odinodin / data-frisk-reagent Goto Github PK

Once my data structures start to get fairly complex, it can be a little difficult to navigate down to the piece of it that I'm interested. It would be nice if data-frisk had an option to sort the keys in maps, rather than just presenting them in map sort order.

Would be nice to have an 'expand all' button next to 'collapse all' :)

It would be good if can add some high z-index to the main div to prevent that elements from the page overlap the content in the frisk.

If I see it right it should be added here https://github.com/Odinodin/data-frisk-reagent/blob/master/src/datafrisk/core.cljs#L17

Would it be possible to add unique CSS classes to the different nodes?

It would be nice to be able to override some of the styles and be able to customize the layout / colours. In my current project I would like to be able to override flex-direction on the KeyValNode so that some of the heavily nested data is easier to see.

Definitely not a critical issue and I can work around it, but I'm also happy to submit a PR if there's any desire for it.

Have you given any thought to allowing users to change the values?
(when viewing ratoms or mutable things)

This would make the context of the data you're inspecting much clearer (visually so, even) when you want to inspect more than one piece of data in a UI. This could even (perhaps) for some apps be used for default error components :-)

Hi, I wrote some features which I personally find useful, if you like any of them I can clean up / fix the code and make a PR. Here is what I added:

1. A text input for searching by path. The tree automatically expands to match whatever path is inputted. I put this in a modal panel so that it's always visible when the tree is large.

2. Clicking on a node will display its path in the modal panel.

3. Ctrl-clicking on a node will copy its path to clipboard.

Here's a demo (I added a large map for testing)
Here's the code

Right it seems (correct me if I'm wrong) that you can only really frisk one data object at a time. This makes it difficult to inspect data nested at some place within an app, where the component with the data in question is representated more than once in the UI. While #1 addresses this to some degree, it does so at the expense of potentially wrecking the UI (why I think the overlay functionality of data-frisk is so valuable).

What would be ideal is for you to call something like d/DataFriskShell that would queue some data for inspection, and ultimately produce a single component with all of the inspected chunks of data organized in some way. Perhaps you could even drop in anchors that let the chunks of data you inspect point to the components with the data you're inspecting.

A hacky/partial/non-optimal implementation would probably still be helpful, but there's probably a lot you could do to make things really slick here.

Is react-with-addons really necessary? Looks like it is not because it is not :required anywhere.

Currently most users will end up with multiple Reacts (in classpath) because Reagent etc. depend on normal version and readme here doesn't ask user to exclude those. At minimum the readme should state that user has to add exclude normal React.

Even better would be to use normal React, or not depend on React directly and just depend on Reagent.

flexsurfer/re-frisk#12

I think it is because button under the other div