Stages
When learning any topic, it's important to be able to break it up into reasonable "chunks". No one wants to start at a wall of learning material they know will take them the better part of an afternoon to get through.
In Antidote, lessons are broken up into Stages. The intention behind Stages is to provide a logical place to "break up" lesson content. You can think of them like chapters in a textbook; where the goal of an Algebra textbook is to teach algebra, we really only care about linear equations in Chapter 1. Those not only provided bounded structures for learning, where the learner is able to more easily wrap their head around the content and feel like they've accomplished something, it also helps them understand the path in front of them.
In any lesson, Stages are defined using the stages stanza in the lesson definition file:
stages:- id: 1description: No BGP config - tests fail- id: 2description: Correct BGP config - tests pass
When a lesson is loaded in the web front-end, these stages show up as a continuum that the user can select, underneath the lab guide:
While the Stage definition seems simple, there's a lot that goes on when a user navigates between Stages by selecting something in that drop-down:
All Endpoints with a
configurationTypewill be reconfigured accordingly. This happenswhen a lesson is initially loaded
Endpoint health checks as described in Presentations are not done between stages. Presentations are static for the whole lesson, regardless of Stage. They're done when the lesson is initially loaded.
When a stage ID is omitted, the default is to load the first one, but this isn't a requirement for users. Each lesson URL includes a Stage ID, which means hyperlinks to any stage in any lesson are honored. What this means for lesson builders is that while your Stages can (and should) have a natural progression, you should not rely on users to have done something themselves in Stage 1 in order for Stage 2 to work. If you have the user accomplish a task in Stage 2, you should still overwrite all configs yourself to the expected value in Stage 2's configurations.
Each Stage has a particular directory structure that you should be aware of. As with most things involving curriculum resource definition, most of this is enforced by syrctl so you can validate this structure yourself, but here are some general rules:
.├── jsnapy_config.yaml├── jsnapy_tests.yaml├── lessondiagram.png├── lesson.meta.yaml├── stage1│ ├── configs│ │ ├── vqfx1.txt│ │ ├── vqfx2.txt│ │ └── vqfx3.txt│ └── guide.md└── stage2├── configs│ ├── vqfx1.txt│ ├── vqfx2.txt│ └── vqfx3.txt└── guide.md
Each stage must have a corresponding directory called
stage<N>whereNis the stage ID.Each stage directory must have a
configsdirectory, where all of the files related to Endpoint configuration should be kept.Each stage directory must also have either a markdown-based lesson guide, or a jupyter notebook to be used for the same. We'll get into the differences between these in the next few sections.
All NRE Labs lessons come with lab guides. These are meant to be instructions the learner can follow along with so they're not spinning your wheels, wondering what to do with a lesson. It is also meant to include snippets of code or commands for them to execute, or have executed for them.
TODO - State in the lesson guide docs to not put any top-level headers, we'll do that for you. Just start right into your first paragraph.
There are two options for lesson guides in Antidote today:
Markdown <Writing Lab Guides with Markdown>Jupyter Notebooks <Writing Lab Guides with Jupyter Notebooks>
You can choose either of these options on a per-stage basis. This means that Stage 1 can have a Markdown lesson guide, Stage 2 a Jupyter notebook, and Stage 3 back to Markdown, if you want. The NRE Labs NAPALM lesson <https://labs.networkreliability.engineering/labs/?lessonId=13&lessonStage=1>_ is a good example of a lesson that leverages both options. The sections below will explain how to use either option.
The simplest option by far is to write lab guides with Markdown. This is an extremely popular, simple formatting syntax for creating rich documents from plain-text sources. Most of the time when you see a README file on a GitHub repository, the chances are very good that it's written in Markdown. What Github does is translate the raw text of the file into richly formatted, rendered versions appropriate for viewing in a web browser.
To enable this same experience for Antidote, lesson guides can be written in Markdown, and antidote-web will take care of translating the source file into HTML to be presented to the user.
You don't have to use self-medicate to preview the HTML version for your Markdown documents. While every Markdown renderer is a bit different, and there might be some minor differences within the Antidote front-end, if you're just looking for some basic HTML preview functionality, there are plenty of tools to do this:
Keeping in the spirit of doing everything in the browser, Dillinger is very handy for
working on lesson guides with a constant preview.
Most popular text editors also have markdown preview functionality built in or available via plugin. For example, there's good support for this in VS Code.
There are a number of CLI tools available as well for doing the conversion yourself, such as
Pandoc, if you are so inclined.
While native Markdown is perfectly fine, there's one feature built into antidote-web you should be aware of that really uplevels your lesson guide. In Markdown, you can wrap a bit of text with triple-backticks, and it will preserve the formatting you use within that block. This is very useful for code or CLI commands, where the structure is very important.
```echo "Hello World!"```
Markdown-based lab guides include the ability to add a "Run this snippet" button to automatically run the contents of a code snippet in a given terminal tab. This is extremely useful and recommended to allow users to quickly execute your examples without having to type it out themselves.
To do this, the lesson author only needs to add some HTML underneath each snippet:
```echo "Hello World!"```<button type="button" class="btn btn-primary btn-sm" onclick="runSnippetInTab('linux1', this)">Run this snippet</button>
Most of the HTML shown above can remain the same, but in the above example linux1 refers to the tab where this snippet should be executed. The front-end will switch to the tab named accordingly and paste that text automatically. So, you'll need to edit this to point to the tab you want.
Also, when you're adding a snippet to a lesson guide, sometimes you may want an extra newline run at the end. For example, if you are executing some Python code, and your snippet ends on a loop, or a conditional, you need an extra newline to get the interpreter to understand you're done defining the loop.
The solution to this is to use <pre> tags in lieu of the traditional triple-backtick for embedding code in Markdown:
<pre>echo "Hello World!"</pre><button type="button" class="btn btn-primary btn-sm" onclick="runSnippetInTab('linux1', this)">Run this snippet</button>
These are rendered exactly the same way in the lesson guide, but the latter is interpreted much more literally when being pasted into the terminal window, meaning the extra newline is executed like any other character.
Finally, when you have a lesson guide ready, place it in the stage directory as guide.md. This is where the Antidote platform will look for this lesson guide.
This section is focused on lesson authors looking to use Jupyter notebooks in the creation of a lesson. See here if you're looking for an overview of how to use lesson guides in NRE Labs.
Many folks who have invested time in education on automation or related topics have some experience with Jupyter notebooks. Jupyter notebooks are awesome in their own right, and it's not fair to force folks to convert that content to Markdown just to get it working with Antidote. So, Antidote natively supports the use of Jupyter notebooks as lab guides in lieu of a Markdown-based guide.
Fair warning - Jupyter notebooks offer a lot more functionality than Markdown-based lab guides, but they do add a layer of complexity as a result. So, if you're starting from scratch, it's probably best to start with Markdown-based lab guides. However, the choice is yours.
Even cooler - any lesson that uses a Jupyter notebook is automatically provisioned a background Endpoint dedicated to running that notebook, that is run alongside all other Endpoints for that lesson. That means that you can take advantage of Kubernetes DNS when referring to other Endpoints in your notebooks. If you want to send a REST API request to an Endpoint with the name of webapp, you can refer to it via the hostname webapp, right in the notebook. No need to figure out IP addresses for stuff.
To use a Jupyter notebook as a lesson guide in an Antidote lesson, you need only add the line jupyterLabGuide: true to each Stage that requires it in your lesson definition. Here's an example of a lesson that uses Jupyter notebooks for stages 1, 2, and 4, but uses the traditional Markdown format for stage 3:
stages:- id: 1description: Get device factsjupyterLabGuide: true- id: 2description: Get information with NAPALM "getter" functionsjupyterLabGuide: true- id: 3description: The NAPALM Command-Line Utility- id: 4description: Make configuration changes with NAPALMjupyterLabGuide: true
When you do this, you will need to make sure that a jupyter notebook titled notebook.ipynb is in all relevant stage directories. This obviates the need for a guide.md file.
If you're starting from scratch and wish to write a Jupyter notebook, your best bet is to follow an online Jupyter notebook tutorial to get it started. Or you can copy one from an existing lesson into your lesson, and once spun up, you can use the Jupyter GUI to edit and download the revised notebook.