SynapseDirect

My employer, US HealthWorks uses an IBM AS/400  network for PMS.  The company has about 114 clinics, so it's a sizable system.  Recently, management decided that the "Activity Prescription Form" used for all work comp cases cannot be faxed or sent hand written from the clinics (as they heretofore have always been).  Docs are being strongly encouraged to fill out the form on AS/400 -- much more efficient.   The clinics run AS/400 emulator clients on Windows PC's.  Very DOS/console looking stuff.

I figured out to use xtn5250, an open source java AS/400 client on Puppy Linux along with Synapse.  For a long time I have pondered the possibility of "screen scaping" patient demographics from the AS/400 client screen and dumping them somehow into Synapse.  I want to keep other notes and fax in prescriptions for patients using Synapse as well as generate required forms.

Today I figured out how to drag and paste patient demographics out of AS/400.  Here's a sample -- I just changed the real name, address, and phone:

  BAAD, BOBBY S                         
  2218 E BAD RD                       
  MEAD                   St WA  Zip 99208
  509  9999999           Cell 000 0000000

Any specific ideas as to how to import this into Synapse?
 

If it's an open source client perhaps you could alter this to dump the demographics into a CSV file ... using the Synapse CSV import format as previously documented.

Graham:

If it's an open source client perhaps you could alter this to dump the demographics into a CSV file ... using the Synapse CSV import format as previously documented.

Or use AutoHotKey to "pull" the data right from the screen and either (1) place it directly into Synapse or (2) generate a .csv file (to import as you suggest).

I think this would be an easy script to make.

If, all the information was on "one screen", and you could TAB to each field, that would make it easier. Essentially the script would systematically Tab across the screen, collecting the data as it went to the needed fields and saving them as needed.

I could certainly generate the data going the other way.

A good AHK scripter (not me really) could do this in 5 minutes.

Rebol isn't 100% AHK friendly, but this script is easily doable.

Another method would be to use something like SnagIt - to "Screen OCR" the Text on the screen.

Or there is another program called "ScreenOCR" 

Jason, 

Fortunately, I don't need to OCR the text.  I can use the "hold left mouse" button in conjunction with a rectangle to define an area in the AS/400 console window to drag, and paste the results from the clipboard directly in a word processor.  That's what I did the example demographic.  Worked fine on Puppy Linux.  It works pretty much exactly like dragging and pasting text from a Command Console window in Windows.

I like your idea of using a "SendKeys" type of script to advance the cursor to each datafield I want it to go to, harvest the data and do something with it.  The IBM AS/400 Client on Windows can use VBScript or it's own internal script to run macros, which would do exactly what you are suggesting.  IBM has a free version of their AS/400 Clients available for Linux, which I wasn't able to get working yet.  VBScript wouldn't work on Linux, but I bet the internal macro language would work.

I wonder, though as to the relative advantage of harvesting the data one field at a time vs using a parser to reformat the data into a CSV file.  I know how to use Tcl/Tk to reformat the data.  Would I be able to use REBOL to do that, and would there be any relative advantage?  Maybe better able to hook up the script to Synapse?

Grabbing data by one field at a time sounds no better than just retyping the data.

I would modify the OS java client to grab the data and save it as a CSV file.

Graham:

Grabbing data by one field at a time sounds no better than just retyping the data. 

Except that it would scripted and the computer can "Tab, Copy, Save to file,  ..... Tab, Copy, Save to file .....    repeat until done " 1000x faster than a human :)