Hacker News Reader

Comment by projectramo | Hacker News Reader

Here are the issues with Jupyter, and most other flavor, of notebook:

1. variables have to be explicitly output

The most important tool for programming, for me, is that window that shows you the current state of all the variables. When I step through a program, I look at the state. 90% of my debugging solutions come from seeing that variable doesn't have the right state.

2. Intellisense

For the love of god, I do not want to remember if it is len(), length(), .len(), .length(), .size(), size(1) or whatever.

That's it. But those two are so big that I have to code and debug in Spyder and then paste the code into notebook. I feel sorry for people who are new who think that all the debugging is happening in the notebook.

View on HN

Hey There! I'm trying to solve the issue of IntelliSense.. I'm building/improving Jupyter Notebooks inside VSCode: https://github.com/pavanagrawal123/VSNotebooks . It's a fork from another extension somebody already built, but all activity is dead, so I'm starting up dev on an active fork. I'd love to hear any feedback y'all have! :)

Also planning to add some nice debug features, plus hopefully integration into the inbuilt VSCode debugger!

How do you see the idea of a vs code notebook comparing to or being different from the goals of the hydrogen editor?

To me, my favorite part of the design of Hydrogen is that it's entirely language agnostic, and can be used with _any_ Jupyter kernel.

I'm planning on adding better language support in a couple of weeks! Don't want to limited to Python and R

One thing you may want to be aware of is that the python language server used for vscode runs pylint which performs static analysis on the code. However, Jupyter Notebook uses autocomplete by actually introspecting the variables as they are defined. This creates large differences when doing things such as selecting a column in a pandas dataframe. In jupyter if you press tab on the column name, it can autocomplete and also assumes you are getting a series, which leads to autocomplete on things like .min, .max, etc... In pylint you don't get any of this autocomplete since pylint cannot statically determine the column names so you lose the intellisense.

yep! this is something I will be focussing on in VSNotebooks. I noticed this as a huge drawback to the current implementation, so I will be fixing it :)

I think the biggest thing for me personally is that I like VScode a lot more than Atom. Also, I'm going to be focussing more on better debugging, which AFAIK, is not heavily emphasized in hydrogen.

Better debugging is something I want to focus on in VSNotebooks.

Personally, I like VSCode more than Atom, so this was one of the reasons I started working on this extension!

Thanks for working on this. We’re converging on the Bret Victor IDE

Microsoft just announced an initiative like this (unfunded, community-based, likely at risk of becoming abandonware), perhaps you could combine your efforts with theirs? (The issue in their code that I'm personally most impacted by in lack of support for conda[0])

[0]https://github.com/lorenzo2897/vscode-ipe/issues/162

HN thread for that announcement: https://news.ycombinator.com/item?id=18346198

Fixed this in my fork :)

I want to combine, but Neuron has stated they will not be accepting any PRs until December.

Re #2, if you haven't tried the newest versions recently (and especially with the jupyterlab beta which has a nicer completion GUI), I'd encourage you to take a look! It's come a long way, along with the library that's doing the completions under the hood.

Thanks for the tip. Will take a look. I thought I was using the latest version.

Re #1: There's a plugin in nbextensions that shows variable values a la spyder.

Good to know. Will check it out.

Reply on news.ycombinator.com