2 sheets tagged with "update":

Open the Resolver document. See how it contains usercode to create a new (cache) worksheet to recieve the incoming data. Switch to this sheet, if it isn't already the current.

More usercode uses the code in server.py to create a TCP server, listening for messages on localhost port 9999. You should see some diagnostic text in the output pane reading:

listening on localhost:9999... ok

(you might have to scroll up to see it)

Now, open a command-line window and cd to the directory of these files. Type the command:

python send-test-data.py < some-json.txt

This will send the JSON in the text file as a TCP message to localhost on port 9999.

The JSON message is recieved by the TCP server, which passes the message to the function 'handler' in the Resolver document usercode.

This function does a quick and dirty conversion of JSON into Python data, and uses that data to populate some cells in the cache worksheet.

If you click on the grid (ie. not on the code pane) then recalcs will
automatically update the sheet with new values every 2 seconds.

In this example, Resolver acts as a TCP client, sending requests for data on every recalc. Recalcs have been set to automatically re-occur every second (note, this auto-recalc stops if you suspend recalcs using the toolbar button,
or if you are editing the code in the code pane.)

Unzip the attached to provide the following three files:

json-over-tcp.rsl - a Resolver document
test-server.py - a command line script to act as test TCP server
README.txt - this file

Open the .rsl document.

Outgoing TCP messages are produced on every recalc, by the function
'request()', in the pre-constants usercode.

These will currently give the following error:

couldnt connect to server
[Errno 10061] No connection could be made because the target machine actively refused it 127.0.0.1:9999

This is coming from the print in the request() function, The error is because nothing is listening to the port on which Resolver is asking for data.

Open a command-line window, and cd to the directory containing these files.

Enter the command:

python test-server.py

It should start a TCP server, and produce the message:

listening on localhost:9999

Now, whenever a Resolver recalc occurs, the function 'request' sends a message, consisting merely of the word 'get', sent to localhost 9999.

The responses, sent back by test-server.py, contain some JSON data, containing a list of (col, row, value) triples.

The JSON response is received by Resolver in 'request()', and printed to the output pane:

-------- Recalculation Started: 16:49:50 --------
response [
[4, 4, "marigold"],
[3, 4, "pot"],
]

This JSON is converted into Python data structures in the pre-formula user code, using the dirty hack:

data = eval(json) # this works because json is valid Python

The decoded data is then used by function 'display()' to display words and colors on the worksheet.

If you click on the grid (ie. not on the code pane) then recalcs will
automatically update the sheet with new values every second.

Alternatively, requests like this could be put into button handlers instead of on every recalc. This would be arguably easier, since button handlers are magically able to persistantly update cell values, using the Formula property,
so there would be no need to use a cache worksheet any more. But button handlers can't be set to trigger every X seconds like recalcs can.