Am I going into the weeds too much?

I think this tutorial, of any, should be more in the weeds since this is genuinely a new concept to most in robotics so very few will even know the vocabulary we're using.

Do I need more pictures or gifs to show what is going on?

These are good, take some groot screen shots or use the BT visualization script in /tools of Nav2 to generate some visualizations of the trees so people can see them. The XML is good, but its much easier for people to understand the tree structure looking at... well... trees. For instance in the "Navigate With Replanning and Recovery" section, you show the main Bt first thing, I think you should show the actual tree before that XML block so they can see how the tree goes into the XML. After that point, for any "complete" BT's I'd show the visualization of the tree. For any sub-trees or specific calls, show the XML.

Reviewing now for more detailed commentary...

Just merged in something that walks through the custom BT nodes, wanted to push it here and make sure it renders properly before I go through and re-edit the navigate_w_replanning_and_recovery portion of the tutorial.

Does this new page (https://navigation.ros.org/behavior_trees/index.html) change the direction of this PR / this section I am about to go and edit?

I think it does, we will have a bit more duplicate ideas

Additionally, need some advice on how to get these images rendering nicely / professionally. (Currently just using a paint program to slap the labels “(running), (idle), (failure)” etc.

I will focus on this issue last, the images for now can serve as a “sketch” of what the end product would be .. but I cannot imagine it being this in the final version.

The Groot trees took a lot of manual manipulation to get into an aspect ratio that would look good on the webpage so I used the bt2img utility.

can you give an example? I think how we've done it elsewhere is visualizing the behavior trees using Groot / that script you mentioned.

can you give an example? I think how we've done it elsewhere is visualizing the behavior trees using Groot / that script you mentioned.

Here is the latest build which has those trees: https://714-240847415-gh.circle-artifacts.com/0/html/tutorials/docs/custom_behavior_tree.html

Right now these are generated using the bt2img.py utility but manually marked up in a paint program for labels.

I think the python script does a reasonable job at making the tutorial-ready tree.

I might add the state on top of the BT node rather than overlapping/below so that its not overlapping in the box itself. Another option would be putting a border around the BT nodes rather than the (words). Green = success, red = failure, grey = idle, blue = running or something. While I think that would be more visually pleasing, I think the words, if put on top, would be totally sufficient and maybe better to just be more explicitly clear.

@SteveMacenski I have significantly edited this now to address your first round of comments, and I think it is worth another round of review from you before I move on to the "custom" portion of the tutorial.

Hey! Just touching base on this

Hey @SteveMacenski , I apologize this slipped.

I believe I have addressed everything else so far, except I just posed some discussion for the PipelineSequence part of the tutorial.

I suggest that we punt the "custom BT" portion of the tutorial to a separate PR so we can at least get the work we did thus far released, as I am afraid I cannot predict a timeline for the other "custom BT" portion of this ticket.

If you agree, I can remove all reference to the "custom BT" part of the tutorial and we can pick that up at a separate time / someone else could pick it up if they so choose.