scikit-tensor dependency issues

[jewettaij]

There are several problems with the scikit-tensor module which SuRVoS depends on:

1) scikit-tensor does not appear to be python3 compatible. For example, when I install it from within a virtual environment (created using "python3 -m venv survos_venv"), the error message I get is: SyntaxError: Missing parentheses in call to 'print'. Did you mean print(mod.version)?

2) scikit-tensor appears to be no longer maintained.

a) The github repository page "https://github.com/mnick/scikit-tensor" is showing a travis build failure. I can clone this and run "pip install .", and there are no errors during installation, but the fact that it has been failing automatic dependency tests for the last 9 months is disconcerting.

b) If I install scikit-tensor using pip (using "pip install scikit-tensor"), I get the following error: pkg_resources.DistributionNotFound: The 'nose' distribution was not found and is required by the application. (Perhaps this is a separate issue. However I've seen this on 3 different python distributions on different computers running freshly installed versions of ubuntu 16.06 and 18.04)

Are there any alternatives to scikit-tensor?

[srikanthnagella]

Please use the pyqt5 branch. master branch is only compatible with python 2.7

On Sat, Mar 24, 2018 at 3:30 AM, Andrew Jewett notifications@github.com wrote:

There are several problems with the scikit-tensor module which SuRVoS depends on:

1.

scikit-tensor does not appear to be python3 compatible. For example, when I install it from within a virtual environment (created using "python3 -m venv survos_venv"), the error message I get is: SyntaxError: Missing parentheses in call to 'print'. Did you mean print(mod.version)? 2.

scikit-tensor appears to be no longer maintained.

a) The github repository page "https://github.com/mnick/scikit-tensor" is showing a travis build failure. I can clone this and run "pip install .", and there are no errors during installation, but the fact that it has been failing automatic dependency tests for the last 9 months is disconcerting.

b) If I install scikit-tensor using pip (using "pip install scikit-tensor"), I get the following error: pkg_resources.DistributionNotFound: The 'nose' distribution was not found and is required by the application. (Perhaps this is a separate issue. However I've seen this on 3 different python distributions on different computers running freshly installed versions of ubuntu 16.06 and 18.04)

Are there any alternatives to scikit-tensor?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/DiamondLightSource/SuRVoS/issues/36, or mute the thread https://github.com/notifications/unsubscribe-auth/AAovcdbyILyOF020NTE6XcnwyLdEBEyTks5thb3MgaJpZM4S5lYI .

[jewettaij]

Thanks. I can confirm that wwitching to the pyqt5 branch and installing via conda gets around this problem. However, the documentation at https://diamondlightsource.github.io/SuRVoS/docs/installation/ needs to be updated. Here's an excerpt from that web page:

With Anaconda:

$> conda update conda pip
$> conda install numpy scipy matplotlib scikit-learn scikit-image cython seaborn networkx libgcc pyqt=4.11.4
$> pip install scikit-tensor tifffile

With another python distribution:

$> pip install --upgrade pip
$> pip install --upgrade numpy scipy matplotlib scikit-learn scikit-image cython seaborn networkx scikit-tensor tifffile

I'll close this issue when the documentation is updated (and when the "pyqt5" branch is hopefully merged) Thanks again

[srikanthnagella]

Not relevant anymore. The latest version doesn't depend on this.