Feature: add prompt load from layer

[ElliottLester]

Scope: We want to restore the state of SDImgPageBase from an already generated layer.

Requests for comment:

• Is this a reasonable design
• Is it worth making an "info" class for holding generation information but no the base64 encoded images. -- this allows to/from json -- central list of properties
• I don't like the way update_prompt_params works at the moment.
• Should we update the model I don't because the user can't see it from the prompt screen. (expand info bar maybe?)

[JasonS09]

• Is this a reasonable design
• Is it worth making an "info" class for holding generation information but no the base64 encoded images. -- this allows to/from json -- central list of properties

I think we can go with the class approach. May be a simpler and less confusing design than having to use the PROMPT_LOADER_KEYS list.

• Should we update the model I don't because the user can't see it from the prompt screen. (expand info bar maybe?)

One possible way to update the model could be adding this in update_prompt_params function:

dockers = Krita.instance().dockers()
for d in dockers:
    if d.objectName() == TAB_SDCOMMON:
    d.page_widget.sd_model_layout.qcombo.setEditText(params[keys[0]])
    break

[ElliottLester]

@JasonS09 So I wrote a replacement for craft_response() that formalizes the response class and pushes some code out of client.py Let me know if you think this is the right direction? There are some TODOs in this PR I would like your thoughts on. client.py:224 prompt.py

Other Querys?

• Are we able to import PyQT directly on other platforms? this makes writing unit tests easier.
• These classes can be printed and serialized to json
• The addition of Base64Image class lets us move the QImage and b64string around together this should stop us from changing formats everywhere.
• I would like to move the workflow generation code out of client.py I'm thinking Prompt.to_workflow()

[JasonS09]

• I would like to move the workflow generation code out of client.py I'm thinking Prompt.to_workflow()

Go for it.

[ElliottLester]

I have removed all the base64 conversions in the response section. I think I'll check this in before moving onto the Prompt/config side.