Published R reports with sourced scripts

[davor]

Do I have an option to publish R scripts as well when publishing reports?

I would like to do:

source('path/to/script.r')

A workable option is to import the script files with 'Get Data' -> 'Folder', and then use something like:

eval(parse(text = as.character(dataset$Content[1])))

where 'Content' is the script file transformed to text. This is OK, as long as you don't pass lots of extra data due to the 'unique(dataset)' expression.

Obviously I could also paste the whole script, but from a code-re-use perspective it is bad practice (and the script is quite long...). 

Anyway, any decent way to load scripts, or am I to hack around it?

[TomMartens]

Hey,

currently it's not possible to reuse r-script files if you are using the r-script visual approach.

If you consider to create a r based custom visual then you are able to source r script files.
I guess this will get you started:
https://github.com/Microsoft/PowerBI-visuals/blob/master/Readme.md

But you have to be aware that there is still no global library that you can use for many visuals.

Hopefully this gets you started

Regards
Tom

[davor]

Hi Tom, can you elaborate on:

"But you have to be aware that there is still no global library that you can use for many visuals."

[TomMartens]

Hey,

what I mean is this:

You can source script files, if you are creating a custom visual, but you to do it for each visual that you want to use your source file.

Meaning for each custom visual you are creating, you have to source file individually.

Hopefully this clarifies my first post a bit.

Regards
Tom

[davor]

Any way to adjust the current R-script-visual (i.e. make a new one based on the CORE R-script-visual)?

I am mainly thinking about removing `unique(dataset)`.

[Anonymous]

Hi @davor. I have the same issue. Did you find a solution?

[davor]

No.

I have one large script that I copy-paste in all the visuals, and then at the end only call those functions that are needed for the visual.

Best I could think of...

[Anonymous]

i found this solution on stack overflow

[davor]

I also tried sourcing from the web:

source('https://raw.githubusercontent.com/.../model.r')

But it seems this is explicitely blocked:

Error in file(filename, "r", encoding = encoding): cannot open the connection 
In addition: 
Warning message: In file(filename, "r", encoding = encoding): 
InternetOpenUrl failed: 'The server name or address could not be resolved'  

[davor]

Each r-visual also seems to operate in its own VM environment, so if you export something to "C:\Reports" with one visual, the other visual doesn't see it in "C:\Reports".

If this was possible, one could use one visual to parse the code from "Get Data -> Folder", and the others could load it from "C:\Reports" when available with something like:

while (!file.exists('parsed.r')) {
  Sys.sleep(0.01)
}
load('parsed.r')

Pff, getting out of ideas...