Prusa link sd card emulation

Sarah_B · March 26, 2025, 3:44pm

hi,
I want to write a plugin to mimic sd card functions for prusa printers via prusaLink. The prusaLink part isn't hard.
I already have code to catch e.g. M20 (via queuing hook). What's missing for my first results is how to put my results in the received queue.

Any idea how to do that?

I saw this Handle M20 - List SD Card for printers don't support M20 but it misses some details for me, so I open a new thread...

thx in advance!

jneilliii · March 26, 2025, 10:02pm

For me and the Bambu plugin I'm able to do it through the serial factory I've added to the plugin to handle all serial communication back and forth. If you're not doing it all and just intercepting the M20 command you would have to call the internal comm layer received command I suspect. Something that uses the comm_instance parameter of the gcode queing callback, which is an instance of the MachineCom class. Seems like it might be difficult to do.

The other option, which I don't know if Prusa supports, is the use of M118 command, which basically tells the printer to send me back this line. So your hook captures queuing, grabs the file list, and then expands the command being sent to include a bunch of M118 commands with the file data lines appended.

Sarah_B · March 26, 2025, 10:28pm

M118 seams to work. nice idea! I want to replace the other not/not well supported SD commands.

I see I can control octoprint with that: "M118 A1 action:cancel" fancy
I looked already in the comm module and also find it hard to use for my purpose.

thx

jneilliii · March 27, 2025, 3:35am

Cool, so yeah that should work. In your gcode queueing hook, instead of returning line, you would return a list of M118 commands. Simple example (untested)...

def prusa_list(self, comm_instance, phase, cmd, cmd_type, gcode, subcode=None, tags=None *args, **kwargs):
    if gcode != "M20":
        return
    
    # code to get list of files and generate return data, ie dosname, date, size, unix timestamp, long filename
    file_list_commands = ["Begin file list"]

    for file in files:
        # dosname size timestamp "filename"
        file_list_commands .append(f"M118 {dosname } {size} {timestamp} \"{filename}\"") 

    file_list_commands.append("End file list")

    return file_list_commands

__plugin_hooks__ = {
    "octoprint.comm.protocol.gcode.queuing": prusa_list
}

you may also have to use the received phase gcode hook to manipulate how the printer sends the information (ie, remove "echo: " from the string possibly)

Sarah_B · March 27, 2025, 5:13am

Exactly what I did. Currently working on getting the M20 L to work for more then msdos filenames.

besides of the gcode commands I have the problem that file manager don't show my file list. I think I have to manipulate the printer capabilities.

jneilliii · March 27, 2025, 6:57pm

Yeah, I enabled long file name support and long file name writing, since I also am using the sd card upload hook.

github.com

jneilliii/OctoPrint-BambuPrinter/blob/d0fd4a5434043670c3fc02a2ce7020fc885be1d5/octoprint_bambu_printer/printer/bambu_virtual_printer.py#L442-L443


      
          self.sendIO("Cap:EXTENDED_M20:1")
          self.sendIO("Cap:LFN_WRITE:1")

jneilliii · March 27, 2025, 7:01pm

Oh more pointers on the dosname and timestamp. For timestamp I'm using OctoPrint's conversion from unix timestamp.

github.com

jneilliii/OctoPrint-BambuPrinter/blob/d0fd4a5434043670c3fc02a2ce7020fc885be1d5/octoprint_bambu_printer/printer/file_system/file_info.py#L7


      
          from __future__ import annotations

          

          from dataclasses import asdict, dataclass

          from datetime import datetime

          from pathlib import Path

          

          from octoprint.util.files import unix_timestamp_to_m20_timestamp

          

          

          @dataclass(frozen=True)

          class FileInfo:

              dosname: str

              path: Path

              size: int

              date: datetime

          

              @property

and for dosname I'm using OctoPrint's get_dos_filename

github.com

jneilliii/OctoPrint-BambuPrinter/blob/d0fd4a5434043670c3fc02a2ce7020fc885be1d5/octoprint_bambu_printer/printer/file_system/remote_sd_card_file_list.py#L8


      
          from __future__ import annotations
          
          import datetime
          from pathlib import Path
          from typing import Iterable, Iterator
          import logging.handlers
          
          from octoprint.util import get_dos_filename
          
          from .ftps_client import IoTFTPSClient, IoTFTPSConnection
          from .file_info import FileInfo
          
          
          class RemoteSDCardFileList:
          
              def __init__(self, settings) -> None:
                  self._settings = settings
                  self._selected_project_file: FileInfo | None = None

erffrfez · April 1, 2025, 4:05am

a meta-comment just to say "thanks". I'm not in this topic, but want to say in general that the world is a better place because of folks who help each other. I think you know that, and I don't think you are doing this just so that you get these "thanks", But it is important, so : kudos....
That is all

Sarah_B · April 1, 2025, 7:40am

Hi, @jneilliii thx for the answer. Main difference as I understand is that you have a complete printer plugin and I only a plugin that tries to modify the standard printer behaviour throug hooks and tricks.
This is my first plugin in octoprint and I'm feeling not ready for making a complete new printer class, albeit it might be better in the end.

My plugin shows files on sd card with long filenames and times (currently not the size of a file). I can load the file (with no analyze of the file) and delete a file. Also sd upload is possible.
I wasn't able to set capabilities, because i can't hook in the capability readout (with complete printer plugin as yours this is easy).

Currently I work on starting the print from sd card.
Any clue how I can hook in the analyze stuff so I can provide more informations of the file?

jneilliii · April 1, 2025, 12:30pm

Unfortunately analysis doesn't work on SD card files, I have the same problem with full virtual printer.

Sarah_B · April 1, 2025, 12:44pm

my plan is to have a custom GcodeAnalysisQueue that download the file and analyse it.

I look in plugin "prusa mini eta" there is a good example.

foosel · April 2, 2025, 7:52am

For the record, when the 1.11.0 release candidate phase is not keeping me busy, I'm currently working on finally getting rid of the tech debt that is having the SD card stuff not be its own storage, and implementing that properly. Which also means capability flags on the storage interface whether files can also be read vs just written, which in turn should enable gcode analysis and the gcode viewer if talking to a printer that actually has a proper file management interface vs just a write-only storage (which basically is what the standard SD card related GCODEs boil down to).

That will be in 1.12.0, and my currently hope is that I'll be able to push this out faster than the past minor releases. But that all depends on the other stuff that comes up.

Sarah_B · April 2, 2025, 8:09am

That sounds great! So I'm concentrating on proper support for prusa network stuff and hope for your changes later.

Btw I can start a print on "sd card". Now the cancel function...

TheDelta · May 14, 2025, 3:40pm

I was fighting some OctoPrint serial print issues on my MK4 without any success. Whenever I print via OctoPrint it does weird things, but works totally fine on USB Stick and currently using PrusaLink and was thinking the whole time about "why not just connect it via OctoPrint".

Looking forward to this, would help me tremendously <3

Would love to contribute / test whenever I have some free time.
Any chance to have this on a public repo already?

Sarah_B · May 15, 2025, 5:46am

Next week I can publish the code - but this is more like a concept rather than a daily driver. Anyway I use it daily and it has it's limitations currently, but this is great to check why this plugin is a useful idea.

Currently I wait for octo 1.12 because I need changes in SD card stuff so I can simulate this without so much quirks.

Perhaps you know the "prusa did not support serial api aka octo so good as in mk3 ages" discussion (e.g. "feature request" show information's of prints from OctoPrint · Issue #189 · prusa3d/Prusa-Firmware-Buddy · GitHub )

I currently used prusalink due to speed issues and I miss informations that I currently have when I print via octo ("feature request" show information's of prints from OctoPrint · Issue #189 · prusa3d/Prusa-Firmware-Buddy · GitHub - this is not my current text, I addd more infos).

using my plugin gives the potential to join both worlds. current state can give a glimpse of what it can be.

so I'm happy to share and happy for helping hands.

TheDelta · May 15, 2025, 7:07am

Oh yeah, having a working screen is something I totally forgot and also appreciate after switching back to PrusaLink.

Sounds great, can't wait to have a look.
You said it well, both worlds joined together would be totally awesome.

Sarah_B · May 21, 2025, 6:38am

this is the repo: GitHub - redPanther/prusalink4octoprint

in init.py you have to set your credentials in function:

def get_settings_defaults(self):
    return dict(
        host="your host ip",
        username="your user name",
        password="your password"
    )

you should see your files on printers usb stick and you can start the print - as long the printer accepts the gcode without warning (nozzle size/filament type)

when you upload a file to octo it is uploaded to the printers memory too.

will give some more infos in some days ...

and remember, this code is just a hacky proof of concept!

the prusa buddy github has lots of outstanding features regarding serial support+sd card. From time to (long) time they add missing gcode commands. so I hope for movement on prusa side - this would make this plugin much easier. Perhaps you should upvote&comment those issues, if you havn't already

jneilliii · May 21, 2025, 4:57pm

you don't have to set this in init.py, you can just add it to config.yaml either via octoprint CLI commands or manually with nano. I'm not sure how the plugin_identifier is extracted when not defined, but typically that is handled in the setup.py that seems to be missing from your repo.

plugins:
  <plugin_identifier>:
        host="your host ip",
        username="your user name",
        password="your password"

I see you started a settings jinja template that has settings.plugins.prusalink.url, which doesn't match up to any of your default settings and you're probably getting binding errors on initial page load, but you can add those same blocks for these settings if you want, although you should probably encode the password in some way so it's not stored in plain text. Same as above, you need to replace <plugin_identiier>.

    <div class="control-group">
        <label class="control-label">{{ _('URL') }}</label>
        <div class="controls">
            <input type="text" class="input-block-level" data-bind="value: settings.plugins.<plugin_identifier>.host">
        </div>
    </div>
    <div class="control-group">
        <label class="control-label">{{ _('URL') }}</label>
        <div class="controls">
            <input type="text" class="input-block-level" data-bind="value: settings.plugins.<plugin_identifier>.username">
        </div>
    </div>
    <div class="control-group">
        <label class="control-label">{{ _('URL') }}</label>
        <div class="controls">
            <input type="text" class="input-block-level" data-bind="value: settings.plugins.<plugin_identifier>.password">
        </div>
    </div>

Sarah_B · May 21, 2025, 5:39pm

I know, this was just a quick get it started thingy. I published it, because there was a request.

Plan is to start from scratch and make it as a printer plug in.