Hi All,
I have broken the back on my personal itch to want low friction markdown based citation management. I have a basic release for the curious at https://github.com/singingfish/Citeproc-Markdown - it’s a bit more than proof of concept but lacks pandoc integration until I actually write some text containing my special citation keys.
Two particularly useful things about this:
- The test file demonstrates how to get useful information out of the csl processor in a reasonably self documenting way.
- It shows how to get reasonably reliable results feeding javascript into mozrepl, so no need to install extra firefox/xulrunner extensions.
There’s a tarball in the github repo that provides code that will install using existing cpan infrastructure. I won’t release it to CPAN until it’s generically testable.
The README is below:
NAME
Citeproc::Markdown
DESCRIPTION
Package for handling human readable Author/Date citations in markdown
format. Designed to be good enough for common use cases, not perfect.
Code is also a good start to a more general Zotero/Perl gateway.
SYNOPSIS
Will scan a plain text document paragraph by paragraph for citations
using a human readable format to key citations. Conversion will die
without modifying the document if there are ambiguous citation keys.
Requires mozrepl
(<https://addons.mozilla.org/en-us/firefox/addon/mozrepl/>) installed
and running to the same Firefox, or XULRunner that your Zotero library
is stored in. Other than that, and a working modern perl ( >= 5.10.0) no
other extensions, firefox or otherwise are required.
Examples of the format are
(c|Fletcher 2003 Mapping stakeholder perceptions)
(c|Law 2008 On sociology)
(s|Law 2008 On sociology)
The final example is to supress author (not yet implemented, but
supported in the regex). The code will warn if more than one keys are
found (maybe it should die ...)
TODO
1. Tests work on my local machine but not elsewhere due to citation
library differences and zotero setup.
2. Tests only work with a running
zotero and mozrepl.
3. Need to write the pandoc integration (need to
write some markdown with citations).
4. Supress author citations not yet
supported (but stubbed (s| form of citation to support this at a later
stage.
5. Consider adding compatible zot4rst citation keys.
However, this module provides the basis for having decent
markdown/zotero integration without the need for intermediate files. In
the final implementation I suppose there will be two different versions
of the script run "markdown_cite --draft" that will keep the author cite
keys untouched, and dump the references to a file, and "markdown_cite
--final" that will replace them with the final CSL generated citations.
BUILD
Loads the citeproc javascript required for this code to work.
json_encoder
JSON::Any object used in data transfer between repl and perl
js_dir
sharedir where we keep the javascript required for mozrepl
add_citation
takes a list of citation ids, and adds them to the csl processor through
the repl.
search
Takes a citation string, parses it returns the item id. Warns if > 1
result is returned.
citations
hashref of citations seen during document processing, keyed by the
citation text provided by the user. Used to store citations for
publication indexed by writer’s citation keys.
repl
MozRepl object for internal use. Run script with env var DEBUG=1 for
verbose info, otherwise only warnings and fatals are emitted. DEBUG=1
will also catch JSON encoding problems.
run
sends javascript commands to the repl and returns the result of the last
command.
run_file
Reads in javascript source code, and sends it to the repl paragraph by
paragraph(delimited by \n\n). WARNING - be sure to ensure that each
paragraph compiles as a standalone entity. If not the repl will hang
then time out.
parse_citation
parses the citation to author title and year hashref
citation_regex
Simple regex for parsing the text string. TODO. consider making a proper
parser. TODO. consider optional doi support.
available_styles
Lazy accessor for the available zotero styles. Returns a hashref: key:
name val: url.
set_style
Uses instantiateCiteProc in citeproc.js to set the current style.
extract_citation_list
Takes a list of cites (c|Whatever 1999 Title fragment)(c|Someone 2002
Stuff) etc and splits into an array for further processing.
make_bibliography
Create the bibliography after all citations have been processed.
process_citation
Given an in-text citation (could be one or more ([cs]| … form
citations, append the final_cite key to its hashref to give the in final
(publisher) citation for that chunk of text.
ACKNOWLEDGEMENTS
Erik Hetzner for the javascript code in zot4rst, which is also used in
this project (with very minor documentation and naming changes).
Frank Bennett for the very useful citeproc documentation
<http://gsl-nagoya-u.net/http/pub/citeproc-doc.html>, which with Erik's
code enabled me get something that was usable running.