Hi,

I am trying to control a robot using HTTP requests. I would like to use [netsend] to send messages from Pd.

Trigger a recorded motion using HTTP requests

This can be done by making an HTTP request to the ikaros web server.

http://127.0.0.1:8000/command/SR.trig/X/0/0

X is the sequence. 0 Is the first sequence.

The local IP number 127.0.0.1 can be replaced by an assigned network IP.

If I write in the web browser address bar it works fine but I have not managed to get [netsend] working. I do notice that by sending [connect 127.0.0.1 8000( it seems connected but I think I am sending messages that are not formatted correctly. How should they look in pd?

Thanks!

@cfry 127.0.0.1 is your localhost..... the computer running Pd.
If the robot is running the web server "ikaros" then I imagine that you need the IP address of the robot (you should fix it... make it static and outside the dhcp range of your router so that there are no conflicts.).
David.

@cfry said:

Thanks for replying.

if I want to send this command through [netsend]

http://127.0.0.1:8000/command/SR.trig/1/0/0

How should it look in Pd/[netsend]?

I can open the connection with [connect localhost 8000(

but sending

[list command/SR.trig/1/0/0(
or
[list command SR.trig 1 0 0(

does not work. I am formatting the ip adress wrong I think.

[netsend] doesn't use HTTP but FUDI protocol.

You can do
but it opens the link in the browser

@FFW thanks. That works sort of but since the browser pops up and create a new tab each time it is not functional in this case. But good to know.

Is there a way to send HTTP request from Pd in some manner without going throw the browser, even if [netrecieve] is not suitable? I would have guessed that it was just a formatting problem.

@cfry said:

I would have guessed that it was just a formatting problem.

It is.
You can try the simplest http 0.9 request:

GET /command/SR.trig/1/0/0

but FUDI add a semicolon at the message end so it can fail…

I think the PuREST_JSON Library could be a good try. There is the [rest] object which can send GET requests! I don't know much about the details there but it can definetly contact a server and get something back without opening a browser!

@rg.kies thanks, Ill explore this tomorrow when I get my hands on the robot again. However I suspect this seems more complicated than it is since I am not sure of the procedure. Below is the github with instructions. Probably better to read that to understand what I am trying to do. I think what I want to achieve in the end is a socket connection from Pd to the robot (web) control interface.
Look here plz: https://github.com/birgerjohansson/Epi/wiki/Additional-instructions

Since I probably have to do this on another computer, the less libraries I have to install the better. Its a hassle with secure Internet connections and so forth.

Thank you so much.

@cfry The Mrpeach library, or the IEM library update has some basic HTTP objects too.
But I am not sure that it is the HTTP protocol that you want to use.
Looking at the first sentence on that page I think you could use midi..... https://github.com/search?q=repo%3Aikaros-project%2Fikaros+midi&type=commits
or better still a socket using [netsend] as you first suggested.... https://github.com/search?q=repo%3Aikaros-project%2Fikaros+socket&type=commits
maybe specifically "Adding basic modules for socket communication"

The project has been running for a long time.... so there is a lot to read...!
Can a socket module use the same message strings as the web interface?
Probably.
David.

@cfry You could use [command] to run curl or wget.

@rg.kies said:

I think the PuREST_JSON Library could be a good try. There is the [rest] object which can send GET requests! I don't know much about the details there but it can definetly contact a server and get something back without opening a browser!

This is a wonderful project, but I always have issues with this lib under MacOS... There is a version in deken but It seems outdated. The project moved to https://codeberg.org/Residuum/PuRestJson

@cfry Actually, looking at MrPeach again..... it is not at all basic, and my experience in the past was that the library worked well and was easy to use. It should be on Deken.
Have a look at [tcpclient]..... the help file has links to string conversions too, although they have mostly been built into Vanilla since.
I haven't checked, but the fudi objects might well do the correct decimal conversions for your messages.

But it looks as though you could send [connect xxx.xxx.xxx.xxx 8000( ... into [tcpclient]
and then [send /command/SR.trig/X/0/0(
although maybe the leading forward slash can be dropped for the send message?

If you will be using an internet connection then don't forget to give Pd public network permission in your firewall.
David.

@whale-av thank you so much. I have contact with the devs but I try to wait asking for favours until I have a decent understanding... as I do not want to strain you guys too much either... its a balancing act.

Unfortunately mrpeach is not available through deken, says "no matching packages for your architecture". Im on a Mac M2 with 12.5.1 Monterey that comes with the robot.

Status at the moment:

I can trigger from the browser and from terminal/zsh with curl.

I have not managed to get [netsend] to work. Tried some combinations with list and symbol. If using tcp I lock out terminal and browser triggering on that port/socket(?) (8000) . Using [netsend -u] (UPD) allows me to continue using terminal/zsh or the browser for triggering. My conclusion is that it would work if could format the message in the right way. But I do not know how the message looks when/if it arrives at the Ikaros WebUI. Can I monitor this in some way?

Now I will try using [command] to see if that works.

EDIT:
This works!! thanks @oid for getting me into [command]!

Now the next step will be to try to control parameters.

My goal is to be able to change the slider parameters of the robot/web ui from Pd by some sound process on the same computer locally. So no plans for connecting over internet at the moment.

It works fine now, thank you for the support!

Quick proof of concept, triggering individual channels with sound:
https://www.instagram.com/p/C2z_v_7iZZt/

I get constant error messages in Pd console from command saying "old process is still running". Probably [console] should be handled a bit different, but it works anyway so all good. Probably [netsend] streaming control data over a socket-channel would be the best solution if it needs to be rock solid.

Cheers & thanks

@cfry You get the error because pd is faster than the shell. A few options, none tested but even if they do not work they show the methodology. First just uses two commands, still may throw errors but it is simple and could be fast enough that there are no errors. Second implements a queue using a [text], each request is written to the [text] and when the request is completed [command] gets the next request from the queue. Third just gets the most recent request which is what you have now since all those errors are ignored requests, each request gets stored in the [list] and when [command] finishes it bangs the [list] to get the last request stored. Third is probably the most sane solution but each have advantages.

http-requests.pd

Edit: Redid the post, it was a mess

Edit: Flaw in the [text] version, fixed? Shouldn't respond when I am procrastinating, always turns out a mess but it is a good way to procrastinate.

@oid Thank you, I used the third version (which outputs only the last request). Seems to work fine.

A thought/question: First I tried to have all messages to all motors of the robot to be sent through a single pd [command] object. This did not work that well since messages was blocked/cancelled because [command] was not keeping up. When I used a separate instance of [command] for each channel I was sending on it worked better. This leads me to think that you could send through command at a higher speed if you had several [command] objects that you cycle through. I have no need for this here but I just want to understand how [command] works.

Another thought- when you do these nice examples/abstractions maybe you could add some comments in the patch too? I can figure out what is happening but I think it would be very beneficial for other ppl browsing and maybe make it more likely that others can take part of what we come up with. If it is not too much extra work for you, hit command+5 and throw in a few words?

Cheers! Good job with the procrastination!

@cfry said:

This leads me to think that you could send through command at a higher speed if you had several [command] objects that you cycle through.

Sure, that is what the first version does, toggle works the same as a counter followed by a [mod 2]. The problem here is that you probably can not be certain the [command]s will finish in order, first one might take 2 seconds and the other 3 could finish in 0.2 seconds causing it to end where it should have started. Shells are not very rigid in their timing and we don't know how the kernel is going to distribute those commands across cores and even on a single core the kernel may decide that something else is more important and cause problems for you.

I really am terrible at commenting, both the act of adding them and writing effective and useful comments.. It is a frustration for me as well, I spend a fair amount of time figuring out old patches. My neat and consistent patching style is partially an attempt to make up for my inability to comment patches which is not real helpful on these smaller patches though. Mainly I hope people will ask questions if they have them.