Moodle Anylitics Python Backend question

Moodle Anylitics Python Backend question

by Stephen Bourget -
Number of replies: 2
Picture of Core developers Picture of Plugin developers Picture of Testers

Hi,

I'm attempting to install the Python Machine Learning Backend on one of My Moodle 3.5.1+ instances.

(The server is Running Windows Server 2008r2)

I've successfully installed Python, but when i Run "pip install moodlemlbackend" as specified in the docs (https://docs.moodle.org/35/en/Analytics#Predictions_processor) it fails with the following:

Collecting tensorflow<1.1,>=1.0.0 (from moodlemlbackend)
Could not find a version that satisfies the requirement tensorflow<1.1,>=1.0.0
(from moodlemlbackend) (from versions: )
No matching distribution found for tensorflow<1.1,>=1.0.0 (from moodlemlbackend)

Here is the full output from running the command:

Has anyone encountered this issue before?  If so, has anyone been able to work around it?

Average of ratings: -
In reply to Stephen Bourget

Re: Moodle Anylitics Python Backend question

by Moodle Admin -

Hello Stephen 

it is related with version of Tensor Flow  it should be between 1.1.0 and 1.0.0 so you can install the related 

version as below 


pip install --upgrade https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.1.0-py2-none-any.whl


after installing the related version then try to install moodlemlbackend again.

Average of ratings: -
In reply to Stephen Bourget

Re: Moodle Anylitics Python Backend question

by David Monllaó -
Hi Stephen,

Sorry, I am not subscribed to this forum and I've just seen your question. I see that you may be using python 3.7. The current version of the python package is only available from python 2.7, 3.4 and 3.5. We are bumping the version of our dependencies at the moment so we can support newer python versions. You can find more info in https://github.com/moodlehq/moodle-mlbackend-python/issues/8. The summary is:

* For stables (Moodle 3.6 and 3.7) we add compatibility for python 3.6
* For master (Moodle 3.8dev) we add compatibility for python 3.5, 3.6 and 3.7, and we drop compatibility with python 2.7 (end of life Jan 2020

Unfortunately there was no combination of our dependencies that would guarantee the compatibility of our package from python 2.7 to python 3.7. We need to guarantee support for python 2.7 in stables so we can't support python 3.7.
Average of ratings: -