I am learning to work with Nbdev.

One of the requirement is to install Anaconda and run nbdev library from the same server where Jupyter Notebook is running.

Uptill now I utilised Google Colab or Paperspace Code for most of my work.However, here it required installing Anaconda, and I followed the documentation available on the site as it is. Without any issues.

Issues I faced and resources that helped:

  • I learnt that nbdev library was test to work on Python 3.7 and hence I needed Python 3.7 instead of Python 3.8 (also I was running through some errors and while searching for solutions I found it is easier to try with Python 3.7). This is where Anaconda comes handy, you can create environments and I wanted to keep Python 3.7 and Nbdev in one environment isolated.
    Anaconda Cheatsheet consists of command on how to create environments and activate and dectivate them.To come out of base environment:conda deactivate

  • Learning how can I add packages to a specific environment in conda only.By now I had setup a Python 3.7 environment but when I was installing Nbdev library, for some reason it was available globally. This is when I went back to Anaconda Documentation page.
    • How I knew it was available globally?
      I could access the nbdev library from the (base) enviroment as well.

  • On the docs page, I learnt about Anaconda Navigator which is a GUI and help me visualize which environment has what packages and libraries installed. Anaconda Navigator also makes it easy installing libraries to environments. For example, Jupyter Notebook is installed during Anaconda installation but is available in the base environment only. When a new enviroment is created it has to be added separately. Solution for a prevelent problem of not able to access Navigator from Desktop as an app

Issues I faced during working with Nbdev tutorial: