PLANNING ISSUE

[tech4GT]

discussed on 29th April 2018

Original proposal: https://publiclab.org/notes/tech4gt/02-19-2018/gsoc-proposal-image-sequencer-v2-processing-on-steroids

New projects:

• [x] meta modules - Meta Modules (node): https://github.com/publiclab/image-sequencer/pull/203
• [x] meta modules in browser (should be possible to do both at once?)

Planning issues (see https://publiclab.org/notes/warren/01-18-2018/software-outreach-modularity-is-great-for-collaboration)

Priority/order of tasks (timeline)

Core

• [x] overall simplifications of new module submission
  • [ ] automate output assignment (via output() function, which can be overridden)
  • [x] automate UI setup
• [x] Dynamic modules
• [x] sequencer.importModule(module)
• [x] sequencer --importModule module-name
• [x] Meta modules - sequencer.importModule(string)
• [x] sequencer
  • in browser and node
  • documentation
• [x] conversion png/jpg/gif with sequencer.run(format)
  • integration into CLI and Browser UI
• [ ] (generally) consider CodeClimate

Demo

• [x] generating "permalinks" to a saved sequence (using toString/toJSON methods + URL generation)
• [ ] compartmentalize demo code into a re-usable module, like sequencer.htmlUI.addStep(name) or something?
• [x] test demo code
• [x] load-image module improvements - i.e. from webcam
• [x] page caching for offline use

CLI Improvements

• [x] naming of the output image

Use cases, modules

• [x] particle sizing
  • [x] region extraction
• [x] Text Overlay
• [x] Dithering
• [x] Flood Fill
• [x] contrast (can build on dynamic module?)

Rest server

• determine URL structure like /run?s=SEQUENCE_HERE&i=IMAGE or u=URL <= although probably if we pass the image itself it'd be POST not GET
• pass sequence to CLI and return new image
• callback for and client-side integration
• documentation & test server?

https://github.com/publiclab/image-sequencer/pulls

Code quality

• Code Climate
• Bithound
• Codefactor
• Codebeat
• Codecov

New Modules

• Text Overlay
• Histogram
• Dithering
• Flood Fill
• Combine different channels of different images
• Encrypt using AES or any other suitable algorithm
• FaceDetection

[jywarren]

Let's update this page too! 👍

[tech4GT]

@jywarren Sure!!

[harshkhandeparkar]

Flood fill, dither done. Text overlay has its own issue. Particle sizing and region extraction have an issue as well. Demo code testing done. Webcam done. UI code refactoring in progress in #825