CK

A command-line tool for managing your bibliography (i.e., .bib files and .pdf files) across multiple machines using Dropbox.

Features:

• add papers, sorted by citation key, to Dropbox
• easily open any paper given its citation key
• organize papers by tagging them
• generate a full .bib file of all your papers
• export papers with a particular tag

Setup

./install-deps.sh

Fill in ck.config and put it in your user_config_dir folder.

For auto tag-suggesting, you can install pdfgrep (optional):

apt install pdfgrep # Ubuntu/Debian
brew install pdfgrep # Mac OS

To install bash auto-completion, run

source bash_completion.d/ck

Other dependencies:

brew install pango libffi # Mac OS

Other useful, related repositories

• zotero translators
• BibFromXplore.sh
• AnyBibTeX
• iacr-dl
• bibcure
• scihub2pdf
• scholrref

Just search GitHub for more: "ieeexplore downloadcitations"

How to use

# add a paper to your library given a paywall URL (e.g., ACM DL, SpringerLink, IEEEXplore)
# or an eprint url (e.g., IACR eprint)
ck add <paper-url> <citation-key>

# add a bib file to your library without a PDF
ck open <citation-key>.bib
# ...and edit the .bib file and save it

# open a paper's PDF
ck open <citation-key>
ck open <citation-key>.pdf

# open a paper's .bib file
ck open <citation-key>.bib

# tag the paper with <tag> (or enter tag manually from keyboard)
ck tag <citation-key> [<tag>]

# search all your .bib files and print matching papers' citation keys
ck search <query>

Best practices

Q: How to deal with multiple published versions of the same paper (eprint, conference, journal)?
A: Have each version as a different CK, since it might contain additional info that needs to be cited.

TODOs

Bugs

• some springerlink URLs don't work because they have .ris citations only
  • e.g., https://link.springer.com/article/10.1007/s10207-005-0071-2

Features

Add PDF and BibTeX separately

Might want separate addbib and addpdf commands, to support adding the PDF and .bib file from different locations. Then, ck add can just call both of them, with overwrite=false

• If called individually, they will leave the library in an inconsistent state, so the user should be warned (since either PDF or .bib file might be missing)

• URL handler should be split into a download_pdf and a download_bib, so we can call them separately in addbib/addpdf
  • this is useful when PDFs are paywalled, but we still want the .bib

Add support for adding PDF from local file too.

Add support for downloading a webpage as a PDF and adding it.

Uncategorized

• ck tag/untag/genbib/copypdfs autocompletion of tags
• figure out how to have a setup.py that installs this thing
  • add a requirements.txt too
• if PDFs are not available from publisher, try sci-hub.tw: see python example here
• Cryptology ePrint updater: need it to update papers to their latest versions
  • ck should run this once a day
  • should move old paper to CK<year>.<ckdateadded>.pdf (make sure no naming conflicts)
• tools for making the .bib files consistent
  • titles should have double brackets
  • same conference shouldn't have different names
  • author names should always be separated by ' and '
  • similar or incomplete author names
• ck command
  • list subcommand
    • display associated file info
      • nopdf
      • nobib
      • md
      • any other files (e.g., ABC19.slides.pdf)
    • list all conferences across papers
    • when invoked without recursion, tell user what other subtags are there in the current subdir?
  • open subcommand
    • add support for various associated files: .bib, .html, .md, .notes.[0-9]*.pdf, .slides.pdf, .etc
    • if you type in a partial citation key, should list all matches
      • if just one match, should just open it, displaying a warning that it only partially matched so as to not train you to use the wrong CK
    • if you type an ambiguous citation key (lowercase / uppercase), maybe you should be prompted for what to open, because there won't be many matches.
  • untag subcommand
    • right now, if we only have a .bib file without any PDFs, untag will not detect any untagged papers. how to handle?