Hello, I have been searching for protocols to communicate matlab and puredata for transfering information using matlab as a client and puredata as a server. I created a matlab code that sends a message to puredata but at the puredata end, I can stablish the connection (I debugged it using a number box) but I am not able to print matlab's message on the editor.
I saw these threads but it didn't work either:
https://forum.pdpatchrepo.info/topic/11188/eof-on-socket-using-netreceive-raspberry-pi
https://stackoverflow.com/questions/17077364/communicating-between-pure-data-and-matlab-using-osc

Here is my matlab code:
clear all;
clc;
tcpipClient = tcpip('127.0.0.1',55002);
set(tcpipClient,'Timeout',30);
fopen(tcpipClient);
fwrite(tcpipClient,'hi, i am matlab');
pause(10);
fclose(tcpipClient);

@greencv I don't know if it is necessary, but in Matlab doc there is a space before the port number.
But Pd is receiving something or there would be just the broadcast setting.
I assume that you do not get the message if your Matlab code is not running.

I think you need to tell Pd that the message has ended...... so.... maybe.... add a semi-colon to your message so that Pd knows it has reached the end of file...?

clear all;
 clc;
 tcpipClient = tcpip('127.0.0.1', 55002);
 set(tcpipClient,'Timeout',30);
 fopen(tcpipClient);
 fwrite(tcpipClient,'hi, i am matlab ;');
 pause(10);
 fclose(tcpipClient);

David.

@whale-av
Hello David,

Yes, you are right! I didn't notice that Pd wasn't able to tell when the message was ending... that appeared to be the problem because now it is working perfectly when I send strings from matlab to pd.

However when I try to send a number from matlab and display it in a number box in pd, it doesn't seem to do anything, I will keep on reading about data type compatibility between the two softwares, perhaps it is related to that...

Thank you really much for the help with the string message!

@greencv Try sending ('float 0.998 ;')
"float" or "list" should work.....
You need to tell Pd what sort of atom it is receiving. Internally in Pd the atom definitions are always there, just not always
shown to us "users"........
For more info on messages in Pd....... http://puredata.info/docs/manuals/pd/x2.htm .... chapter 2.3.
Or actually...... this is better...... https://puredata.info/dev/PdMessages although the first link is good bedtime reading...
David.

@whale-av
I see, then I will see the file formats and try what you suggest.
Awesome, I will reveiw both links, thank you really much for all the help!

Carlos

Updates:

I was now able to figure out how to send real time float numbers from matlab to puredata.

Because puredata needs to be told which kind of atom it is receiving, (in this case float), In matlab you need to use the functon "sprintf" to send a variable with a stored number inside a string and still be able to tell puredata that the variable is a float.

Also, pd needs to know when the message is over, so in matlab you need to put a semicolon at the end of the string.. Here is the code that worked:
In matlab:
clc;
tcpipClient = tcpip('127.0.0.1',3000);
set(tcpipClient,'Timeout',30);
fopen(tcpipClient);
a=rand;
% b=num2str(a)+';';
fwrite(tcpipClient,sprintf('float %f;', a));
fclose(tcpipClient);

This sends a random number from matlab to pd

@greencv You can probably forget about "pointer", but if you figure out how to send the types "bang" "symbol" and "list then please post them back here.......
"symbol" might be the type you started with though. You can check that with [list split 1] which will split off "one" from the message "one two three" if it is a list, but not if "one two three" is a symbol.....

Probably, for most people, "list" will be the most important, as they will be trying to send data sets like "0.998 0.5 -2" for which Pd need to see "list 0.988 0.5 -2"..........
A set of atoms starting with a float (first atom) is automatically a list in Pd.... e.g. 0.22 hello matlab.... is a list, but hello matlab 0.22 is not.
David.

@whale-av
Of course, I'll see what I get when trying to send arrays or more than one number at a time and try to see if I can split the results in pd...

I will post how it goes, in here