doc(examples): 01. Quickstart Jupyter Notebook

[oktak]

#93
Added Quickstart notebook.

[whitead]

Thank you for the contribution! We'll have it reviewed shortly.

[whitead]

Looks good, a few small changes:

1. With the simple compiling instructions, you shouldn't need to do the complex conda instructions even if you are in a separate conda environment. Those instructions are more for if you have lots of conflicting python versions installed, like on a shared compute cluster. I'll add some text to the README about this and let's just refer to the simple instructions for your notebook.
2. Please remove the code amended. That's good for me to know, but probably not someone trying out hoomd-tf. You're correct that the graph.save order was wrong, thank you!

[oktak]

Looks good, a few small changes:

1. With the simple compiling instructions, you shouldn't need to do the complex conda instructions even if you are in a separate conda environment. Those instructions are more for if you have lots of conflicting python versions installed, like on a shared compute cluster. I'll add some text to the README about this and let's just refer to the simple instructions for your notebook.

2. Please remove the code amended. That's good for me to know, but probably not someone trying out hoomd-tf.  You're correct that the `graph.save` order was wrong, thank you!

Thanks for your reply. So your suggestion is to remove the texts, and leave the title and those code blocks. Am I correct?

1. It's because I want to keep my OS's python environment clean. And I am not sure if the others build the code without conda. So I've chosen to mention it.

2. It's ok for me to remove that. But I think I should add the inline comment to hoomd.context.initialize("--mode=cpu") line, because when I call that without any parameters, there will be errors that saying '-f' parameter is unexpected. Therefore, I suspect that the coder should choose mode GPU or CPU explicitly.

[whitead]

Great, thanks. Sure leave the conda instructions you mentioned if you found that using the simple instructions caused issues. 2. Ok please leave the initialization flag in then.

[oktak]

Okay, I updated and pushed another commit.

[dilnoza92]

I also think you need add a section to the README saying that the jupyter command that you will be using also agrees with the python version that has hoomd-tf and tensorflow installed. Maybe it is intuitive but it might be unintuitive for some novice users.

[whitead]

@dilnoza92 adding running instructions to the readme sounds like a good idea, but maybe that could be separate from this PR.

[dilnoza92]

@whitead yes it can be a new PR.