shiny-le

Lake Erie R Shiny

See:

• Lake Erie - 2013 - RShiny App on shinyio.

Environment

These were tested in R Studio.

Required libraries for reformatting

• ncdf4
• lubridate
• rlist
• dplyr

Required libraries for R Shiny App

• shinydashboard
• RColorBrewer
• leaflet
• dplyr

Get the repo

You can get the repo with git clone:

git clone https://github.com/l3-hpc/shiny-le.git
cd shiny-le

Or by downloading the repo by going to shiny-le, click the green 'Code' button, then select 'Download ZIP'.

Open in R Studio

I did not test these as batch scripts.

In R Studio, in the 'Files' window(bottom right panel, first tab), navigate to the shiny-le folder(this repo). Click 'More'(next to the gear icon), and select 'Set as Working Directory'.

Open global.R, ui.R, or server.R. Any of these will have a Run App button that the top (instead of Source). Click Run App to run R Shiny locally. It should work, after installing the proper packages

Add and reformat new model files

Add FVCOM model outputs to a directory named starting with output. (The .gitignore file ignores everything starting with output.)For the example, I am using:

output_sinkout
output_nosinkout

Open ReformatModelData_allnodes.R, set the path for the FVCOM outputs. They should contain the variable "TP".

#--Model data
var1file <- "output_nosinkout/leem_0001.nc"
var2file <- "output_sinkout/leem_0001.nc"

If you want to try these out, you can get them from OSN. (Change the directory paths or move the directories to the proper place.)

wget https://renc.osn.xsede.org/ees210015-bucket01/outputs_shiny-le.tar
tar -xvf outputs_shiny-le.tar

If you don't want to write over the example rds in this repo, set shiny_dir to something else. To run these instructions in the github repo without tracking the new data, name the directory with data_. (The .gitignore file ignores everything starting with data_.)

Click Source. This should make new rds files in you shiny_dir, or overwrite the ones already in data.

Modify directory paths and labels.

If you made a different shiny_dir, change data_dir in global.R.

Open labels.R and change the variable names and titles to reflect your data.

Open global.R, ui.R, or server.R. Any of these will have a Run App button that the top (instead of Source). Click Run App to run R Shiny locally.

Additional steps for LEEM data

Using this example...

Make a directory starting with output. (The .gitignore file ignores everything starting with output.)

Try a sample file:

mkdir output_leem
cd output_leem
wget https://renc.osn.xsede.org/ees210015-bucket01/LEEM_2013_Calculated_Time0.nc

Create netCDF with just TP. (This could probably be improved...)

Copy:

ncks -O LEEM_2013_Calculated_Time0.nc TP_leem.nc

ncap2 can't add variables and divide at the same time (or at least, I couldn't figure out how), so make and add three new ZOO variables. Append to the existing file (a).

ncap2 -s"ZOO1=ZOO1/50.;ZOO2=ZOO2/50.;ZOO3=ZOO3/50.;" -v LEEM_2013_Calculated_Time0.nc TP_leem.nc
a

Add the TP variable, append to existing file.

ncap2 -s"TP=RPOP+LPOP+RDOP+LDOP+PO4T+LPIP+RPIP+ZOO1+ZOO2+ZOO3" -v TP_leem.nc TP_leem.nc
a

Make a small version, just TP:

ncks -v TP,h,iint,lat,latc,lon,lonc,nprocs,nv,partition,siglay,siglay_shift,siglev,time,x,xc,y,yc,zeta TP_leem.nc TP.nc

Use VisIt to test by opening TP.nc and LEEM_2013_Calculated_Time0.nc, with TP defined as:

DefineScalarExpression("TPtot", "RPOP + LPOP + RDOP + LDOP + PO4T + LPIP + RPIP + (ZOO1 + ZOO2 + ZOO3)/50.0")

They should be exactly the same.

The file LEEM_2013_Calculated_Time0.nc has data every 6 hours. Crop it, since I did daily outputs on Expanse: Make a small version, just TP:

ncks -d time,0,979,4  TP.nc TP_crop.nc

Wilson's files

There were 3 more timesteps in LEEM (from 6, 12, and 18 hrs probably) so I cut that:

ncks -d time,0,976 TP.nc TP_crop.nc