kwarc / bibs Goto Github PK

The main problem with the current publication list generation process is the crossref field in the bibTeX. This leads to wrong results and inflated counts, see brucemiller/LaTeXML#800 for details. Once this is fixed, we can acually use the publication lists in kwarc.info.

We should generate publications for all the KWARC people (or at least all with publications).
So we should know all the places where we have to add the name.
I would guess that Theresa and Jonas will come soon.

It would be good to publish the generated *.bib.xml. (and the logs for cleaning up the bib sources) on a branch (normally, you do not want the XML; only the bibTeX).
I suspect that this should be quite easy to do.

In several places we have "Normen Müuller". It looks to me like that is a typo and should actually be Normen Müller.

@kohlhase Can you confirm?

In the merged PR #26 I tried integrating masters and bachelors projects on almost all pages.
However, although nothing broke, no pages are showing the projects, e.g. https://kwarc.github.io/bibs/nroux/ should show a project on diagram operators.

Debugging so far:

• the gh-pages branch has an mscproject folder with an index.html. That index.html is, except boilerplate, empty.

Apparently, kwarc.bib contains a syntax error I cannot see. Perhaps integrating validation in the build process would be a goot mitigation action.

Anyway, running biber --tool -V kwarc.bib gives me

ERROR - BibTeX subsystem: C:\<temp path>\kwarc.bib_6404.utf8, line 43527, syntax error: found ".", expected end of entry ("}" or ")") (skipping to next "@")
INFO - WARNINGS: 1738
INFO - ERRORS: 1

where the indicated line is the one with publisher = {Routledge}, in the following excerpt from the file referenced in the error message above:

@techreport{w3c04:rdf-testcases,
  title = {{RDF} Test Cases},
  institution = {World Wide Web Consortium ({W3C})},
  year = {2004},
  date = {2004-02-10},
  type = {{{W3C} Recommendation}},
  month = {2},
  author = {Jan Grant and Dave Beckett},
  url = {http://www.w3.org/TR/2004/REC-rdf-testcases-20040210/},

}

@book{Harding:FeministStandpoint,
  title = {{The Feminist Standpoint Theory Reader: Intellectual and Political Controversies}},
  publisher = {Routledge},
  year = {2004},
  editor = {Sandra Harding},

}

@book{Harding:FeminismAndMethodology,
  title = {{Feminism and Methodology}},
  publisher = {Indiana University Press},
  year = {1987},
  editor = {Sandra Harding},

}

We should have a short paragraph in the README.md of master to say that we are building publication listings for persons and projects/systems via LaTeXML and say that to see things you should just go to the gh-pages branch.
Then we should document in a README.md on gh-pages how the generation works.
Otherwises I will always forget.

This should be relatively easy to add to bibs/src/pubs/publist.xsl.

I have changed the Makefile so that only the parts XML files are rebuilt that need to be rebuilt.
But the comparison mechanism when things should be built in travis is not updated yet.
Here is what should happen.
kwarcpubs.bib --> rebuild kwarcpubs.bib.xml and pubs
extpubs.bib --> rebuild extpubs.bib.xml
kwarccrossrefs.bib --> rebuild kwarccrossrefs.bib.xml and pubs
extcrossrefs.bib -- rebuild extcrossrefs.bib.xml and pubs

for the new kwarc web site, we would like to have an overall publication page (i.e. for all the group). (see https://gl.kwarc.info/kwarc/website/issues/12)
This should probably replace the somewhat useless https://kwarc.github.io/bibs/

We should move away from Travis and instead go to GitHub Actions; this is one of the few repos still using Travis.

If we had that, we could retire the old process and move from the SVN to git. We want to get rid of SVN, and this is one of the places where we still have it.
@tkw1536 @jbs1

Example tuples (deployed title, bib source code) from https://kwarc.github.io/bibs/ and this repo:

• FrameIT: detangling knowledge management from game design in serious games

  {{FrameIT}: Detangling Knowledge Management from Game Design in Serious Games}

• GLIF: a declarative framework for symbolic natural language understanding

  {{GLIF}: A Declarative Framework for Symbolic Natural Language Understanding}

• KAT: an annotation tool for STEM documents

  {{KAT}: an Annotation Tool for {STEM} Documents}

Is the lower-casing intended?

See also https://gl.kwarc.info/kwarc/website/issues/10 for the application. We have to coordinate project names for this to work.

Currently, the deploy.sh script running on Travis abort if it notices that no bib source files have changed in the commit for which Travis is running:

This procudes too many false negatives when pushing multiple commits at once. In both cases below, Travis only considers the most recent commit.

• consider two local-only commits, where the first changes a source file, and the second one something else. When pushing those commits (two-in-one), the latter commit will make the script above prematurely abort.
• consider doing a local commit, then merging upstream changes via git pull, and then pushing.

Probably this is too minor to warrant fixing; feel free to close.

We should really keep this repository up-to-date from SVN automatically via a POST-commit hook, similar to the one we have for the MMT repository.

@kohlhase What do you think? If needed I can set one up.

cc @richardmarcus

e.g. on https://kwarc.github.io/bibs/rmarcus/, the FrameIT paper doesn't appear. I think this regression appeared after I added additional bibliographic details (publisher, volume, abstract, ...). Not sure why it broke.

Are there any error logs available?

e.g.

Wrote dist/html/cjucovschi-thesis.html
src/html/generate-html: line 27: 29362 Segmentation fault: 11  latexmlc $LTOPT --destination=${DIST}html/$i-$j.html ${DIST}tex/$i-$j.tex 2> >(tee ${DIST}tex/$i-$j.ltxlog >&2)

I have annotated all of kwarcpubs.bib with the three keywords phdthesis, mscthesis, bscthesis, from which I would like to generate four publication gh-pages: the three obivous ones and a combined one.
I guess that this should be easy to do for Tom.

It would be really good to have this soon.

We have started to use the pubs= feature for project publications, e.g. http://kwarc.github.io/bibs/MaMoReD/ But these bibliographies are only generated from kwarcpubs.bib and *crossrefs.bib. But the projects also have outside publications that need to be cited, and these are (currently) in extpubs.bib and are thus not considered. We could

• move them kwarcpubs.bib and reinterpret thhttp://kwarc.github.io/bibs/MaMoReD/at as "publications of KWARCies and in projects KWARCies are in"
• leave them there and build from extpubs.bib as well.

If I understand correctly, then travis already builds the html files. Would it be possible to commit them back so that we can use them from kwarc.info? Then we could fully shut down http://old.kward.info, which would be very good.
@jbs1 @m-iancu