import stream as st st.__version__

[sailseem]

Dear，
Thank you for this good package, well, I can install this successfully, but when I run the jupyter I cannot even import stream as st. Always returns to "The kernel appears to have died." I can import pandas or other packages without issues.
Can you help me with it? Thanks

[qrzhang]

Hi Sailseem,

It's hard to say what exactly the problem is. But it should be some problem related to your environment. It should be easier to directly use the docker image we provided. It's simple with only two steps, installing Docker and loading stream image. You can find detailed instructions from our GitHub homepage.

[huidongchen]

Dear， Thank you for this good package, well, I can install this successfully, but when I run the jupyter I cannot even import stream as st. Always returns to "The kernel appears to have died." I can import pandas or other packages without issues. Can you help me with it? Thanks

Hi, Sorry about the delay. Can you add the following code before importing 'stream' and give it another try?

import os 
os.environ['KMP_DUPLICATE_LIB_OK']='True'

[sailseem]

thanks, but not solve my problem, i guess I will try to use docker instead

[Roger-GOAT]

can't download daemon.
Can anyone solve this "The kernel appears to have died." ?

[xindong95]

After applying a workaround, I found the problem might be due to rpy2.robjects import errors. This occurs when using a Jupyter environment that differs from the environment where stream is installed. For example, if stream is installed in the env_stream conda environment, but you launch Jupyter Lab from the base environment, rpy2 may fail to locate the correct R installation.

A resolution is that setting the R lib before import stream:

import os
os.environ['R_HOME'] = '/path/to/conda/envs/env_stream/lib/R'

import stream as st

Another resolution is to change the environment variable in jupyter kernel settings in ~/.local/share/jupyter/kernels/env_stream/kernel.json:

{
 "argv": [
  "/path/to/conda/envs/env_stream/bin/python",
  "-m",
  "ipykernel_launcher",
  "-f",
  "{connection_file}"
 ],
 "display_name": "stream",
 "language": "python",
 "metadata": {
  "debugger": true
 },
 "env": {
  "R_HOME": "/path/to/conda/envs/env_stream/lib/R"       // change or add this line!
 }

This could be the resolution to this issue.

@Roger-GOAT @sailseem