Skip to content

Spacelog/kallisto

Repository files navigation

Kallisto, crowd-sourced transcript cleaning

Kallisto is a simple web app that allows many people to contribute to cleaning a transcript, most likely for Spacelog (or another site using the same Artemis software). Input is in the form of a set of PNGs of transcript pages, and (optionally) a text file for each to act as an initial "uncleaned" version. In turn, someone using Kallisto will "clean" a single page, attempting to get the text representation to match the original image.

Notes

This is very early days, but:

  • Ubuntu (14.04) + postgresql-9.3 + imagemagick + ghostscript + libpng12-0 + python-dev
  • Create a database called "kallisto" accessible by "user" with no password
  • Django 1.7 app; pip install -r requirements/dev.txt into a virtualenv, then python manage.py migrate should get you started
  • You need to make a mission, which must include a mission patch (because we want things to look pretty!)
  • python manage.py import_pages will import the text and PNG pages; see its help for how (and how to generate them)
  • run tests with python manage.py test

Deploying a live instance

Via fabric. Check out fab -l for help, then:

$ fab -H $HOST setup:{ SCARY AMOUNTS OF CONFIGURATION }
$ fab -H $HOST deploy

should do what you want. Won't help you with setting up a database, or directing your website to forward to the gunicorn instance.

Screen sizes

This works well at 1080p full screen in Chrome with presentation mode, and various smaller resolutions. Firefox doesn’t have presentation mode, and at 1080p the page is slightly too tall unless you turn off the bookmarks bar (at which point you can juggle things in full screen mode).

If you can’t manage 1080p, the best width is probably a little under 1500px (which triggers the two column layout), since it makes the text nice and large.

There may be some resolutions where the textarea scrolls, although they should be minimal. This hasn’t been tested on real tablets, only emulated. All the sizing is based on early pages of MA8, and so will undoubtedly fail on other missions and possibly on later pages of that one.

Authors

With assistance and ideas from the rest of the Spacelog team.

About

Crowd-sourced transcription cleaning for Spacelog

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •