Reagent | Experiments for future reactive libraries | Reactive Programming library

I've constructed a data.frame using the inefficient code below. Can you improve it, noting that if you can think of a better starting point, please include that answer.

My code takes data from the first two data frames and combines them to give the third. The first data.frame is a grid of 1s and -1s representing low or high values. The second data.frame includes all the information for me to calculate the high or low values. Note that each column has similar calculations but the calculation may differ from column to column.

This is my jQuery script so far:

I am sending a map via a POST using cljs-http and I want to receive it with a Compojure POST handler and store it in a database. I'm totally confused as what I am supposed to do with the map in transit... there are a lot of options... pr-str, prn, read-string and on and on. On the sending side, it seems cljs-http might coerce my map into edn on it's own. I have no way of knowing without a working Compojure handler. Actually, to be honest, I'm still a bit unclear on what edn is. It's Clojure's data format, that I gather. As far as I can tell, what I am dealing with is actually just converting edn into a string format for shipping and back out. However, the :data value on the request is an object, not a string. A working example of a cljs-http POST and a Compojure POST handler that allows me to work with the original map sent by the client would be awesome. I don't know the purpose of :edn-params in my

This might sound a stupid question, but how do I turn a piece of hiccup into html in a ClojureScript reagent application?

I want something like this :

Using clojurescript 1.10.758 and reagent 1.0.0, I am running into an error in which a file index.js tries to reference $jscomp, which is not defined.

I've seen a number of Stackoverflow and Github issues related to $jscomp being undefined in the context of shadow-cljs, but I'm not using that.

The problem occurs when I use a development mode build with figwheel (using Leiningen with cljsbuild and the figwheel plugin), and also occurs if I use cljsbuild for a once-only development build. Strangely, if I use webpack to create a bundle, the problem does not occur.

Before I tried to make webpack work, I did have working code without webpack. Something I changed seems to have affected the non-bundled build. The only change I can thing of was to install react and react-dom using npm, and exclude those packages from reagent in Leiningen's dependencies. But undoing the exclusion didn't make the non-bundled code work again.

Any suggestions for how to cause $jscomp to be defined when it's first needed?

I'm trying to do some (simple) plots in Clojure and it seems like Oz would be a great long term solution. However, I'm running into problems just trying to get the example code to run. PS I'm completely new to ClojureScript / Reagent / Hiccup ...

On the website it shows the following example code:

What would be an elegant way to find the ba4 values for :foo in a data structure like:

[[{:foo "ba1"}{:foo "ba2"}] [{:foo "ba3"}{:foo "ba4"}]]

and add the :bif "baf" key and value so the map(s) containing the key/value I sought so I had:

[[{:foo "ba1"}{:foo "ba2"}] [{:foo "ba3"}{:foo "ba4" :bif "baf"}]]?

This is the main thing I am trying to figure out; I do wonder about how you would do this if there were multiple levels of nested maps, but that will be next for me to understand.

I have, I think, seen various libraries out there for tackling this sort of thing and I'm not at all against using them, though they would have to work in ClojureScript.

I am trying to update a Reagent atom such that classes will change on a UI element.

I am struggling quite a bit with making the labels of a plot look a certain way. I am using ggplot2 and tidyverse.

This is what I have:

I would like to have two headlines (=name) for the legend, one for the cell type HCT, and one for the cell type RKO. Then for HCT and RKO each, I want to have the legend for the Reagent with the respective color, linetype and shape. So basically, I want to break up the color legend into two separate legends. I just can't wrap my head around how to code it. Here is a drawing of what I would like to have instead (for the figure legend; please imagine the orange square is filled in):

Do I need to change my geom_line and geom_point code in order to achieve the legend style I'd like? Or is there another way to do it? I tried searching for a way to do it but couldn't find anything (maybe I am just not using the correct terms). I already tried following what was done here: How to merge color, line style and shape legends in ggplot and Combine legends for color and shape into a single legend but I couldn't get it to work. (In other words, I tried changing scale_shape_manual etc. to accommodate my wishes with no success. I also attempted to use interaction())

Note: I decided not to use facet_wrap since I want to show both of the cell types on the same plot. The plot of the real data looks a little different and it's not as overwhelming. I was able to successfully plot a "facet_wrap" plot with ggpubr.

Note2: I also did not use stat_summary() because I need to take the mean of the same reagent concentration, reagent and cell type. With my data, I did not find a way to make stat_summary work.

Here is the code that I currently have:

What is the best way in Clojure (actually ClojureScript) to accomplish the following:

• A mutable collection of elements xs, each of type T (T is a map, if we want to be specific)
• A "selected element" among them, x, which alternates between various xs or a "none selected" state. The number of selected items is either 1 or zero.
• xs and x will have event listeners listening to them: some listen for any change in xs; some listen to any change in x. (That is, they listen for both an update in the state of the selected item, as well as a switching of which item is selected.) (I'm actually using the Reagent wrapper for React, so this affects when components will update.)
• Methods to edit x (don't need to edit non-selected xs, though need to be able to add new xs)
• Given an element in xs, a method to select it. (Including the case of select "none")

Possibilities I've thought of so far:

1. Make xs an atom which is a vector of Ts, and make sure every T element knows its own index. Replace x with x_idx which stores the index of the selected item (or nil). Selecting an element just means getting its index and changing x_idx to that index, both of which are constant-time ops. Downsides to this are that it makes the structure a little less elegant and simple: I'm always passing around the index of the selected item, and every operation I want to do has to work with the index instead of the item itself, as it would like to. Since my T objects are small, it would be more pleasing to just have a "selected object" variable which is of type T.
2. Make xs an atom vector and x an atom which stores a T. Now x is a T, which is nice, but when I want to update info about x, I have to make two calls to reset! or swap!: one for x and one for the element in xs which x represents. Inelegant for obvious reasons. And I have to do these in quick succession or else there will be an inconsistency in the data: in between the two calls, the event listeners listening to xs will see the selected item in one state, and the ones listening to x will see it in another state.
3. Give T elements a field to tell if they're selected or not, and get rid of x. This would be right if multiple items could be selected at once, but since only one item can be selected, it just makes me do a linear search every time I want the selected item, which sucks.

The point of this question is not to solve some particular issue (any of the possibilities above work fine for the scope of the small project I'm working on), but to learn about Clojure data structures. This seems like a common enough situation that there would be some structure around it. Responses along the lines of "you should be trying to answer a different question entirely..." are welcome.