ImJoy - 用于部署深度学习应用的混合计算平台，支持插件

• Support plugin-level permission control:
  • camera
  • midi
  • geolocation
  • microphone
  • encrypted-media
  • full-screen
  • payment-request
  • same-origin
• provide a standalone library imjoy-lib.js, which can be used in third-party applications
• https://imjoy.io/lite is the lite version run the ImJoy app in standalone mode.
• Add the plugin repository/gallery page at https://imjoy.io/repo

Support uploading files to the plugin engine

• drag and drop files/folders to the plugin engine
• new api functions: api.requestUploadUrl, api.uploadFileToUrl and api.downloadFileFromUrl
• remove file and folder in the plugin engine
• goto any folder
• download file from the engine file dialog

Enhanced window management

• shortcut icons to up to 9 windows
• click to focus window
• standalone mode for windows

Improved Engine Dialog

• seperate to new engine dialog and about engine dialog
• add api_version to engine
• show system information for plugin engines
• Reset engine button

More pythonic way of using imjoy api

• support from imjoy import api

Create window plugin in web-python

• support new plugin type web-python-window
• allows access to dom in python
• plot directly with matplotlib

Support cover image(s)/GIFs

• add cover filed to <config> block which allows to add animations to plugins
• demo: https://tinyurl.com/yxdtse8l

Other improvements

• Use tight layout on small screen
• add api.prompt
• show Restart ImJoy button to allow restart manually in PWA
• popup restart button when loading timeout

API version: 0.1.5

Breaking changes

• window.onclose has been renamed to window.onClose
• api.showFileDialog now return an object instead of a string
• if user cancelled api.showFileDialog, it will return false instead of trigger an error

• Support Web Python, run python plugins without the plugin engine through WebAssembly
• Single file plugin file format (.imjoy.html)
• Plugin editor for editing, test and run your plugin code in the browser
• Installation tags for configurable plugin behavior
• Workflows can be saved and loaded
• Support plugin deployment with Github/Gist, support url parameters for for the distribution of plugins
• Support Python plugins via the Plugin Engine
• Use conda to manage the virtual environments and packages for Python Plugins
• Native support transferring n-dimensional arrays, conversion between numpy/numjs/tensorflow
• Support background tasks and reattach the interface to the Python plugin process
• Share plugin with plugin URI
• Use JSON-schema to recognize inputs and outputs
• Support mobile device
• Plugin engine support (multiple): https://github.com/oeway/ImJoy-Engine
• Support desktop app with https://github.com/oeway/ImJoy-App
• Preliminary offline support with service workers

• Support Web Python, run python plugins without the plugin engine through WebAssembly
• Single file plugin file format (.imjoy.html)
• Plugin editor for editing, test and run your plugin code in the browser
• Installation tags for configurable plugin behavior
• Workflows can be saved and loaded
• Support plugin deployment with Github/Gist, support url parameters for for the distribution of plugins
• Support Python plugins via the Plugin Engine
• Use conda to manage the virtual environments and packages for Python Plugins
• Native support transferring n-dimensional arrays, conversion between numpy/numjs/tensorflow
• Support background tasks and reattach the interface to the Python plugin process
• Share plugin with plugin URI
• Use JSON-schema to recognize inputs and outputs
• Support mobile device
• Plugin engine support: https://github.com/oeway/ImJoy-Engine
• Support desktop app with https://github.com/oeway/ImJoy-App

• Support Web Python, run python plugins without the plugin engine through WebAssembly
• Single file plugin file format (.imjoy.html)
• Plugin editor for editing, test and run your plugin code in the browser
• Installation tags for configurable plugin behavior
• Workflows can be saved and loaded
• Support plugin deployment with Github/Gist, support url parameters for for the distribution of plugins
• Support Python plugins via the Plugin Engine
• Use conda to manage the virtual environments and packages for Python Plugins
• Native support transferring n-dimensional arrays, conversion between numpy/numjs/tensorflow
• Support background tasks and reattach the interface to the Python plugin process
• Share plugin with plugin URI
• Use JSON-schema to recognize inputs and outputs
• Support mobile device
• Plugin engine support: https://github.com/oeway/ImJoy-Engine
• Support desktop app with https://github.com/oeway/ImJoy-App