atom-54 / xenonjs Goto Github PK

Firebase, Fission, spreadsheet, etc...

add support for image editing models:
https://huggingface.co/models?pipeline_tag=image-to-image

non-web graph runner

• Values in text-input with options
• Image upload support
• Fix numeric range element
• Connecting Select field output to a Number input (e.g. radius in Points of interest demo)
• Error handling: typing NaN value in a Number field causes an error
• Open Style inspector: add word-wrap/white-space controls

improve publishing, sharing, and TBD

• Verify Listen works
• Knowledge space:
  • tags don't fit the screen
  • last column shows empty
  • remove checkboxes from table
• Points of interest:
  • Location field doesn't fit on the screen (not showing)
  • SplitPanel - doesn't work
• Improve text wrapping:
  • Image search / VectorDB demos: input text (at the top) is clipped
  • Image completion: the input's label is clipped

explore possible options:

• connect inputs of a newly added node to existing outputs
• connect outputs of newly added node to existing inputs
• connect only if type and name match? for fields (e.g. TextField) also consider label?
• add additional hints / tags in node types?

upload and image, scrape an area and perform image complete

menu, carousel/gallery, accordion..

• Select a Graph with deviceUx, turn on the camera
• Choose another graph
  The camera remains on, but it should be turned off

• text based
• w/video streaming
• multi user

uses locally stored selected ID, and that defaults to local$

to reproduce:

• open a public graph
• click + and add a node
• the node appears in the Designer, but not in Inspector or Nodegraph, which is confusing and inconsistent