A revived project that converts different document formats to a prettified output, such as PDF or latex.
Instead of writing our own complete converters we decided to a well-established Pandoc framework http://pandoc.org/ to do most of the heavy lifting of the conversion.
Pandoc is not capable of providing prettified documents out-of-the-box so we have to perform some optimizations and document manipulations in order to be Pandoc friendly. This means that we perform some pre-processing and post-processing of the documents before and after Pandoc.
- Simple Build Tool - install the latest version of sbt
- Pandoc - requires a local build of the master branch that contains our latest contributions (there was no release that contains them, yet). Since some of our PRs may still be in the queue to the official master we have a fork of Pandoc and a special branch called
master-typesetrthat contains the latest changes. Same applied to Pandoc's definitions, pandoc-types. - Scala - (optional)
sbthandles all the compilation process and provides a console but you may still find it useful
Then enter sbt's console
> sbt
(wait for the Internet to download if that's the first time you did it)
sbt> project core
sbt> run
This will display a number of parameters (required or optional).
We also need to checkout the old typesetr where the existing
templates are located.
For example, you can transform an ODT document into a pdf using the
following command:
sbt> run -i odt -o pdf --infile $INPUTFILE.odt --outfile $OUTPUTFILE.pdf -s lshift/pitch --style-base $TYPESETRSTYLES
where
$INPUTFILE.odt- is some ODF-compliant file$OUTPUTFILE.pdf- is the desired location where the pdf is to be createdlshift/pitch- determines the template/style to use$TYPESETRSTYLES- is the location where all styles are defined
To test, simply run:
sbt> test
- [o] headers - no sections above heading level 5
- inline code - ODT Parser in Pandoc does not recognize any code blocks (other input formats support it). See PR
- removal of bogus paragraph elements - won't be implemented (Pandoc understands those correctly)
- list items - custom typesetr's styling - desired?
- list items - starting/restarting numbering - currently Pandoc hardcoded 1 as a starting number for ODT documents (see PR)
- bullet items - only a single character representing bullet points (original typesetr had the possibility of 'disc', 'circle', 'square', 'hyphen'). Will reuse lshift style.
- tables are always empty (see PR)
- images - not supported in the ODT Parser in Pandoc (see PR and follow ups)
- [i] floating figures - not supported in Pandoc's LaTeX writer (see our PR)
- labels - needs to translate local href references to bookmark references
- inline math formulas - Pandoc translates backslashes to \textbackslash{}, which breaks the latex conversion
- simple tables - currently ignored
- inner quoations (American style) - unknown state
- negative numbers - (also broken in the old typesetr)
- [i] page breaks - Pandoc limitation, see our PR
- page numbering - total number of pages is broken
- table of contents
- links as new commands - remains to be decided whether it is necessary with the new design
- deeply nested items - currently not supported (Pandoc limitation). We added a workaround in our templates.
- [o] odt - no native support for inline math text - we can add Pandoc support for it potentially. Partially supported by doing some postprocessing.
- [o] multiple text style changes - Pandoc only picks the first one (see issue). Looks like it is partially the fault of Typesetr's templates.
- underline formatting - Pandoc limitation Legend:
- empty - not supported
- o - partially supported
- i - implementation in progress
- w - won't be implemented
- X - implemented
- duplicate entries for references, bookmarks etc (see reported bug)
Unlike the old oversion of Typesetr, Pandoc does not handle heading styles above the level 5. While a workaround can be integrated, currently no styling is applied above level 5.