Incorrect handling of {\L}

[blegat]

Create bug.bib:

@inproceedings{vaswani2017Attention,
  title = {Attention Is {{All}} You {{Need}}},
  author = {Kaiser, {\L}ukasz},
}

and run

using DocumenterCitations
bib = CitationBibliography("bug.bib")
key = "vaswani2017Attention"
DocumenterCitations.format_bibliography_reference(bib.style, bib.entries[key])

The error is:

ERROR: ArgumentError: Premature end of tex string: BoundsError("{. Kaiser", 11)
Stacktrace:
 [1] tex_to_markdown(tex_str::String; transform_case::Function, debug::Base.CoreLogging.LogLevel)
   @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:135
 [2] tex_to_markdown
   @ ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:125 [inlined]
 [3] 
   @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/formatting.jl:178
 [4] format_names (repeats 2 times)
   @ ~/.julia/dev/DocumenterCitations/src/formatting.jl:93 [inlined]
 [5] format_labeled_bibliography_reference(style::Val{…}, entry::BibInternal.Entry; namesfmt::Symbol, urldate_accessed_on::String, urldate_fmt::Dates.DateFormat{…}, title_transform_case::Function, article_link_doi_in_title::Bool)
   @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/labeled_styles_utils.jl:253
 [6] format_labeled_bibliography_reference(style::Val{:numeric}, entry::BibInternal.Entry)
   @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/labeled_styles_utils.jl:244
 [7] format_bibliography_reference
   @ ~/.julia/dev/DocumenterCitations/src/styles/numeric.jl:22 [inlined]
 [8] format_bibliography_reference(style::Symbol, entry::BibInternal.Entry)
   @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/expand_bibliography.jl:107
 [9] top-level scope
   @ ~/.julia/dev/DocumenterCitations/test.jl:7

caused by: BoundsError: attempt to access 10-codeunit String at index [11]
Stacktrace:
  [1] checkbounds
    @ ./strings/basic.jl:216 [inlined]
  [2] codeunit
    @ ./strings/string.jl:138 [inlined]
  [3] getindex
    @ ./strings/string.jl:461 [inlined]
  [4] _collect_group(tex_str::String, i::Int64)
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:407
  [5] _process_tex(tex_str::String; transform_case::DocumenterCitations.var"#73#75", debug::Base.CoreLogging.LogLevel)
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:196
  [6] tex_to_markdown(tex_str::String; transform_case::Function, debug::Base.CoreLogging.LogLevel)
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:131
  [7] tex_to_markdown
    @ ~/.julia/dev/DocumenterCitations/src/tex_to_markdown.jl:125 [inlined]
  [8] 
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/formatting.jl:178
  [9] format_names (repeats 2 times)
    @ ~/.julia/dev/DocumenterCitations/src/formatting.jl:93 [inlined]
 [10] format_labeled_bibliography_reference(style::Val{…}, entry::BibInternal.Entry; namesfmt::Symbol, urldate_accessed_on::String, urldate_fmt::Dates.DateFormat{…}, title_transform_case::Function, article_link_doi_in_title::Bool)
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/labeled_styles_utils.jl:253
 [11] format_labeled_bibliography_reference(style::Val{:numeric}, entry::BibInternal.Entry)
    @ DocumenterCitations ~/.julia/dev/DocumenterCitations/src/labeled_styles_utils.jl:244

[goerz]

I'll have a look, but TeX escape codes are only partially supported. The recommended approach is to use unicode, as in

@inproceedings{vaswani2017Attention,
  title = {Attention Is All You Need},
  author = {Kaiser, Łukasz},
}

You should avoid those extra TeX groups inside the title as well. See the Syntax for the .bib file

[blegat]

I don't generate this manually, this is the output of Zotero's Better BibTeX extension. It's encoded with unicode in Zotero but the exporter turn them into LaTeX friendly escaped symbols.

[goerz]

Zotero produces broken .bib files and is absolutely not supported as a source for DocumenterCitations. As a full entry,

@inproceedings{vaswani2017Attention,
  title = {Attention Is {{All}} You {{Need}}},
  author = {Kaiser, {\L}ukasz},
}

is not valid, since it's missing the mandatory year and booktitle fields.

See also #76.

However, the handling of the Ł in this particular case is a bug in DocumenterCitations, and I'll work on a fix for that.

[blegat]

I actually removed fields from the output of Zotero to reduce the MWE

[goerz]

Sure! That's fine!

I just wanted to stress the point that DocumenterCitations is intended to be used with "hand-written" .bib files that use unicode. They should avoid TeX commands, since the strings are not actually processed by TeX.

Of course, I encourage having automatic tooling for generating suitable .bib files, but many export tools from Bibliography managers do not generate suitable output. Based on #76, Zotero seems to be particularly bad and produces files that aren't even valid in a LaTeX context. Downloaded .bib files from publisher websites are even worse (which is why I wrote myself a getbibtex script).