Skip to content

Latest commit

 

History

History
79 lines (57 loc) · 2.65 KB

README.md

File metadata and controls

79 lines (57 loc) · 2.65 KB
Raw

hyperpixel

Preview of system

Text dominates the software experience. Visual aesthetics are so difficult to program in the web, that commodity blogging templates have come to dominate. Everything looks the same.

Hyperpixel is an image-centric authoring specification.

It is built on two core primitive: images and transcluded links.

  • Each image is hosted at a URL
  • Sub-regions of an image can link to other URLs
  • Backlinks are present

You navigate through the graph of images by clicking on links, much like traditional hypertext. What makes this different is the authoring experience: instead of opening up a text-editor, pull out a pen and paper, open up photoshop, or take some screenshots.

authoring

Hyperpixel is built with a very simple local server pattern. You run the server on your local computer, and ask it to point to some directory of files on your filesystem. It more-or-less follows this vanilla js micro-framework.

> python3 server.py <directory>

The system has two kinds of pages:

  1. Inbox at localhost:1234/, corresponding to inbox.html
  2. Image pages at localhost:1234/<image-id>, corresponding to edit.html

To add a new image into the system, you can navigate to an localhost:1234/<url> and either use the file picker or drag-and-drop onto page.

Once on an image page with an image, you can create a link by clicking and dragging and then typing a URL in the white address bar. Local URLs to other images are prefixed with /, but you can also just link to an existing URL on the world wide web.

You can delete links with a shift-click.

The system is synchronized with a single <directory> of your filesystem. It will listen for new files in <directory> as well as persist uploaded images to <directory>. The metadata with associated OCR, links, and backlinks is saved to <directory>/db.json.

Existing images show up in the inbox. If an image is already within the link network, it'll show up in the left column. Unlinked images show up in the right.

You can search for images from the inbox. This relies on OCR detection.

requirements

  • pip: tornado,pillow
  • brew/apt: tesseract-ocr

publishing

cd publish
python3 publish.py <directory>
cd build
python3 -m http.server
open localhost:8000

The creation of the static build is done with: python3 publish.py <directory>. For publishing, index.template and read.template are roughly equivalent to inbox and edit, with editing functionality stripped out. This allows for a read-only static build which can be uploaded to a personal server.