
    raise ValueError(f"The model at path {self.model_path} is not valid.")
ValueError: The model at path C:\Users\dobo\knime-workspace-5.2\llm_models\all-MiniLM-L6-v2-f16.gguf is not valid.

ERROR GPT4All Embeddings Connector 3:1979     Execute failed: The model at path C:\Users\dobo\knime-workspace-5.2\llm_models\all-MiniLM-L6-v2-f16.gguf is not valid.
2 Likes
2

Hi @dobo ,

Thank you very much for reporting this. I can reproduce the error only for the GPT4All Embeddings Connector node.

The other GPT4All Connector nodes work fine for me (without Internet connection). If that’s not the case for you, can you please provide the error trace also for those nodes?

I’ve opened an internal ticket for our dev team to investigate the issue (AP-21736).

Sorry for the inconvenience,
Roberto

4 Likes
3

Hi @roberto_cadili,

sorry for the confusion, I thought it was also occurring in the GPT4All Chat Model Connector, but instead it was the Vector Store Retriever!

WARN  Vector Store Retriever 3:1900     Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 203, in _new_conn
    sock = connection.create_connection(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\util\connection.py", line 60, in create_connection
    for res in socket.getaddrinfo(host, port, family, socket.SOCK_STREAM):
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\socket.py", line 962, in getaddrinfo
    for res in _socket.getaddrinfo(host, port, family, type, proto, flags):
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
socket.gaierror: [Errno 11001] getaddrinfo failed

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

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 790, in urlopen
    response = self._make_request(
               ^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 491, in _make_request
    raise new_e
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 467, in _make_request
    self._validate_conn(conn)
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 1096, in _validate_conn
    conn.connect()
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 611, in connect
    self.sock = sock = self._new_conn()
                       ^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connection.py", line 210, in _new_conn
    raise NameResolutionError(self.host, self, e) from e
urllib3.exceptions.NameResolutionError: <urllib3.connection.HTTPSConnection object at 0x0000020B18F37B50>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)

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

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\adapters.py", line 486, in send
    resp = conn.urlopen(
           ^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\connectionpool.py", line 844, in urlopen
    retries = retries.increment(
              ^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\urllib3\util\retry.py", line 515, in increment
    raise MaxRetryError(_pool, url, reason) from reason  # type: ignore[arg-type]
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
urllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='gpt4all.io', port=443): Max retries exceeded with url: /models/models2.json (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0x0000020B18F37B50>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)"))

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Software\knime_5.2.0\plugins\org.knime.python3.nodes_5.2.0.v202311290857\src\main\python\_node_backend_launcher.py", line 718, in execute
    outputs = self._node.execute(exec_context, *inputs)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\indexes\base.py", line 362, in execute
    db = vectorstore.load_store(ctx)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\indexes\base.py", line 200, in load_store
    embeddings = self.embeddings_model.create_model(ctx)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\models\gpt4all.py", line 537, in create_model
    return _Embeddings4All(
           ^^^^^^^^^^^^^^^^
  File "pydantic\main.py", line 339, in pydantic.main.BaseModel.__init__
  File "pydantic\main.py", line 1102, in pydantic.main.validate_model
  File "C:\Software\knime_5.2.0\plugins\org.knime.python.llm_5.2.0.202312061509\src\main\python\src\models\gpt4all.py", line 476, in validate_environment
    values["client"] = _GPT4All(
                       ^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 97, in __init__
    self.config: ConfigType = self.retrieve_model(model_name, model_path=model_path, allow_download=allow_download, verbose=verbose)
                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 149, in retrieve_model
    available_models = GPT4All.list_models()
                       ^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py", line 118, in list_models
    resp = requests.get("https://gpt4all.io/models/models2.json")
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\api.py", line 73, in get
    return request("get", url, params=params, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\api.py", line 59, in request
    return session.request(method=method, url=url, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\sessions.py", line 589, in request
    resp = self.send(prep, **send_kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\sessions.py", line 703, in send
    r = adapter.send(request, **kwargs)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Software\knime_5.2.0\bundling\envs\org_knime_python_llm\Lib\site-packages\requests\adapters.py", line 519, in send
    raise ConnectionError(e, request=request)
requests.exceptions.ConnectionError: HTTPSConnectionPool(host='gpt4all.io', port=443): Max retries exceeded with url: /models/models2.json (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0x0000020B18F37B50>: Failed to resolve 'gpt4all.io' ([Errno 11001] getaddrinfo failed)"))
1 Like
4

Hi @dobo,

Thanks for further reporting. I’ve added this information to the open ticket.

Our dev team is already investigating the issue.

Best,
Roberto

3 Likes
5

I ran into this, too, with the Embeddings4All node. Online, but behind our corporate proxy. The problem is that even if you provide a model, it calls home for the JSON file when a GPT4All object instance is created. And then wrongly claims the model file is invalid.

I fixed it manually by adding “allow_download=False” to that GPT4All constructor call. Of course ultimately GPT4All should be fixed to respect proxy settings when available.

@KNIME team: proxy settings are a permanent source of trouble in an enterprise setting. I’d really suggest to have some of your devs/testers work from behind a proxy by default to spot those recurring issues.
But overall great job making GPT4All available so quickly!! :muscle:

4 Likes
6

@notsure where would you enter this option?

EDIT: I think I found the option in:

...\bundling\envs\org_knime_python_llm\Lib\site-packages\gpt4all\gpt4all.py

image
image718×258 6.52 KB

and:

image

1 Like
7

Yes that’s the option. I edited it in the Embeddings class in plugins/org…llm…/src/main/python/src/models/gpt4all.py around line 480. But your way should work, too, and disable network for all GPT4All calls.

2 Likes
8

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.