Given a mediawiki where multiple documents (each spread across multiple pages) are developed. Provide an easy way to translate that into a separate PDF per document. Use pandoc/latex to do so.
Specific requirement: supported embedded plantUML in the pages, use plantUML to produce the figures.
- Provide a page in the wiki where you list all the documents you want to have translated.
- In that page, put a bullet list with names of wiki pages, one line per document
- These pages again contain a bullet list of wiki pages. These pages will be contained in the final document, in the order in which they are listed on this page
- Feel free to use plantuml in your pages.
- Put configuration data in config.py: user name, password, etc.
- run build.py
- it will download wiki pages, figures, etc.
- extracts uml, translates it via plantuml
- uses pandocs to translate into latex
- runs pdflatex
- uploads the pdfs and results page
- For each document mentioned in the documentlist, there will be a file .pdf uploaded and a page with the name of documnet and BuildReport added.
- BuildReports contains error messages as well as a link to the PDF. They are also added to a category BuildReport
- They also have a link to the generated and uploaded PDF.
- Better handling of figures: captions, size, ...
- plantuml jar file is included here, but look for a more up-to-date version
- python modules needed (see requirements.txt): mwclient, pypandoc, pandocfilters, hunspell
- installation: python, java, pandoc, hunspell, libhunspell-dev
- So basically:
- git clone mw2pdf
- think about whether you want a virtualenv; if so, set it up and activiate it
- pip install -r requirements.txt
- python build.py -h
Should not take longer than five minutes if you are roughly familiar working with a python environment.
- python build.py --download --latex --ignore-fingerprint
- python build.py --download --latex --ignore-fingerprint --document D4.1_Orchestrator_prototype
- python build.py --download --latex --ignore-fingerprint --document WP2_Deliverable_2.2
- python build.py --download --latex --ignore-fingerprint --document HolgersDocument