This issue was generated by todo based on a TODO comment in f20da86 when #41 was merged. cc @nteract.

The now github integration allows us to automate our deployments including easy PR review with demo links: https://zeit.co/github

If one double clicks to launch nteract, nteract does not launch within the conda environment (if activated). Instead it launches in the global space.

If one launches nteract via the CLI after activating the conda environment, nteract is working within the conda environment.

Currently flow-typed typings aren't used by flow. The flow-typed folder needs to be added to [libs] in our .flowconfig.

This will likely introduce some flow errors.

Alternatively we could move to typescript 😜

Need Circle CI for testing and typing. Can set it up following the steps here: https://circleci.com/setup-project/gh/nteract/nteract.io

Our README is out of date. Deployments are now automatically happening via the now GitHub integration.

https://components.nteract.io/#source

Under the "Bring your own editor" subheader:

You can ass React component(s)...

I tried to find where the text was stored in the repo so I could open a PR, but couldn't find it anywhere.

Getting an issue for something that @MSeal, @wadethestealth, and I recently discussed on one of the team sites.

The nteract website hasn't undergone any major content changes for a couple of years now. In that time, the nteract ecosystem has evolved a ton. We now have a popular ecosystem of libraries, like Papermill and Scrapbook, a standardized JavaScript SDK for building notebook UIs, and a growing set of applications. We need to update our website to reflect these changes.

I went through each page of the website and made an audit of the segments of content that need to be changed.

• Update text in hero segment to highlight core SDK first then desktop app second
• Update three content headings in the homepage
  • First content section should be about core SDK
  • Second content section should be about applications like Hydrogen, nteract desktop, nteract play
  • Third content section should be about libraries like Papermill and Scrapbook
• Update Desktop heading to "Applications"
  • Update text on heading to be "A beautiful notebook experience everywhere"
  • Update first content section to be about desktop app in general
  • Add section for nteract web app (currently in development but once it is done we should add it here)
  • Add section for nteract Jupyter extension
• Update Atom heading to "Hydrogen"
  • I'll let @wadethestealth figure out what specific things to add under here.
• Update components section to "core SDK"
  • First section should be about core SDK
  • Second section should be about React components
• Add section for nteract libraries
  • First section should be about papermill
  • Section section should cover scrapbook
  • Extended section on bookstore, etc.
• Update mission statement under "About" to be a paragraph and remove "We're people, not software" label
• Fix whatever is going on with the contributors list that keeps breaking

We should include a Scala kernel on the https://nteract.io/kernels page.

/cc @alexarchambault -- What could we put in the install lines like we do for Python, R, etc.?

It would be lovely to put all the links from https://github.com/nteract/examples into an Examples page

On a /kernels page, list out kernels we support well:

• IPython
• IJavaScript
• IJulia
• IRkernel

I'm imagining a toggle to go between languages. I'll make a mockup later on.

We should have a page linking to slidedecks, videos (when available) and blog posts.

Right now we show double-click to open. What else should we show to people?

Note that it's "as simple" as creating another React component like the other ones:

Ideas:

• Compatibility with Jupyter Notebooks
• vdom support
• Other output type support

For other ideas, also check out @captainsafia's post about the notebook

Quite a few nteract projects already use prettier. We should use it here as well.

For reference: nteract/commuter#158

We show the NumFocus logo in the bottom left corner, we should link to numfocus.org as well.

I followed the instructions for installing the cling (C++) kernel here:

• https://nteract.io/kernels/cplusplus

Unfortunately, the kernel that is installed with these instructions is a Python one, not a C++ one.

If I source activate cling, then use jupyter notebook, then I see multiple C++ kernels labelled C++11, C++14, C++17. If I use jupyter lab, I see none of these. If I use nteract, I see none of these. Is there a way to expose these kernels to nteract?

Mostly to reduce the number of binary deps, build standalone reusable components.

From discussion on Slack #random:

the nteract project, a set of JavaScript packages and applications for interacting with the Jupyter ecosystem

I also love this description of nteract. I've always thought of nteract as a set of tools first and a collection of applications of those tools (desktop, jext, hydrogen) second

This seems like a good thing to incorporate on the website landing page.

cc/ @MSeal @BenRussert

nteract.io:

This PR:

(Pardon the little 1.00 on the screen there, that's a chrome extension for video speed controller)

Anywho, I say this is fit to ship. Let's go for it!

Originally posted by @rgbkrk in #54 (comment)

Loading over https can be slow as Chrome complains:

Mixed Content: The page at 'https://nteract.io/' was loaded over HTTPS, but requested an insecure prefetch resource 'http://nteract.io/'. This content should also be served over HTTPS.

and indeed

  <!-- Prefetching, preloading, prebrowsing -->
  <link rel="dns-prefetch" href="//nteract.io/">
  <link rel="preconnect" href="http://nteract.io/">
  <link rel="prefetch" href="http://nteract.io/">
  <link rel="prerender" href="http://nteract.io/">
  <link rel="subresource" href="/assets/stylesheets/main.css">

Though I haven't found this in the source code.

The page /desktop does not contain a <title> element.

Note that nteract.github.io doesn't have "Desktop", "Atom", "Kernels" links.

nteract.io:

nteract.github.io:

Here's all the bits we need, though you'll probably have to change _document.js

gtag.js

// @flow
// nteract.io's GA Tracking ID (I just set this up)
export const GA_TRACKING_ID = "UA-129108362-1";

/**
 * pages/_document.js will load the google analytics adapter and use
 * nteract.io's tracking ID above.
 *
 * The following functions are helpers for submitting events and page views
 *
 * Usage:
 *   gtag.event({
 *      action: 'clone',
 *      category: 'workflow',
 *      label: 'test'
 *   })
 *
 */

// https://developers.google.com/analytics/devguides/collection/gtagjs/pages
export const pageview = url => {
  window.gtag("config", GA_TRACKING_ID, {
    page_location: url
  });
};

// https://developers.google.com/analytics/devguides/collection/gtagjs/events
export const event = ({ action, category, label, value }) => {
  window.gtag("event", action, {
    event_category: category,
    event_label: label,
    value: value
  });
};

pages/_app.js:

import * as React from "react";

import App, { Container } from "next/app";
import * as gtag from "../gtag";

import Router from "next/router";

Router.events.on("routeChangeComplete", url => gtag.pageview(url));

// This is the default setup, only placing it here to make the container nicer
// to work with
export default class MyApp extends App {
  static async getInitialProps({ Component, router, ctx }) {
    let pageProps = {};

    if (Component.getInitialProps) {
      pageProps = await Component.getInitialProps(ctx);
    }

    return { pageProps };
  }

  render() {
    const { Component, pageProps } = this.props;

    return (
      <Container>
        <Component {...pageProps} />
      </Container>
    );
  }
}

_document.js

import React from 'react'
import Document, { Head, Main, NextScript } from 'next/document'
import flush from 'styled-jsx/server'

import { GA_TRACKING_ID } from '../lib/gtag'

export default class extends Document {
  static getInitialProps ({ renderPage }) {
    const { html, head, errorHtml, chunks } = renderPage()
    const styles = flush()
    return { html, head, errorHtml, chunks, styles }
  }

  render () {
    return (
      <html>
        <Head>
          {/* Global Site Tag (gtag.js) - Google Analytics */}
          <script
            async
            src={`https://www.googletagmanager.com/gtag/js?id=${GA_TRACKING_ID}`}
          />
          <script
            dangerouslySetInnerHTML={{
              __html: `
            window.dataLayer = window.dataLayer || [];
            function gtag(){dataLayer.push(arguments);}
            gtag('js', new Date());
            gtag('config', '${GA_TRACKING_ID}');
          `}}
          />
        </Head>
        <body>
          <Main />
          <NextScript />
        </body>
      </html>
    )
  }
}

This is a tiny ops log / plan. I'll update it as I go along.

Steps:

1. Deploy nteract.io following guidance
2. Deploy slackin to nteract account
3. Attempt to switch domain
4. Set DNS for deployments in (1) and (2)

On the contributor section, members are not being populated
https://nteract.io/about

Here is one of the older issues that was closed but might help someone debug.
#101

Node 10.12 does not support a clean install out of the box. Latest LTS node 8.12.0 works.

• large white space in first section
• contributors gets cut off at about 1/2 the people and unable to scroll for rest

I've enabled imgbot to send us PRs to suggest best practices for image page load.

We could have snapshot testing to ensure that regressions are less likely to happen.

Click tests are harder, and probably would benefit from automated testing.

Nitpicking but when inertia-scrolling you can get some white at the top of the page (or bottom).
You can have a "cleaner" website by setting the background color as being the same as the header:

See example:

That's screwing other sections which needs to be colored, so no PR (for now), and the same happen in the footer at the bottom of the page... Not sure how to deal with it... gradient maybe ?

It looks like some kernel options are hidden behind the terminal image on the kernels page. You see the bug, if you visit https://nteract.io/kernels and reduce the screen width to about 800px.

I was able to reproduce on both Chrome and Firefox.

Screenshot of the bug

I believe that most users follow one of the two modern methods of managing their Python installations: virtualenvs or Anaconda environments. Hopefully almost nobody runs

sudo pip install numpy

nowadays anymore.

For those users I find that the instructions on https://nteract.io/kernels/python follow an installation model that is not suitable to them.

The instructions ask the user to install an ipython kernel in their home directory using

python -m ipykernel install --user

This kernel will then become visible to all Hydrogen instances everywhere and overrides all other kernels in all other virtual environments, even if a local kernel is available. This can can create very messy situations with code inside a virtualenv run by an interpreter outside of it. And uninstalling an ipykernel kernel is much much harder than installing one.

The way I understand it, these instructions are only meant for users that do not run Atom from the commandline, and thus outside a virtualenv (which, admittedly, is probably a large number). I see that for them these instructions may be the only way to get Python kernels to work.

However for users that are using virtualenvs and are running Atom from inside one, these instructions can be very destructive. Instead of following the instructions, all they had to do was run

pip install jupyter

I believe the page should show more of the common possible scenarios and have installation instructions for them.

Would be great to add to the about.

I am starting to become really suspicious about this project. In the past, I got some virus warnings about its installer and now the website is down. What's wrong there?

When running the site, next.js throws the following warning:
Warning: <title> should not be used in _document.js's <Head>. https://err.sh/next.js/no-document-title

It looks like our Next.js setup can be greatly simplified since we're now fully on styled-components

This issue was generated by todo based on a TODO comment in f20da86 when #41 was merged. cc @nteract.

From https://nteract.io/atom it would be great if there were a link to the Hydrogen documentation website https://nteract.gitbooks.io/hydrogen/. Currently I have to remember to go to the Github repo to find the link on the Readme.

I'm less sure about this but it might also be useful for https://nteract.io/desktop to have a link to something like https://components.nteract.io/?

Our server currently serves robots.txt and sitemap.xml.

Both files aren't generated at the moment. We should either remove this behavior or correctly generate them.

Copying over from our meeting minutes, I thought this would be good to track here.

• add landing page for jupyter server extension at /web
• edit front page to mention web, desktop, components, hydrogen
• feature Data Explorer, Papermill, workflows

404 Response on all kernel documentation except python.

Convert kernels section to use the components created in the /components route. This enables us to present richer data in the kernels page.

Running into following error post npm install, and while trying run - npm run start

/code/nteract.io/next.config.js:5
    webpack: (config, { dev }) => {
                      ^

SyntaxError: Unexpected token {
    at exports.runInThisContext (vm.js:53:16)
    at Module._compile (module.js:404:25)
    at Object.Module._extensions..js (module.js:432:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:311:12)
    at Module.require (module.js:366:17)
    at require (module.js:385:17)
    at loadConfig 

I really think it would help Nteract users if there was a help button on the footer or top of the page. We direct users to use Nteract and they are not aware of the Slack channel where questions can be asked.

I'm trying to make some plots with Holoviews, but the plots are placed at the bottom of my nteract window, outside the cell that generated them, and I can't seem to clear them.
I'm not sure if this is an issue with Holoviews or nteract, but things work as expected when running the notebook in my browser, so I'm guessing it's nteract.

I'm using nteract version 11.7 as an AppImage on Fedora 28. I tried the 11.6 AppImage as well, and the problem was there too.

Some example code:

import numpy as np
import holoviews as hv
hv.extension('bokeh')

data = [(chr(65+i), chr(97+j),  i*j) for i in range(5) for j in range(5) if i!=j]
hv.HeatMap(data).sort()

The plot gets shoved way to the bottom of the notebook, so much so that at normal zoom, I can't even capture the cell and the plot in the same screenshot. If I zoom out (Ctrl -), the distance between the cell and the plot actually increases - the cells seem to shrink towards the top while the plot shrinks towards the bottom. If I zoom in (Ctrl +) to comical levels, I can manage to get the bottom of the cell and the top of the plot in the same screenshot (attached).

Any idea what's going on?

It looks like the About page isn't displaying the full alpha list of contributors (contributors with gh user names >= 's')

Or maybe I've been demoted 😉

We should no longer proxy anything from nteract.github.io, as currently nteract.github.io redirects to nteract.io. This will cause a redirect loop.