`!connect jupyter` directive error

[tonidy]

Describe the bug

After upgrading Polyglot to 1.0.6070010 and testing with .NET 9, I'm getting this error when trying to use #!connect jupyter

Although I've already installed the deno kernel:

also tried with another kernel, e.g: python, but the issue still persists.

Please complete the following:

Which version of .NET Interactive are you using? (In a notebook, run the #!about magic command. ):

• OS
  • Windows 11
  • Windows 10
  • macOS
  • Linux (Please specify distro)
  • iOS
  • Android
• Frontend
  • Jupyter Notebook
  • Jupyter Lab
  • nteract
  • Visual Studio Code
  • Visual Studio Code Insiders
  • Visual Studio
  • Other (please specify)

[jonsequitur]

Are those all present in your base environment (which is the default for #!connect)?

[tonidy]

@jonsequitur how to check if all presents in the base environment?

[kennedysousa]

Yesterday I came across the very same issue. All the Polyglot notebooks that I use with Jupyter kernel stopped working - even the ones that were working perfectly the night before.

I'm working with Python and Jupyter directly without Anaconda, so I use the following magic command:

#!connect jupyter --kernel-name pythonkernel --kernel-spec python3

I don't have the screenshots, but I was getting the same error as shown in @tonidy's message, when the command jupyter kernelspec list returned something like:

PS C:\Users\usr> jupyter kernelspec list
Available kernels:
  python3    C:\Users\usr\AppData\Local\Programs\Python\Python313\share\jupyter\kernels\python3

I dropped the kernel with jupyter kernelspec remove python3 and recreated it with python -m ipykernel install --user --name=python3, which changed the path to:

PS C:\Users\usr> jupyter kernelspec list
Available kernels:
  python3    C:\Users\usr\AppData\Roaming\jupyter\kernels\python3

Then, the error message changed to the following:

Error: System.InvalidOperationException: Cannot start process because a file name has not been provided.
at System.Diagnostics.Process.Start()
at Microsoft.DotNet.Interactive.Utility.CommandLine.StartProcess(String command, String args, DirectoryInfo workingDir, Action`1 output, Action`1 error, ValueTuple`2[] environmentVariables) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\Utility\CommandLine.cs:line 172
at Microsoft.DotNet.Interactive.Jupyter.CondaEnvironment.StartProcess(String command, String args, DirectoryInfo workingDir, Action`1 output, Action`1 error) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive.Jupyter\CondaEnvironment.cs:line 84
at Microsoft.DotNet.Interactive.Jupyter.ZMQ.JupyterConnection.CreateKernelProcess(KernelSpec spec, String connectionFilePath) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive.Jupyter\ZMQ\JupyterConnection.cs:line 128
at Microsoft.DotNet.Interactive.Jupyter.ZMQ.JupyterConnection.CreateKernelConnectionAsync(String kernelSpecName) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive.Jupyter\ZMQ\JupyterConnection.cs:line 92
at Microsoft.DotNet.Interactive.Jupyter.JupyterKernelConnector.CreateKernelAsync(String kernelName) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive.Jupyter\JupyterKernelConnector.cs:line 27
at Microsoft.DotNet.Interactive.Jupyter.ConnectJupyterKernelDirective.ConnectKernelsAsync(ConnectJupyterKernel connectCommand, KernelInvocationContext context) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive.Jupyter\ConnectJupyterKernelDirective.cs:line 70
at Microsoft.DotNet.Interactive.CompositeKernel.ConnectKernel[TCommand](TCommand command, ConnectKernelDirective`1 connectDirective, KernelInvocationContext context) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\CompositeKernel.cs:line 253
at Microsoft.DotNet.Interactive.CompositeKernel.<>c__DisplayClass20_0`1.<<AddConnectDirective>b__0>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\CompositeKernel.cs:line 238
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.Kernel.HandleAsync(KernelCommand command, KernelInvocationContext context) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\Kernel.cs:line 365
at Microsoft.DotNet.Interactive.CompositeKernel.HandleAsync(KernelCommand command, KernelInvocationContext context) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\CompositeKernel.cs:line 194
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<BuildPipeline>b__6_0(KernelCommand command, KernelInvocationContext context, KernelPipelineContinuation _) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 60
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_1.<<BuildPipeline>b__3>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 75
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.App.KernelExtensions.<>c__DisplayClass5_0.<<UseTelemetrySender>b__0>d.MoveNext() in D:\a\_work\1\s\src\dotnet-interactive\KernelExtensions.cs:line 296
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_1.<<BuildPipeline>b__3>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 75
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.App.KernelExtensionLoader.<>c__DisplayClass0_0.<<UseNuGetExtensions>b__0>d.MoveNext() in D:\a\_work\1\s\src\dotnet-interactive\KernelExtensionLoader.cs:line 25
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_1.<<BuildPipeline>b__3>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 75
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.App.KernelExtensions.<>c__DisplayClass4_0.<<UseSecretManager>b__0>d.MoveNext() in D:\a\_work\1\s\src\dotnet-interactive\KernelExtensions.cs:line 227
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_0.<<BuildPipeline>g__Combine|2>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 73
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_0.<<BuildPipeline>g__Combine|2>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 73
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.<>c__DisplayClass6_0.<<BuildPipeline>g__Combine|2>d.MoveNext() in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 73
--- End of stack trace from previous location ---
at Microsoft.DotNet.Interactive.KernelCommandPipeline.SendAsync(KernelCommand command, KernelInvocationContext context) in D:\a\_work\1\s\src\Microsoft.DotNet.Interactive\KernelCommandPipeline.cs:line 41

[jonsequitur]

@kennedysousa Are you starting VS Code from a Conda environment? (The previous documentation said to do this, but it should no longer be necessary.)

[kennedysousa]

Hi @jonsequitur! I'm not starting VS Code from a Conda environment. In fact, I do not use Conda at all within my devbox.

[jonsequitur]

Additional logs would be helpful. These issues tend to be environment-specific and I'm not confident in this heuristic for robustly locating a Conda path.

If you can, please follow the directions here: https://github.com/dotnet/interactive/blob/main/DEVELOPER-GUIDE.md#use-a-local-build-of-the-dotnet-interactive-tool. Just add just the following arguments to your kernelTransportArgs settings (no local build needed):

      "--log-path",
      "/path/to/a/folder/for/your/kernel-logs/",
      "--verbose",

If you can repro the bug and post back the relevant section of the log file, that would be helpful.

Thanks!

[kennedysousa]

Hi @jonsequitur! Thanks for the quick response.

While reviewing the instructions you shared, I realized I don't have admin privileges on my company laptop to run some of the steps. So, I decided to reproduce the issue on my personal computer. I followed the instructions on the Polyglot extension page to install the latest .Net 8 SDK. However, when I created a new Polyglot notebook, it tried to install .Net Interactive 1.0.607001 but got stuck and never completed.

I then referred back to the developer instructions to generate the logs you mentioned and found that the project depends on .Net 9.0.

After installing .NET 9.0, the installation of .NET Interactive completed successfully, and I was able to use #!connect jupyter --kernel-name pythonkernel --conda-env base --kernel-spec python3 in my Polyglot notebook on my personal computer.

So, I'm guessing the auto-update bumped the .Net Interactive up but it no longer works with the .Net 8.0 SDK (LTS).

I've raised a ticket to get .Net 9.0 installed on my work computer and will let you know if that helps as soon as the service desk fulfills the request.

[tonidy]

@jonsequitur I've attempted to install conda, and after selecting it

it's still not working:

What am I missing? I can see @kennedysousa can connect to the python kernel? Maybe this issue only happened on macos/linux?

[jonsequitur]

@tonidy Could you follow these instructions and post your logs? Thanks.

#3813 (comment)

[tonidy]

@jonsequitur here is the logs:

2025-01-26T06:53:34.1646370Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-891cf0a985e94123-00] [Microsoft.DotNet.Interactive.Jupyter.CondaEnvironment]  ℹ Found Conda on path: /opt/homebrew/Caskroom/miniconda/base/bin
2025-01-26T06:53:34.1673700Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-14ac0883b0975c1b-00] [CommandLine] [Execute]  ▶ > conda activate base&jupyter kernelspec list --json 
2025-01-26T06:53:34.1678010Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-2412da768328dfe5-00] [CommandLine] [StartProcess]  ▶  
2025-01-26T06:53:34.1679930Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-2412da768328dfe5-00] [CommandLine] [StartProcess]  ℹ (0.193ms) > conda activate base&jupyter kernelspec list --json 
2025-01-26T06:53:34.2149730Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-2412da768328dfe5-00] [CommandLine] [StartProcess]  ⏹ (47.175ms)  
2025-01-26T06:53:34.4359590Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-14ac0883b0975c1b-00] [CommandLine] [Execute]  ⏹ -> ❌ (268.924ms) Microsoft.DotNet.Interactive.Utility.CommandLineInvocationException: Exit code 2

StdErr:
usage: conda [-h] [-v] [--no-plugins] [-V] COMMAND ...
conda: error: unrecognized arguments: --json

StdOut: Microsoft.DotNet.Interactive.Utility.CommandLineInvocationException: Exit code 2

StdErr:
usage: conda [-h] [-v] [--no-plugins] [-V] COMMAND ...
conda: error: unrecognized arguments: --json

StdOut:
2025-01-26T06:53:34.4384650Z [00-2e33d7f0573ef2308ddd0a3d697dcd27-891cf0a985e94123-00] [Microsoft.DotNet.Interactive.Jupyter.JupyterKernelSpecModule]  ⚠  +[ Failed to list kernelspecs., [ usage: conda [-h] [-v] [--no-plugins] [-V] COMMAND ..., conda: error: unrecognized arguments: --json ] ] 

[jonsequitur]

Thanks!
 
#!connect jupyter is executing this (as indicated in your log): conda activate base&jupyter kernelspec list --json. It does this assuming a compatible conda is present on your path. Could it be that you have more than one installation present and an older version that doesn't support the --json option is actually getting invoked? Is the installation at /opt/homebrew/Caskroom/miniconda/base/bin the one you would expect to see in use?