Allow all user input to be simulated via the Local Http Server API

It would be really cool if every possible user input could be performed via the Local Http Server API as if the user were doing them directly via mouse and keyboard on the Logseq app.

Having a centralized dispatch endpoint for every possible command that usually would be performed by triggering keybindings, using the command palette, clicking a context menu item, using a slash command, etc. would make it a lot easier to “blackbox-automate” logseq from a user perspective. But instead of telling the API to perform low-level input events like “click X” or “type Y”, it could be semantic, like “Go to tomorrow”, “Move block up”, etc.