Group_4_workflow.md

Group 4 workflow

1. We set up a GitHub repository for the project: https://github.com/djbpitt/tufts_2018 ; with the following folder structure:

• (at the root):

  • index.html (the main page of our website, see below)

  • site.css: main css for the website (but not the only one)

  • README.md: markdown file that describes the repo

• (and these subfolders:)

  • composite: holds the html files for the pages of our digital edition of each plant (generated by XSLT, see below)

  • images: holds the images of the four plants from the French edition (edited by hand to create a transparent background with Gimp; format needs to be PNG, because jpeg does not allow transparent backgrounds)

  • meta: holds the html files for the metadata pages for each historical edition (generated by XSLT, see below)

  • old: files we don't need anymore (similar folders can be found at other levels)

  • reading_views: holds folders for each language, each of which holds for every plant

    • tt.xml files (converted from the original EXMARaLDA treetagger export) that have a numbered list of the sentences in the plant description (see below)

    • html files that were converted from these tt.xml files (see below)

  • sources: holds folders for each language, each of which contains data provided by the other teams (and Thomas), which had to be sometimes manually corrected:

    • .exb export files for every plant from EXMARaLDA

    • .tt (treetagger) xml files for every plant, converted from .exb via pepper by Thomas

    • metadata xml files for the book level

    • a pdf of the relevant pages of the historical edition

  • stemma: holds the GraphViz .dot file that produces the stemma, and the stemma svg itself (lightly edited by hand)

  • util: holds the xslt and python files, with some support files, for development (see below)

  • wireframes: contains a number of tests and sketches for the design of the website

2. We set up a GitHubPages website for the project (go to the properties of the project and set it to allow the setup of a GitHubPages site): https://djbpitt.github.io/tufts_2018

• When you create GitHubPages, a .yaml file is created; in GitHubPages you can use Jekyll templates (but we didn't). For more on Jekyll, see: https://jekyllrb.com , https://programminghistorian.org/en/lessons/building-static-sites-with-jekyll-github-pages

3. Homepage workflow:

• We sketched out on paper what the website should look like (see pictures in the wireframes directory); thinking about how users would like to interact with the site, we decided to make separate pages for every plant, linked from the main page

• We put the index.html file, which we built by hand, in the Github repository's root directory

• Three main sections on the main page:

  • Links to the plant pages: we took images from the French edition to hold the links to the plant pages

  • Stemma: We made a quick (but thoughtful) manual comparison of the texts of one plant to build a provisional stemma

  • Bibliography table with links to the pdf editions + metadata of every edition

4. Metadata of the editions workflow:

• converted the metadata files the groups produced to (well-formed) xml files, partly with a python script (in /tufts_2018/util), partly by hand

• Annis link in the metadata: in Annis, we made a search for "sentence" in the relevant corpus, copied the link to the search results, and added it to the metadata file (xml base file)

• With an XSLT file (/tufts_2018/util /meta-to-html.xsl), we transformed these xml metadata files into html pages linked from the index.html page, and residing in the meta subdirectory

5. Digital edition workflow:

• Thomas used pepper to convert the EXMARaLDA .exb files to treetagger (.tt) xml files

• We started from .tt files (in /tufts_2018/sources (renaming them to be consistent, and fixing a number of xml errors by hand))

• Using the tt_to_html.xsl file, we transformed these to .tt.xml files that have numbered list of sentences; these were put into the language subdirectories of the reading_views directory.
  $ saxon --s:botrys.tt --o:botrys.tt.xml --xsl:../../util/tt_to_html.xsl

  (we later automated this by writing a shell script, reading_views.sh, in the util directory; it has to be run from the main directory)

• We hand-built an html table for every plant, that provides a comparison of the sentence numbers in every language (/tufts_2018/util/mapping.html). Canonic line numbers were added to these tables to make comparison possible.

• Using the util[/add_sentence_numbers.xsl]{.underline} (running it from the language subdirectories of the reading_views directory), we added the canonic line numbers in the @class attribute of the ordered lists:

  $ saxon --s:botrys.tt.xml --o: botrys.html --xsl:../../util/add_sentence_numbers.xsl

• The final plant html files were built with combine_[plantname].xsl scripts (in the ideal world, we would have made just one xslt script that built all these pages), and put into the composite directory.
  Run "[combine_artemisia_herba-alba.xsl]{.underline}" (from the util directory): (BUT: adapt the file names + the number of files we are looping over + title+header)

  $ saxon --it --xsl:combine_artemisia_herba-alba.xsl --o:../composite/ambrosia.html
  NB: -it = initial template: to use if there is not a direct input file for the xslt

• In order to have corresponding sentences light up when hovering over a sentence, we produced a javascript file (in /composite/scripts.js)

• We also added the metadata on each plant in each language, in a popup that appears when clicking the small circled ⓘ after each language title; and a link to a pdf of a scan of the relevant historical edition (which is located in the sources/[language] folder).

NB: saxon is an xslt engine (the same one used by oXygen)

Running saxon from Windows:

Download Saxon HE9 and save it in a directory (make sure that the path does not include spaces)

Instead of simply the command saxon, use:
java --jar C:/Programme/SaxonHE9-8-0-12J/saxon9he.jar

You can create an alias for saxon by typing:
doskey saxon= java --jar C:/Programme/SaxonHE9-8-0-12J/saxon9he.jar

Mac users can install saxon using homebrew:

$ brew install saxon