more documentation

[org-ref.git] / org-ref.org

#+TITLE: The org-ref manual
#+AUTHOR: John Kitchin
#+DATE: 2015-03-15 Sun
#+OPTIONS: toc:nil ^:{}
#+LATEX_HEADER: \usepackage{natbib}
\maketitle
\tableofcontents

* Introduction to org-ref
Org-ref is a library for org-mode cite:Dominik201408 that provides rich support for citations, labels and cross-references in org-mode. org-ref is especially suitable for org-mode documents destined for LaTeX export and scientific publication. org-ref is also extremely useful for research documents and notes. org-ref bundles several other libraries that provide functions to create and modify bibtex entries from a variety of sources, but most notably from a DOI.

The basic idea of org-ref is that it defines a convenient interface to insert citations from a reference database (e.g. from a bibtex file(s)), and a  set of functional org-mode links for citations, cross-references and labels that export properly to LaTeX, and that provide clickable functionality to the user. org-ref interfaces with helm-bibtex to facilitate citation entry, and it can also use reftex.

org-ref provides a fairly large number of utilities for finding bad citations, extracting bibtex entries from citations in an org-file, and functions for interacting with bibtex entries. We find these utilities indispensable in scientific writing.

org-ref is [[id:32B558A3-7B48-4581-982B-082017B0AEE8][customizable]].

org-ref has been in development since sometime in 2013. It was written to aid in the preparation of scientific manuscripts.  We have published a lot of scientific articles with it so far  cite:hallenbeck-2013-effec-o2,mehta-2014-ident-poten,xu-2014-relat,xu-2014-probin-cover,miller-2014-simul-temper,curnan-2014-effec-concen,boes-2015-estim-bulk,xu-2015-linear-respon,xu-2015-relat. Be sure to check out the supporting information files for each of these. The org source for the supporting information is usually embedded in the supporting information.

There has been a lot of recent discussion on the org-mode mailing list about a citation syntax for org-mode. It is not clear what the impact of this on org-ref will be. The new syntax will more cleanly support pre/post text, and it will be separate from the links used in org-ref. It is not clear if the new syntax will support all of the citation types, especially those in biblatex, that are supported in org-ref. The new citations are likely to be clickable, and could share functionality with org-ref. The new citation syntax will not cover labels and cross-references though, so these links from org-ref will still exist. We anticipate a long life for org-ref.

** Basic usage of org-ref

*** Bibliography links
org-ref provides a bibliography link to specify which bibtex files to use in the document. You should use the filename with extension, and separate filenames by commas. This link is clickable; clicking on a filename will open the file. Also, if your cursor is on a filename, you will see a minibuffer message about whether the file exists or not. On export, the bibliography will appear at the position where the link is defined. Usually this link goes near the end of your document, e.g. like [[bibliography link][here]].

There is also a nobibliography link, which is useful for specifying a bibliography file, but not listing a bibliography in the exported document.

There is also a bibliographystyle link that specifies the style. This link does nothing but export to a LaTeX command.

If you use biblatex, then you use an addbibresource link. biblatex support is not very well tested by me, because we do not use it.

*** Citations
    :PROPERTIES:
    :CUSTOM_ID: citations
    :END:
org-ref uses the [[bibliography link]] to determine which bibtex files to get citations from, and falls back to the bibtex files defined in the variable  reftex-default-bibliography or org-ref-default-bibliography if no bibliography link is found.

For simple citation needs, org-ref is simple to use. At the point you want to insert a citation, you select the "Org -> org-ref -> Insert citation" menu (or use the key-binding C-c ] by default), select the reference(s) you want in the helm-bibtex buffer and press enter. The citation will be inserted automatically into your org-file. You "select" an entry by using the arrow keys to move up and down to the entry you want. You can narrow the selection by typing a pattern to match, e.g. author name, title words, year, keyword, etc... You can select multiple entries by pressing C-spc to mark enties in the helm-bibtex buffer.

If the cursor is on a citation key, you should see a message in the minibuffer that summarizes which citation it refers to. If you click on a key, you should see a helm selection buffer with some actions to choose, including opening the bibtex entry, opening/getting a pdf for the entry, searching the entry in Web of Science, etc...

The default citation type is [[id:32B558A3-7B48-4581-982B-082017B0AEE8][customizable]], and set to "cite". If you want another type of citation type, then type C-u before pressing enter in the helm-bibtex selection buffer. You will be prompted for the type of citation you actually want.

Here is a list of supported citation types. You can customize this if you want. If you do not know what all these types are, you probably do not need them. The default cite is what you need. See http://ctan.unixbrain.com/macros/latex/contrib/natbib/natnotes.pdf
 for the cite commands supported in bibtex, and http://ctan.mirrorcatalogs.com/macros/latex/contrib/biblatex/doc/biblatex.pdf
 for the commands supported in biblatex. For most scientific journals, only bibtex is supported.

#+BEGIN_SRC emacs-lisp
org-ref-cite-types
#+END_SRC

#+RESULTS:
| cite | nocite | citet | citet* | citep | citep* | citealt | citealt* | citealp | citealp* | citenum | citetext | citeauthor | citeauthor* | citeyear | citeyear* | Citet | Citep | Citealt | Citealp | Citeauthor | Cite | parencite | Parencite | footcite | footcitetext | textcite | Textcite | smartcite | Smartcite | cite* | parencite* | supercite | autocite | Autocite | autocite* | Autocite* | Citeauthor* | citetitle | citetitle* | citedate | citedate* | citeurl | fullcite | footfullcite | notecite | Notecite | pnotecite | Pnotecite | fnotecite | cites | Cites | parencites | Parencites | footcites | footcitetexts | smartcites | Smartcites | textcites | Textcites | supercites | autocites | Autocites | bibentry |

If the cursor is on a citation, or at the end of the citation, and you add another citation, it will be appended to the current citation.

If you want to /replace/ an existing key in a citation, put the cursor on the key, run the insert citation command, and type C-u C-u before pressing enter in the helm-bibtex selection buffer. The key will be replaced. Of course, you can just delete it yourself, and add a new key.

Finally, if you do not like the order of the keys in a citation, you can put your cursor on a key and use shift-arrows (left or right) to move the key around. Alternatively, you can run the command org-ref-sort-citation-link which will sort the keys by year, oldest to newest.

org-ref has basic support for pre/post text in citations. We have very little need for this in scientific publishing; we write pre text before the citation, and post text after it. However, you can get pre/post text by using a description in a cite link, with pre/post text separated by ::. For example, [[cite:Dominik201408][See page 20::, for example]]. It is a little awkward to see this because you cannot see the key. It is a low priority to find a fix for this, because it is not common in scientific publishing and you can always fall back to the old-fashioned LaTeX: \cite[See page 20][, for example]{Dominik201408}.

You may want to bind a hydra menu to a key-binding or key-chord. For example:

#+BEGIN_SRC emacs-lisp
(key-chord-define-global "kk" 'org-ref-cite-hydra/body)
#+END_SRC

This will allow you to quickly press kk while on a cite link to access functions that can act on the link.

*** label links
LaTeX uses labels to define places you can refer to. These can be labels in the captions of figures and tables, or labels in sections. We illustrate some uses here.

label links are "functional" if you put your cursor on the link, you will get a message in the minibuffer showing you the number of occurrences of that label in the buffer. That number should be one! It is most preferable to put a label link into a caption like this.
#+caption: Another simple table. label:tab-ydata
| y |
| 4 |
| 5 |

org-ref can help you insert unique labels with the command elisp:org-ref-helm-insert-label-link. This will show you the existing labels, and insert your new label as a link. There is no default key-binding for this.

*** ref links
A ref link refers to a label of some sort. For example, you can refer to a table name, e.g. Table ref:table-1.

#+tblname: table-1
#+caption: A simple table.
| x |
| 1 |
| 2 |

Or you can refer to an org-mode label as in Table ref:table-3.

#+label: table-3
#+caption: A simple table.
| x |
| 1 |
| 2 |

You can also refer to an org-ref label link as in Table ref:tab-ydata.

To help you insert ref links, use the "Org->org-ref->Insert ref" menu, or run the command elisp:org-ref-helm-insert-ref-link. There is no default key-binding for this.

ref links are functional. If you put the cursor on a ref link, you will get a little message in the minibuffer with some context of the corresponding label. If you click on the ref link, the cursor will jump to the label.

A brief note about references to a section. This only works if you put a label in the org-mode headline. Otherwise, you must use a CUSTOM_ID and a CUSTOM_ID link. For example section [[#citations]] has a CUSTOM_ID of citations. Section ref:sec-misc has a label link in the headline. That works, but is not too pretty.

**** Miscellaneous ref links  label:sec-misc
org-ref also provides a pageref, nameref and eqref link.

Note for eqref, you must use a LaTeX label like this:

\begin{equation}
e^x = 4 \label{eq:1}
\end{equation}

Then you can refer to Eq. eqref:eq:1 in your documents.

** org-ref customization of helm-bibtex
org-ref adds a few new features to helm-bibtex. First, we add keywords as a searchable field. Second, org-ref modifies the helm-bibtex search buffer to include the keywords. Since keywords now can have a central role in searching, we add some functionality to add keywords from the helm-bibtex buffer as a new action.

We change the order of the actions in helm-bibtex to suit our work flow, and add some new actions. We define a format function for org-mode that is compatible with the usage defined in section [[#citations]]. Finally, we add some new fallback options for additional scientific search engines.

** Some basic org-ref utilities
The command org-ref does a lot for you automatically. It will check the buffer for errors, e.g. multiply-defined labels, bad citations or ref links, and provide easy access to a few commands through a helm buffer.

org-ref-clean-bibtex-entry will sort the fields of a bibtex entry, clean it, and give it a bibtex key. This function does a lot of cleaning:

1. adds a comma if needed in the first line of the entry
2. makes sure the doi field is an actual doi, and not a url.
3. fixes bad year entries
4. fixes empty pages
5. Escapes & to \&
6. generate a key according to your setup
7. Runs your hook functions
8. sorts the fields in the entry
9. checks the buffer for non-ascii characters.

This function has a hook org-ref-clean-bibtex-entry-hook, which you can add functions to of your own. Each function must work on a bibtex entry at point.

#+BEGIN_SRC emacs-lisp
(add-hook 'org-ref-clean-bibtex-entry-hook 'jmax-replace-nonascii)
#+END_SRC

#+RESULTS:
| jmax-title-case-article | jmax-replace-nonascii |

org-ref-extract-bibtex-entries will create a bibtex file from the citations in the current buffer.

** LaTeX export
All org-ref links are designed to export to the corresponding LaTeX commands for citations, labels, refs and the bibliography/bibliography style. Once you have the LaTeX file, you have to build it, using the appropriate latex and bibtex commands. You can have org-mode do this for you with a setup like:

#+BEGIN_SRC emacs-lisp
(setq org-latex-pdf-process
      '("pdflatex -interaction nonstopmode -output-directory %o %f"
        "bibtex %b"
        "pdflatex -interaction nonstopmode -output-directory %o %f"
        "pdflatex -interaction nonstopmode -output-directory %o %f")
#+END_SRC

** Other exports
There is some basic support for HTML and ascii export. Not all bibtex entry types are supported, but basic support exists for articles and books. For a markdown export, the cite links are exported as Pandoc style links.

* Other libraries in org-ref
These are mostly functions for adding entries to bibtex files, modifying entries or for operating on bibtex files. Some new org-mode links are defined.

** doi-utils
This library adds two extremely useful tools for getting bibtex entries and pdf files of journal manuscripts. Add this to your emacs setup:
#+BEGIN_SRC emacs-lisp
(require 'doi-utils)
#+END_SRC

The provides two important commands:

- doi-utils-add-bibtex-entry-from-doi
This will prompt you for a DOI, and a bibtex file, and then try to get the bibtex entry, and pdf of the article.


- doi-utils-add-entry-from-crossref-query
This will prompt you for a query string, which is usually the title of an article, or a free-form text citation of an article. Then you will get a helm buffer of matching items, which you can choose from to insert a new bibtex entry into a bibtex file.

** jmax-bibtex
These are functions I use often in bibtex files.

*** Generate new bibtex files with adapted journal names
The variable jmax-bibtex-journal-abbreviations contains a mapping of a short string to a full journal title, and an abbreviated journal title. We can use these to create new versions of a bibtex file with full or abbreviated journal titles. You can add new strings like this:

#+BEGIN_SRC emacs-lisp
(add-to-list 'jmax-bibtex-journal-abbreviations
  '("JIR" "Journal of Irreproducible Research" "J. Irrep. Res."))
#+END_SRC

- jmax-bibtex-generate-longtitles :: Generate a bib file with long titles as
     defined in `jmax-bibtex-journal-abbreviations'
- jmax-bibtex-generate-shorttitles :: Generate a bib file with short titles as
     defined in `jmax-bibtex-journal-abbreviations'

*** Modifying bibtex entries

- jmax-stringify-journal-name :: replace a journal name with a string in
     `jmax-bibtex-journal-abbreviations'
- jmax-set-journal-string :: in a bibtex entry run this to replace the journal
     with a string selected interactively.

- jmax-title-case-article :: title case the title in an article entry.
- jmax-sentence-case-article :: sentence case the title in an article entry.

- jmax-replace-nonascii :: replace nonascii characters in a bibtex
     entry. Replacements are in `jmax-nonascii-latex-replacements'. This
     function is a hook function in org-ref-clean-bibtex-entry.

The non-ascii characters are looked up in a list of cons cells. You can add your own non-ascii replacements like this. Note backslashes must be escaped doubly, so one \ is \\\\ in the cons cell.

#+BEGIN_SRC emacs-lisp
(add-to-list 'jmax-nonascii-latex-replacements
  '("æ" . "{\\\\ae}"))
#+END_SRC

These functions are compatible with bibtex-map-entries, so it is possible to conveniently apply them to all the entries in a file like this:

#+BEGIN_SRC emacs-lisp
(bibtex-map-entries 'jmax-title-case-article)
#+END_SRC


*** Bibtex entry navigation
- jmax-bibtex-next-entry :: bound to M-n
- jmax-bibtex-previous-entry :: bound to M-p

*** Hydra menus for bibtex entries and files
- Functions to act on a bibtex entry or file
  - jmax-bibtex-hydra/body :: gives a hydra menu to a lot of useful functions
       like opening the pdf, or the entry in a browser, or searching in a
       variety of scientific search engines.
  - jmax-bibtex-new-entry/body :: gives a hydra menu to add new bibtex entries.
  - jmax-bibtex-file/body :: gives a hydra menu of actions for the bibtex file.

You will want to bind the hydra menus to a key. You only need to bind the first one, as the second and third can be accessed from the first hydra.
You can do that like this before you require jmax-bibtex:

#+BEGIN_SRC emacs-lisp
(setq jmax-bibtex-hydra-key-binding "\C-cj")
#+END_SRC

Or this if you like key-chord:

#+BEGIN_SRC emacs-lisp
(key-chord-define-global "jj" 'jmax-bibtex-hydra/body)
#+END_SRC

** isbn
#+BEGIN_SRC emacs-lisp
(require 'isbn)
#+END_SRC

This library provides some functions to get bibtex entries for books from their ISBN.

- isbn-to-bibtex

** pubmed
#+BEGIN_SRC emacs-lisp
(require 'pubmed)
#+END_SRC

This library provides a number of new org-mode links to Pubmed entries. See http://www.ncbi.nlm.nih.gov/pmc/about/public-access-info/#p3 for details of these identifiers.

pmcid:PMC3498956

pmid:23162369

nihmsid:NIHMS395714

Also, you can retrieve a bibtex entry for a PMID with

- pubmed-insert-bibtex-from-pmid

** arxiv
#+BEGIN_SRC emacs-lisp
(require 'arxiv)
#+END_SRC

This library provides an org-mode link to http://arxiv.org entries:  arxiv:cond-mat/0410285, and a function to get a bibtex entry and pdfs for arxiv entries:

- arxiv-add-bibtex-entry
- arxiv-get-pdf

** sci-id
#+BEGIN_SRC emacs-lisp
(require 'sci-id)
#+END_SRC

This package just defines two new org-mode links for http://www.orcid.org, and http://www.researcherid.com. Here are two examples:

orcid:0000-0003-2625-9232

researcherid:A-2363-2010

* Appendix
** Customizing org-ref
   :PROPERTIES:
   :ID:       32B558A3-7B48-4581-982B-082017B0AEE8
   :END:
You will probably want to customize a few variables before using org-ref extensively. One way to do this is through the Emacs customization interface: [[elisp:(customize-group "org-ref")]].

Here is my minimal setup:
#+BEGIN_SRC emacs-lisp
(setq reftex-default-bibliography '("~/Dropbox/bibliography/references.bib"))

(setq org-ref-bibliography-notes "~/Dropbox/bibliography/notes.org"
      org-ref-default-bibliography '("~/Dropbox/bibliography/references.bib")
      org-ref-pdf-directory "~/Dropbox/bibliography/bibtex-pdfs/")
#+END_SRC

For jmax-bibtex I like:

#+BEGIN_SRC emacs-lisp
(setq jmax-bibtex-hydra-key-chord "jj")
#+END_SRC

** Other things org-ref supports
*** org-completion
Most org-ref links support org-mode completion. You can type C-c C-l to insert a link. You will get completion of the link type, type some characters and press tab. When you select the type, press tab to see the completion options. This works for the following link types:

- bibliography
- bibliographystyle
- all cite types
- ref

*** Storing org-links
    :PROPERTIES:
    :ID:       AD9663C7-1369-413F-842A-157916D4BB75
    :CUSTOM_ID: sec-store-links
    :END:
If you are on a label link, or on a table name, or on an org-mode label you can "store" a link to it by typing C-c l. Then you can insert the corresponding ref link with C-c C-l.


* References
bibliographystyle:unsrtnat
# <<bibliography link>>
bibliography:org-ref.bib