We have a set of R scripts, which process some data and produce some results. We want to make these scripts available to basic users, which are not used to commandline of R - we want to provide them some nice GUI, which would allow to:
import/export data from MS Excel/Access easily (also supporting Copy/Paste if possible)
allow user to change settings/parameters of the process
should be running in MS Windows.
Is there any simple, scripting environment which would allow to develop such nice GUI for our R scripts as fast as possible? Need not necessarily be in R language.
There is the RExcel tool that incorporates R as an Excel plugin so the main interface is Excel with R doing the computations in the background. You could set up a sheet so that the user enters their data, then highlights a box and then chooses a menu item or clicks a button and the results are placed in another cell (or set of cells). Note however, that RExcel and the comunication program it user are not free.
Another option is to create your own gui function in R, then have that gui run automatically when you start R (see ?STARTUP) and set this up on the users machine. I have done this for clients before that did not know anything about R, they just double clicked on the icon on the desktop (windows), minimized the main R window when it opened, interacted with the gui that I had programmed to run (I used tcltk, but there are others) and saw the output provided.
You can get data copied from Excel by having the user select the data and click on copy, then in your program run newdata <- read.delim('clipboard') and the data will be in the data frame called 'newdata', you can use write.table(outdata, file='clipboard', delim='\t') to put the data from data frame 'outdata' onto the clipboard and the user can then paste it into Excel (or other programs).
There is also the Rcmdr package which provides a general GUI for R (basic tools) but also has a mechanism where you can create your own menus and dialog boxes for use with the GUI.
You didn't say it had to be a desktop program. So Jeroen Ooms' hilarious openCPU project might be worth a look. He basically calls it statistical computing in the cloud. The guys has been really active recently (now that I checked the website again, I realized it's new again).
Also, his earlier work stockplot or ggplot demo is very interesting. Especially a brief look at stockplot gives you an impression quickly of his approach.
Afaik, RApache is used and the nice frontend GUI is created with EXTJS .
I think the documentation can explain the approach much better than I do.
I think R on a webserver – particularly for intranet use is a good solution (depending on the size of your company), because:
it's platform independent, clients can use their favorite browsers
local development and deployment of script and even whole R pakacges is easy
Reporting / Publishing is big strength of R and can be facilitated using a web based architecture (see packages like knitr or sweave
In this question I just come around R shiny:
http://www.rstudio.com/shiny/
Related
I have an old DOS application which accepts some files as input, does some calculations and saves results into file system. This app uses terminal as sort of GUI, where you can choose input files, types of calculations to perform and choose where to save the result. I don't know the logics behind calculations and am not able to reuse them in a new project.
The problem is that the users of this app want a modern looking GUI which will be easier to work with.
That is why, I have an idea to create an adapter which will translate button clicks into commands in DOS and grab text output to show in modern GUI.
Is it possible and where should I start from?
It is possible. How to start from depends on your programming Tools. If you use a RAD tool like Delphi or Lazarus or Visual Besic or ... then make your GUI design first and define Events after. For a Button click it is ButtonXClick(); In the RAD tool you will find a object inspector with properties and ther values and Events and their values. Go to Events page there, look for onClick-event. Double click there in the value line and you will get an empty Event handler, wehre you can write your Code for your application.
If you dont have or use such RAD tool, take a GUI Framework for DOS. Create your frontend and write your Code which is to call in Dependance of your button clicks.
I'm working on a unique project using terminal/Shell but I've hit a little bit of a roadblock I haven't been able to work around.
I want to be able to read the content of the location of the cursor.
For example, if the cursor is currently located on line 2, column 5 which contains an E, I want to be able to read that E and create a variable with it.
Can you explain what your project entails? It might help if we knew what you're trying to accomplish.
No tools exist to do this in the shell, as far as I know. To actually read a remote screen would require this as a feature of the remote terminal (or emulator).
Neither do any compiled language support this. All applications that appear to do this fake it by keeping an internal copy of what they assume is displayed on the screen.
Lookup the curses* library for more information. This toolkit allows a programmer to address the screen as a random accessible grid, and hides all of the updates to the actual terminal screen.
See also: ncurses
I am looking for a solution to a specific use case:
When I read something on my browser or pdf reader, I want to take notes without switching windows. I want to type right on my browser or pdf reader, but the typed text should go to the background text editor like notepad.
Is this possible?
Do you know any existing automation script that handles this use case?
You should create an application that uses some keylogger-like techniques (e.g. global hooks) to monitor the keypresses and, depending on some condition (a setting you may have set, the currenctly active window, ...), it may pass them normally to the application or suppress them to store them in a buffer. Such buffer, then, would be shown to the user as needed.
Still, in my opinion a much more convenient way to perform a similar task would be to create an application consisting of a semitransparent edit box, that could be shown and hide simply with a hotkey. This would avoid all the hooks stuff and the potential problems that may arise from them.
you might try using autohotkey scripting language... I can write a little script that would do exactly what you need and afaik in this particular case it wouldnt need keyboard hooks.
--EDIT--
Autohotkey is VERY simple to use/learn so even if you want to do the script yourself you can do it in a very short time even if you dont have any knowledge of ahk. Then again, I can help you with it.
I wanted to have a GUI front-end for a script that accepts numerous command-line options, most of them are UNIX paths. So I thought rather than typing them in (even with auto-completion) every time, I'd create a GUI front end which contains text boxes with buttons beside them, which when clicked will invoke the file browser dialogue. Later, I thought I'd extend this to other scripts which would sure require a different set of GUI elements. This made me think if there's any existing app that would let me create a GUI dialog, after parsing some kind of description of the items that I want that window should contain.
I know of programs like Zenity, but I think it's doesn't give me what I want. For example, if I were to use it for the first script, it'll end up flashing sequence of windows in succession rather than getting everything done from a single window.
So, basically I'm looking at some corss-platform program that lets me create a window from a text description, probably XML or the like. Please suggest.
Thanks
Jeenu
Mozilla's XUL is a cross platform application framework - . You could write an app as a Firefox plugin or a standalone XUL application.
mono and monodevelop could work for this. Or even something super simple like shoes.
We've got several web-based applications that are launched from our ERP system (SAP R/3 in this case, but since we're using ShellExec, I don't consider this a SAP issue). The problem is that we can either set the IE to "open every URL passed from the ERP system in a new window" or to "reuse any one of the existing IE windows" (same problem with tabs). Both settings are not acceptable for our users: one of the web applications is a rather bulky medical image viewer applet that wreaks havoc if invoked multiple times. The other option is a no-go either because this way opening the image browser for a patient automatically displaces the lab result display for the very same patient and vice versa.
I'd like to have some control over which window may or may not be reused. My idea would be a kind of "window tag" and a helper program that checks if a window with that tag exists - if it does, reuse it, if not create it. So basically
IELauncher SEARCH http://www.google.com
would open a new window and tag it as SEARCH.
IELauncher DEVEL http://stackoverflow.com
would open a second window, tagging it as DEVEL. Then
IELauncher SEARCH http://www.wikipedia.org
would replace the contents of the first (google) window only.
Is this possible at all? Do you have any pointers for me where to start? I don't even know what to look for...
I had to resort to some rather ugly methods - dynamically assembling a HTML page that uses JavaScript and window.open() to open and access a window by name, then close the page executing the JavaScript. Ugly, but at least this way you get the additional benefit of being able to control certain aspects of the browser window like switching of the menu bar, toolbar and location bar.