Opened 2 years ago

Closed 2 years ago

#841 closed defect (fixed)

INSPIRE: Random outstanding LaTeX format issues

Reported by: jblayloc Owned by: jblayloc
Priority: major Milestone:
Component: *general* Version:
Keywords: INSPIRE formats Cc:

Description

authors should be written like foo, bar and baz rather than foo, bar, baz

major because users keep complaining about it

Change History (20)

comment:1 Changed 2 years ago by jblayloc

I've just deployed Mike's branches to -dev.

Jan, please double-check my integration; everything looks clean and kwalitee looks good to me on top of current inspire master, but I particularly want to be sure this integrates well with the formats work that Lewis and you have been doing lately. You can check out the cleaned, kwalitee-checked, squashed branch from my inspire github repo, branch '829-multiple_format_fixes-sul'.

Heath, please check that the following things look right:

  • . added for pub-note/arxiv numbers in HTML Latex format: it should be there when it's supposed to be, but not when its not. (I'm afraid I'm not certain when that is. Mike?)
  • multiple authors rendered as foo, bar and baz rather than foo, bar, baz in HTML Latex output: we've been hearing about this a lot lately so let's be sure.
  • pub-notes produces multiple note outputs rather than just one. Multiple pub-notes stack vertically in the detailed format. Similarly there should be no blank line when there are no pubnotes. (See also #829)
  • Journal title text added to Journal Server link on detailed record page. e.g., "Journal Server - Phys.Rev.Lett." or something like that.

If everything looks good to you Heath and you Jan, then I'm satisfied that it can be merged into inspire master and deployed to production.

comment:2 Changed 2 years ago by jblayloc

  • Status changed from new to in_merge

comment:3 Changed 2 years ago by hoc

Hi Joe,

  1. THere should be no [] around the eprint number (except the primarch part, e.g. [hep-ex]) when there is no journal publication note.

e.g. the first entry is correct, the second is not.
%\cite{} *I'll note here that there is nothing inside the {}, start putting the eprint number there*
\bibitem{}

R.~Longo and E.~Witten,
%An Algebraic Construction of Boundary Quantum Field Theory, Commun.\ Math.\ Phys.\ {\bf 303 } (2011) 213-232 [arXiv:1004.0616 [math-ph]].

%\cite{Witten:2009mh}
\bibitem{Witten:2009mh}

E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny, [arXiv:0905.4795 [hep-th]].

  1. When there is no eprint or pub note, a report number or conference note should be used. Compare INSPIRE

%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,

with SPIRES
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
%%CITATION = HUTP-79/A078;%%

Wow! This is wrong in SPIRES too, ok we have a chance here to improve on SPIRES. It should say:
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
HUTP-79/A078.
%%CITATION = HUTP-79/A078;%%

Good job on the "foo, bar and baz" I think that's all perfect now. The construction of the journal notes looks great. There is one issue with volume letters. They typically go outside the \bf part (finnicky, I know). See SPIRES
%\cite{Papadopoulos:1982zz}
\bibitem{Papadopoulos:1982zz}

D.~Papadopoulos and L.~Witten,
%Static axisymmetric solution of the Einstein equations,
Phys.\ Rev.\ D {\bf 25} (1982) 1465.
%%CITATION = PHRVA,D25,1465;%%

vs INSPIRE
%\cite{Goswami:2004fx}
\bibitem{Goswami:2004fx}

R.~Goswami, P.~S.~Joshi, C.~Vaz and L.~Witten,
%A time-like naked singularity,

Phys.\ Rev.\ {\bf D70 } (2004) 084038 [gr-qc/0410041].

Are we going to put the %%CITATION back in now? How long would that take?

comment:4 follow-ups: Changed 2 years ago by jblayloc

  • Status changed from in_merge to assigned
  1. THere should be no [] around the eprint number (except the primarch part, e.g. [hep-ex]) when >there is no journal publication note.


e.g. the first entry is correct, the second is not.
%\cite{} *I'll note here that there is nothing inside the {}, start putting the eprint number there*
\bibitem{}

R.~Longo and E.~Witten,
%An Algebraic Construction of Boundary Quantum Field Theory, Commun.\ Math.\ Phys.\ {\bf 303 } >(2011) 213-232 [arXiv:1004.0616 [math-ph]].

%\cite{Witten:2009mh}
\bibitem{Witten:2009mh}

E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny, [arXiv:0905.4795 [hep-th]].

So the second one should be written:
%\cite{Witten:2009mh}
\bibitem{Witten:2009mh}

E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny, arXiv:0905.4795 [hep-th].

??

comment:5 Changed 2 years ago by jblayloc

  • Keywords onDevPleaseReview removed
  • Owner changed from sul to jblayloc

comment:6 in reply to: ↑ 4 ; follow-up: Changed 2 years ago by jblayloc

And also, how important is point #1? It's not apparent that there is a way to pass data between different format elements in the format system in something resembling a clean way, so having the arxiv output element know whether or not the pubnote element actually output anything is tricky and will take some thought.

comment:7 follow-up: Changed 2 years ago by jblayloc

  • Summary changed from LaTeX format: author foo, bar and baz to INSPIRE: Random outstanding LaTeX format issues
  1. When there is no eprint or pub note, a report number or conference note should be used. Compare >INSPIRE

%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,

with SPIRES
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
%%CITATION = HUTP-79/A078;%%

Wow! This is wrong in SPIRES too, ok we have a chance here to improve on SPIRES. It should say:
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
HUTP-79/A078.
%%CITATION = HUTP-79/A078;%%

Ok, so %%CITATION should fall back to report number, too. Do you know where the report number or conference number are stored? I think conference is in 773__w. I don't know about report number though. Having report number fallback should be reasonably easy; right now I think it's already falling back to 773__x.

There is one issue with volume letters. They typically go outside the \bf part >(finnicky, I know). See >SPIRES
%\cite{Papadopoulos:1982zz}
\bibitem{Papadopoulos:1982zz}

D.~Papadopoulos and L.~Witten,
%Static axisymmetric solution of the Einstein equations,
Phys.\ Rev.\ D {\bf 25} (1982) 1465.
%%CITATION = PHRVA,D25,1465;%%

vs INSPIRE
%\cite{Goswami:2004fx}
\bibitem{Goswami:2004fx}

R.~Goswami, P.~S.~Joshi, C.~Vaz and L.~Witten,
%A time-like naked singularity,

Phys.\ Rev.\ {\bf D70 } (2004) 084038 [gr-qc/0410041].

Ok. This is just because it's stored in the database as D70. We should be able to do a little light parsing to peel off the leading letters from the volume subfield and stick them on the end of the name. I can do this quickly.

Are we going to put the %%CITATION back in now? How long would that take?

I'm not sure how hard this is. It takes some thought. Perhaps a day? I'm going to start working on it wherever it falls in the priority queue with these other issues.

Could you please give me rough ordering of desirability for these issues:

  • %% CITATION
  • Fallback to report number
  • volume letter outside {
    bf}
  • bracket placement on arxiv ref for missing pubnotes

I suspect %%CITATION is most important?

comment:8 in reply to: ↑ 4 Changed 2 years ago by hoc

Replying to jblayloc:

  1. THere should be no [] around the eprint number (except the primarch part, e.g. [hep-ex]) when >there is no journal publication note.


e.g. the first entry is correct, the second is not.
%\cite{} *I'll note here that there is nothing inside the {}, start putting the eprint number there*
\bibitem{}

R.~Longo and E.~Witten,
%An Algebraic Construction of Boundary Quantum Field Theory, Commun.\ Math.\ Phys.\ {\bf 303 } >(2011) 213-232 [arXiv:1004.0616 [math-ph]].

%\cite{Witten:2009mh}
\bibitem{Witten:2009mh}

E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny, [arXiv:0905.4795 [hep-th]].

So the second one should be written:
%\cite{Witten:2009mh}
\bibitem{Witten:2009mh}

E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny, arXiv:0905.4795 [hep-th].

??

Yes, except something weird happened with the format. The author, title and whatever paper-code note should be on separate line:
E.~Witten,
%Geometric Langlands And The Equations Of Nahm And Bogomolny,
arXiv:0905.4795 [hep-th].

"%" is a comment to latex and it wipes out everything until the end of the line. So the title should be commented out because typically people don't want it but if they do, it's there for them.

comment:9 in reply to: ↑ 6 Changed 2 years ago by hoc

Replying to jblayloc:

And also, how important is point #1? It's not apparent that there is a way to pass data between different format elements in the format system in something resembling a clean way, so having the arxiv output element know whether or not the pubnote element actually output anything is tricky and will take some thought.

OK. I think in SPIRES what we did was create variables like $lbracket and $rbracket which were initialized as "" but got the values 'and?' if there was a pub note. Then the eprint was always written sandwiched between the variables. Don't know if this works in INSPIRE.

comment:10 in reply to: ↑ 7 Changed 2 years ago by hoc

Replying to jblayloc:

  1. When there is no eprint or pub note, a report number or conference note should be used. Compare >INSPIRE

%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,

with SPIRES
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
%%CITATION = HUTP-79/A078;%%

Wow! This is wrong in SPIRES too, ok we have a chance here to improve on SPIRES. It should say:
%\cite{Witten:1979pi}
\bibitem{Witten:1979pi}

E.~Witten,
%The 1 / N Expansion In Atomic And Particle Physics,
HUTP-79/A078.
%%CITATION = HUTP-79/A078;%%

Ok, so %%CITATION should fall back to report number, too. Do you know where the report number or conference number are stored? I think conference is in 773__w. I don't know about report number though. Having report number fallback should be reasonably easy; right now I think it's already falling back to 773__x.

I think the report number is stored in 037__a, see http://inspirehep.net/record/143469/export/hm

There is one issue with volume letters. They typically go outside the \bf part >(finnicky, I know). See >SPIRES
%\cite{Papadopoulos:1982zz}
\bibitem{Papadopoulos:1982zz}

D.~Papadopoulos and L.~Witten,
%Static axisymmetric solution of the Einstein equations,
Phys.\ Rev.\ D {\bf 25} (1982) 1465.
%%CITATION = PHRVA,D25,1465;%%

vs INSPIRE
%\cite{Goswami:2004fx}
\bibitem{Goswami:2004fx}

R.~Goswami, P.~S.~Joshi, C.~Vaz and L.~Witten,
%A time-like naked singularity,

Phys.\ Rev.\ {\bf D70 } (2004) 084038 [gr-qc/0410041].

Ok. This is just because it's stored in the database as D70. We should be able to do a little light parsing to peel off the leading letters from the volume subfield and stick them on the end of the name. I can do this quickly.

Are we going to put the %%CITATION back in now? How long would that take?

I'm not sure how hard this is. It takes some thought. Perhaps a day? I'm going to start working on it wherever it falls in the priority queue with these other issues.

Could you please give me rough ordering of desirability for these issues:

  • %% CITATION
  • Fallback to report number
  • volume letter outside {
    bf}
  • bracket placement on arxiv ref for missing pubnotes

I suspect %%CITATION is most important?

Yes. I'd probably move bracket up from 4 to number 2 position.

comment:11 Changed 2 years ago by jblayloc

Ok, my branch 841-manifold_fixes from github has been deployed on DEV for review. Heath, please look over Latex (US) and Latex (EU) and see if they seem like they're up to snuff. The issues that I think this branch should address are:

  • back button works properly when receding from detailed record pages
  • %% CITATION is created
  • When pubnote and arxiv numbers aren't available, things fall back to report numbers
  • volume letter is outside of {
    bf } for Latex formats
  • arxiv tags only get square brackets if there is also a pubnote

Also the above items should still be true:

  • . added for pub-note/arxiv numbers in HTML Latex format: it should be there when it's supposed to be, but not when its not
  • multiple authors rendered as foo, bar and baz rather than foo, bar, baz in HTML Latex output
  • pub-notes produces multiple note outputs rather than just one. Multiple pub-notes stack vertically in the detailed format. Similarly there should be no blank line when there are no pubnotes. (See also #829)
  • Journal title text added to Journal Server link on detailed record page. e.g., "Journal Server - Phys.Rev.Lett." or something like that.

Jan, if you get a minute it would be nice if you could check to see how this integrates with the work Lewis and you have been doing.

comment:12 Changed 2 years ago by hoc

Hi Joe,

Looks really good, we're almost there!

A few comments:

  1. Bibitem{}/cite{} empty (should have arXiv number or something else if no arXiv number)

%\cite{}
\bibitem{}

R.~Longo and E.~Witten,
An Algebraic Construction of Boundary Quantum Field Theory,

Commun.\ Math.\ Phys.\ \ {\bf 303 }, 213-232(2011)
[arXiv:1004.0616 [math-ph]].

cf.

%\cite{Larkoski:2010am}
\bibitem{Larkoski:2010am}

A.~J.~Larkoski and M.~E.~Peskin,
Top Quark Amplitudes with an Anomalous Magnetic Moment,

Phys.\ Rev.\ D\ {\bf 83 }, 034012(2011)
[arXiv:1012.0552 [hep-ph]].
%%CITATION = PHRVA,D83,034012;%%

  1. Above record with no bibitem also has no %%CITATION
  1. Above record has open space before "}", e.g. "{\bf 83 }" should be "{\bf 83}"
  1. Default for title should be to have a "%" in front, e.g. %Top Quark Amplitudes with an Anomalous Magnetic Moment, because people don't typically include the title but if they want it its then a simple regex to get it.
  1. No space between the page number and the date AND don't want page ranges, only first page, e.g. "213-232(2011)" should be "213 (2011)"
  1. Spaces around things inside square brackets are not wanted, e.g. \bibitem{Behnke:2007gj} T.~Behnke, (Ed.) {\it et al.} [ ILC Collaboration ], ILC Reference Design Report Volume 4 - Detectors, arXiv:0712.2356 [physics.ins-det]. %%CITATION = ARXIV:0712.2356;%%

"[ ILC Collaboration ]" should be "[ILC Collaboration]"

  1. Records can get confused about %%CITATION, e.g.

%\cite{Witten:1993xi}
\bibitem{Witten:1993xi}

E.~Witten,
The Verlinde algebra and the cohomology of the Grassmannian,

In *Cambridge 1993, Geometry, topology, and physics* 357-422
[hep-th/9312104].
%%CITATION = .....;%%


Even though there's an eprint number, the system does not give a %%CITATION

The file I tested it on is at:
http://www.slac.stanford.edu/~hoc/peskin.pdf
which has source
http://www.slac.stanford.edu/~hoc/peskin.tex

Heath

comment:13 Changed 2 years ago by jblayloc

Quick update: I have 3, 4, 5 addressed in my branch. I'm still trying to reproduce 6, 7. I understand 1 but haven't decided what to do about it yet.

comment:14 Changed 2 years ago by hoc

Showstopper: 1. Bibitem{}/cite{} empty
Can wait another week: 7 (it only fails for slightly weird records).
Nice to have done at some point: 6.

comment:15 Changed 2 years ago by jblayloc

Ok, thanks for the priorities Heath. This has been deployed to DEV again, with 1, 3, 4, 5 fixed. I have not addressed 2, 6, or 7 from immediately above. Specifically for number 2, %%CITATION is usually there, but not always. I need to spend a bit more time tracking down the precise reason why, but figured some %%CITATION was better than none.

Please confirm:

  • \bibitem{} and \cite{} look reasonable
  • no space inside \bf
  • commented-out title
  • page numbers formatted correctly

And also please check that I didn't break any old things:

  • Detailed format page looks correct (I had to manually merge changes here)
  • back button works properly when receding from detailed record pages
  • %% CITATION is created (at least sometimes)
  • When pubnote and arxiv numbers aren't available, things fall back to report numbers
  • volume letter is outside of {bf} for Latex formats
  • arxiv tags only get square brackets if there is also a pubnote
  • . added for pub-note/arxiv numbers in HTML Latex format: it should be there when it's supposed to be, but not when its not
  • multiple authors rendered as foo, bar and baz rather than foo, bar, baz in HTML Latex output
  • pub-notes produces multiple note outputs rather than just one. Multiple pub-notes stack vertically in the detailed format. Similarly there should be no blank line when there are no pubnotes. (See also #829)
  • Journal title text added to Journal Server link on detailed record page. e.g., "Journal Server - Phys.Rev.Lett." or something like that.

Thanks for your patience.

If this looks ok to you then I'll mark this ticket for closing, merge this branch into master and deploy it to PROD, then spin off a new ticket with the remaining outstanding issues 2, 6, 7.

comment:16 Changed 2 years ago by hoc

Hi Joe, just a few comments (in order of importance)

  1. {\bf 60}(2010) 439 -:- there should be a space -> {\bf 60} (2010) 439
  2. %%CITATION = ARNUA,60,439-462;%% -:- should be %%CITATION = ARNUA,60,439;%%
  3. Long titles could cause copy and paste errors due to wrapping, e.g.

\bibitem{arXiv:0810.0673}

M.~Pohl, A.~A.~Abdo, A.~Atoyan, M.~G.~Baring, J.~F.~Beacom, R.~Blandford, Y.~Butt and A.~Bykov {\it et al.},
%Section on Supernova remnants and cosmic rays of the White Paper on the Status and Future of Ground-based Gamma-ray Astronomy,

arXiv:0810.0673 [astro-ph].

should be so that the wrap part doesn't become exposed

%Section on Supernova remnants and cosmic rays of the White Paper on the Status and
% Future of Ground-based Gamma-ray Astronomy,

if this is difficult, it's not a big issue, just a nicety.

comment:17 Changed 2 years ago by jblayloc

I think most of this is now fixed, and is available on DEV for checking over. I'd like to mark this ticket 'fixed' and put this corresponding branch (841-manifold-fixes) into production, and then continue discussion about format problems over on #847. Heath, please advise.

comment:18 Changed 2 years ago by hoc

OK, Joe, this is ready to go. See you over on #847.

Heath

comment:19 Changed 2 years ago by jblayloc

  • Status changed from assigned to in_merge

This is now deployed into production.

comment:20 Changed 2 years ago by jblayloc

  • Resolution set to fixed
  • Status changed from in_merge to closed
Note: See TracTickets for help on using tickets.