Cancel Object Support

[fidoriel]

There is a way to do cancel object like octoprint does: https://github.com/mchrisgm/bambulabs_api/blob/ea7c190d1a6c730e541fc45a2128db5626a63012/bambulabs_api/mqtt_client.py#L379

What is needed:

• Figure out if it is possible to pull the object list and object placement from the printer
• If not pull the gcode to local cache and parse it to obtain object placement information
• Nice UI with placed objects and numbers/buttons to cancel a part.

[thunderbolt93]

greghesp/ha-bambulab#866

there's also skip-object support for the ha-bambulab homeassistant plugin in development

I created a proof of concept in with quart and svelte https://github.com/thunderbolt93/bbl-skip-object

the basic working principle for gettin the file is:

• wait for the printer to change from idle to printing state
• grab the filename via MQTT
• connect to the FTP (implicit TLS, same login data as MQTT)
• grab the modification date of the file via the MDTM command (prevent duplicate download if an existing file is printed again via on screen controls on the printer)
• download the 3mf file, extract Metadata/pick_1.png (and optionally Metadata/slice_info.config)

now we have the pick file available, which actually contains most of the data we need
the slice_info.config contains a list of object IDs and corresponding names,

the color data in the pick-image corresponds to the IDs, so we can have the user click an object, get the ID via the pixel color in the original pick.png (see handleImageClick function in App.svelte) and present a postprocessed version (pixels that correspond to an ID are colored depending on their presence in the list of skipped objects, see the updateSkip function in App.svelte)

the s_obj list in the printer status should also be factored in, as it contains already skipped objects

[fidoriel]

Are you interesting in porting it? If not, this project has a AGPL License, how should I handle it when getting inspired by your poc?

• What do you think about local caching for images and especially gcode. If files are uploaded they should be cached locally and only pulled via ftp if print started via orcaslicer. Also images and metadata should be cached locally.

[thunderbolt93]

feel free to use what you want from my proof of concept

regarding local caching: yeah, that's pretty much what I did, I just grabbed the filename from MQTT and the timestamp from FTP and saved the downloaded 3mf file with the filename including the timestamp, that way, if that version of a file already exists i can just use the local copy

the reason for the timestamp is that orca slicer will overwrite existing files on the SD-Card if you reupload the file (e.g. you changed some slicer settings or moved objects around and then reuploaded the file to the printer)

[fidoriel]

feel free to use what you want from my proof of concept

regarding local caching: yeah, that's pretty much what I did, I just grabbed the filename from MQTT and the timestamp from FTP and saved the downloaded 3mf file with the filename including the timestamp, that way, if that version of a file already exists i can just use the local copy

the reason for the timestamp is that orca slicer will overwrite existing files on the SD-Card if you reupload the file (e.g. you changed some slicer settings or moved objects around and then reuploaded the file to the printer)

Thank you. I will probably go with ULID as my unique factor in caching and will have to create a local cache for each printer.