Re: WIP: better citation processing

[John MacFarlane <jgm-TVLZxgkOlNX2fBVCVOL8/A@public.gmane.org>]

Joseph Reagle <joseph.2011-T1oY19WcHSwdnm+yROfE0A@public.gmane.org> writes:

> On 8/16/20 6:59 PM, John MacFarlane wrote:
>> I've used this library to create a new filter,
>> new-pandoc-citeproc, which now passes most of the pandoc-citeproc
>> test suite but runs around 6 times faster.
>
> Awesome! I've topped 10k items in my YAML file. Even though you kindly sped up some citeproc bottlenecks for me in the past, I continue to "subset" the 10k file into document-specific YAML files when building a document. That is, it's faster for me to regex my document for citations and pull their entries out of the 10k YAML file before handing it off to pandoc than just using pandoc itself.

There are two separate issues here:

1.  Parsing the YAML metadata (this is done in pandoc's markdown reader)
2.  Processing the CSL (this is done by pandoc-citeproc)

The new library will speed up 2, but it won't affect 1, and I
suspect 1 is the bottleneck for you.

For #1, we have this open issue:

https://github.com/jgm/pandoc/issues/6084

The issue could be fixed by switching back from the pure Haskell
HsYaml to the wrapped C library yaml.  I hate to do that, though,
because I've been trying to remove all pure C library
dependencies from pandoc (both for security reasons and
because they don't work with e.g. compiling to JavaScript
with ghcjs).

Here's a workaround that should work right now.  Reading CSL JSON
is fast, so you could try using pandoc-citeproc -j to convert
your YAML bibliography to CSL, then refer to the CSL bibliography
in your pandoc metadata.

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/m2v9hbbyyu.fsf%40johnmacfarlane.net.