[ncl-talk] emphasize every parts of a line with different colors according to its y value.

[David Brown]

HI Alan,
This is a good reminder that we need to document gsSegments and
gsColors properly. We will try to do that soon.
 -dave

On Fri, Dec 30, 2016 at 12:59 PM, Alan Brammer <abrammer at albany.edu> wrote:
> ok, so there are multiples ways to plot a line.
> Here’s my method which is over complicated for a single line but has massive
> speed benefits if you want to plot multiple lines.
>
> Similar to the previous replies, you can use the relatively undocumented
> gsSegments and gsColors resources to do pretty much what Rick describes but
> with better scaleability.
>
> Attached function with hopefully sufficient comments to explain whats going
> on and includes an inline demo function at the bottom I put together a while
> back.
>
> To run the demo either load the script and call colored_line_demo() or
> comment out lines 127/128 and then `ncl add_colored_line.ncl` will run it as
> the main script.
>
> Note to devs, attached in this script is an extended version of
> get_color_index() to accept an array of "value". Could this be included in
> the main function, it's a trivial edit and retains backwards compatibility.
>
>
>
> Alan.
>
>
> On 30 Dec 2016, at 14:21, Guido Cioni <guidocioni at gmail.com> wrote:
>
> I was going to reply with a similar recipe :)
>
> Basically, when drawing the line you have to split every segment and assign
> a color based on the average value of the y in that segment.
> This is just an hint, you will have to complete the code.
>
> Assuming you have x,y which are your data values:
>
> levels=fspan(min(y), max(y), 20)  ; define the levels for the colors, you
> can change this but it will always depend on the number of segments
>
> cmap = read_colormap_file("amwg_blueyellowred") ; define the colormap that
> you want to use
>
> do i=0,dimsizes(y)-1  ; loop over the segments
>    res at xyLineColors=GetFillColor(levels,cmap,avg((/y(i),y(i+1)/)))
>    gsn_polyline(wks,plot,(/x(i),x(i+1)/),(/y(i),y(i+1)/),lres)
> end do
>
> Guido Cioni
> http://guidocioni.altervista.org
>
> On 30 Dec 2016, at 17:35, Rick Brownrigg <brownrig at ucar.edu> wrote:
>
> Hi,
>
> I don't know of any way of doing this other than to draw the line piecewise.
> What about something like:
>
> datadims = dimsizes(data)  ; assuming your data is 2xN
> numlinesegs = datadims(1) - 1
> linesegs = new((/ numlinesegs/), graphic)
> do i=1,numlinesegs
>   color = ....appropriate logic to determine the color here...
>   res at gsLineColor = color
>   linesegs(i) = gsn_add_polyline(wks, plot, data(0, i-1:i), data(1, i-1:i),
> res)
> end do
>
> This is not particularly efficient, and there may be better ways of doing
> this, but that should work.
>
> HTH...
> Rick
>
>
> On Fri, Dec 30, 2016 at 12:48 AM, grace <313695096 at qq.com> wrote:
>>
>> Hi:
>>   All,I am trying to  draw a line  that emphasize every parts of it with
>> different colors.
>> I have found the example xy_12.ncl which highlight the wanted part by
>> split the data.
>> But I want to draw a line which have different colors according to its y
>> numberical value,just like the plot below.
>> <0356412F at 5DA70B25.65116658.png>
>> If I split the data ,that would be a huge work.
>> Have anybody drawed plot like this ?Any information will be appreciated.
>>
>>
>>
>> _______________________________________________
>> ncl-talk mailing list
>> ncl-talk at ucar.edu
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>