abline!

[diegozea]

Hi!
It would be great to have an abline function, similar to R.
Best,

[tbreloff]

Right now you can do hline, vline, or add a regression term with smooth. Here's some examples:

abline could be done manually of course, but I can see the use for it, so I'll leave this open.

[pkofod]

Just to be clear, how would you create the line y=0.5+0.1*x using the mentioned functions? I mean, as abline(a=0.5, b=0.1) does.

[tbreloff]

Right now you would have to add the line manually:

[diegozea]

Also you can do
plot!(0:100, x -> 0.5 + 0.1 * x, linetype=:l)
but it's annoying when you don't know the limits for x.

[tbreloff]

Or better yet: plot!([0,100], x->0.5+0.1x) which doesn't plot the [1,2,3,...,99] points.

[pkofod]

Okay, so I just misunderstod the "could be done manually of course". So I guess something like

function abline!(a, b, ...)
    fetch x limits
    plot!([xmin, xmax], x-> b+a*x)
end

I can do a PR if this would be useful to have.

[tbreloff]

Sure if you want to give it a go... this probably belongs in recipes.jl

On Tue, Feb 2, 2016 at 10:53 AM, Patrick Kofod Mogensen <
[email protected]> wrote:

Okay, so I just misunderstod the "could be done manually of course". So I
guess something like

function abline!(a, b, ...)
fetch x limits
plot!([xmin, xmax], x-> b+a*x)end

I can do a PR if this would be useful to have.

—
Reply to this email directly or view it on GitHub
#125 (comment).

[tbreloff]

Please merge the PR into the dev branch, not master. It makes my flow a
little easier. Thanks!

On Tue, Feb 2, 2016 at 11:03 AM, Tom Breloff [email protected] wrote:

Sure if you want to give it a go... this probably belongs in recipes.jl

On Tue, Feb 2, 2016 at 10:53 AM, Patrick Kofod Mogensen <
[email protected]> wrote:

Okay, so I just misunderstod the "could be done manually of course". So I
guess something like

function abline!(a, b, ...)
fetch x limits
plot!([xmin, xmax], x-> b+a*x)end

I can do a PR if this would be useful to have.

—
Reply to this email directly or view it on GitHub
#125 (comment).

[joshday]

Could there be a method for plotting a function which takes the x limits from the current plot? It's a common use case, and a little more flexible than abline. plot!(x -> .5 + .1x)

[pkofod]

@joshday how should we then choose how to sample the domain?

[joshday]

I'm not sure I understand the question. Maybe my comment needs clarification. This already works:

scatter(randn(100))
plot!(x -> .5 + .1x, 0, 100)

Intuitively, I'd expect this to work:

scatter(randn(100))
plot!(x -> .5 + .1x)

The second two arguments for plot!(f::Function, ...) could be optional using whatever trick you're using with abline to get the maximum and minimum of current(). It'd be a very short PR. I'd be happy to add it.

[pkofod]

My bad! I get it.

[tbreloff]

Josh this is a nice idea. To implement, you should implement this existing function: https://github.com/tbreloff/Plots.jl/blob/master/src/plot.jl#L415

Look at the other methods defined nearby if you get confused, but you just need to call the same method with an x-value.

On Feb 2, 2016, at 4:38 PM, Josh Day [email protected] wrote:

I'm not sure I understand the question. Maybe my comment needs clarification. This already works:

scatter(randn(100))
plot!(x -> .5 + .1x, 0, 100)
Intuitively, I'd expect this to work:

scatter(randn(100))
plot!(x -> .5 + .1x)
The second two arguments for plot!(f::Function, ...) could be optional using whatever trick you're using with abline to get the maximum and minimum of current(). It'd be a very short PR. I'd be happy to add it.

—
Reply to this email directly or view it on GitHub.

[pkofod]

Should I add something like

xmin(plt::Plot) = minimum([minimum(plt.seriesargs[i][:x]) for i = 1:length(plt.seriesargs)])
xmax(plt::Plot) = maximum([maximum(plt.seriesargs[i][:x]) for i = 1:length(plt.seriesargs)])

in #126 if @joshday is going to look up the minimum and maximum value as well in his addition? So we avoid several different approaches of finding the min and max throughout the code.

[tbreloff]

That sounds good. It might look nicer as:

minimum([minimum(d[:x]) for d in plt.seriesargs])

(Sorry I'm such a nitpicker)

On Feb 2, 2016, at 4:53 PM, Patrick Kofod Mogensen [email protected] wrote:

Should I add something like

xmin(plt::Plot) = minimum([minimum(plt.seriesargs[i][:x]) for i = 1:length(plt.seriesargs)])
xmax(plt::Plot) = maximum([maximum(plt.seriesargs[i][:x]) for i = 1:length(plt.seriesargs)])
in #126 if @joshday is going to look up the minimum and maximum value as well in his addition? So we avoid several different approaches of finding the min and max throughout the code.

—
Reply to this email directly or view it on GitHub.

[pkofod]

Much better, please nitpick - it's partly why I like to contribute where I cannot just commit myself - to learn from and be inspired by others more skilled than me.

[tbreloff]

Cool. And to add to the convenience, maybe implement "Base.extrema(plt::Plot) = ..."

On Feb 2, 2016, at 5:18 PM, Patrick Kofod Mogensen [email protected] wrote:

Much better, please nitpick - it's partly why I like to contribute where I cannot just commit myself - to learn from and be inspired by others more skilled than me.

—
Reply to this email directly or view it on GitHub.

[pkofod]

abline! Is now on the dev branch @diegozea .

[tbreloff]

The relevant PR is #126.

[diegozea]

¡That's great! Thanks @pkofod :D