2

If I am looking at the Anaconda Packages (http://docs.continuum.io/anaconda/pkg-docs), there is a package named "nbconvert" in version 4.0.0. On the github site of nbconvert (https://github.com/jupyter/nbconvert/releases) there is a newer version (i need this one) 4.1.0. Now how to install this version on my machine? If i try to 

conda install nbconvert

It will install version 4.0.0. but I need this new one

If I try to search for this package with 

anaconda search -t conda nbconvert

It only finds the 4.0.0 version

Raphael
  • 673
  • 1
  • 9
  • 27
  • you can ask for a bump in the anaconda issue tracker, but from my experience the developers are very busy and only selectively choose bugs to work on. But probably worth a try. You can find the issue tracker here: https://github.com/ContinuumIO/anaconda-issues/issues If you need the new package urgently, you can uninstall conda's version and switch to `pip`, which should always ship the latest version. – cel Dec 09 '15 at 08:51
  • I have now pip on my machine, but how to evaluate which library (pip or conda) my machine uses when i type "jupyter nbconver --to PDF file" to my command line? – Raphael Dec 09 '15 at 09:06
  • `jupyter nbconvert --version` will show the version. If you see `4.1.0` this is definitely not conda's version. However I am not perfectly sure if installing nbconvert via pip will do or the whole jupyter stack has to be replaced (which would be rather cumbersome). In the latter case it may be easier to create an empty conda enviroment just for nbconvert and install everything related to jupyter with pip. This I would have to check, though. – cel Dec 09 '15 at 09:10
  • 1
    https://github.com/ContinuumIO/anaconda-issues/issues/579 – cel Dec 09 '15 at 12:00

1 Answers1

3

The easiest way is to temporarily replace conda's nbconvert with the latest version of pip.

Here's how you can do that:

conda remove nbconvert
pip install nbconvert

After that you can check the version of nbconvert by typing jupyter nbconvert --version. If everything worked out, you will see 4.1.0.

Note that this is just meant to be a temporary workaround until continuum analytics' developers update the package. Don't forget to pip uninstall and conda install when switching back to conda's version.

cel
  • 30,017
  • 18
  • 97
  • 117