The Biblio system consist of a few perl scripts plus one simple
convenient Bash-script for accessing PubMed from the command
line. They allow for searching and viewing hits in the Medline
database, and most importantly, also convert records to BibTeX. There
are at least one other shell tool for doing these things but, hey, it
was Not Invented Here! And besides, I like my solution and have
further plans for it.
The scripts are found in ~arve/bin. Here is a quick summary:
pmsearch
-- Search Medline and have article identifiers (PMID) returned.
pmid2text
-- Take PMIDs (on the command line or from STDIN) and retrieve the
entries as text in a couple of formats, including partial HTML
for building web pages with.
pmid2bibtex
-- Works like pmid2text, but writes bibtex records to STDOUT.
pubmed
-- This is a simple script that essentially executes
"pmsearch search terms | pmid2text -a".
pmid2seq
-- Download sequences from PubMed that are refereced, linked, to an
article.
I have appended the usage texts below. They are also available by
using the "-u" flag. An important and practical feature (at least to
me) is that Medline records are cached on your local computer (in /tmp
actually), making repeated viewing very fast by reducing calls to
PubMed.
Some examples:
How many papers were published by von Heijne last year?
>pmsearch -c year=2002 author=von+Heijne
9
List them:
>pubmed year=2002 author=von+Heijne
11976343 Vilar M et al (2002) "Insertion and topology of a plant viral
movement protein in the endoplasmic reticulum membrane."
J Biol Chem 277(26), 23447--23445
12441395 Nilsson J, Persson B, Von Heijne G (2002) "Prediction of
partial membrane protein topologies using a consensus approach."
Protein Sci 11(12), 2974--2978
[snip, 7 more]
I am open to suggestions that would make theses tools more useful for you.
Cheers,
Lasse
---------------------------------------------------------
Usage: pmid2text [<options>] [<pubmed identifier>+]
Options:
-u This text.
-i No indentation of citations.
-y Sort by publication year.
-a Show abstract.
-n <int> Max number of authors without using "et al". Set to
zero to show all authors.
-h Output as HTML for inclusion on a web page.
-q Quiet. No extraneous output.
If no PMID is given on the command line, STDIN is read until EOF
and parsed for PMID:s. The expected format is a PMID first on each
line, possibly followed by a space and any string.
----------------------------------------------------------
Usage: pmid2bibtex [<option>] [<pubmed identifier>+]
Options:
-u This text.
-y Sort by year.
If no PMID is given on the command line, STDIN is read until EOF
and parsed for PMID:s. The expected format is a PMID first on each
line, possibly followed by a space and any string.
----------------------------------------------------------
Usage: pmsearch [<options>] <search terms>+
OPTIONS
-u This text.
-c Return the number of matching items.
-d <int> The maximum number of PMID:s to report. Default: 100.
-t <int> The number of days from today to include in search.
-w <int> Warn when more than <int> identifiers are returned. Default: 100.
SEARCH TERMS
A search terms is either a lone word that can match anything in a PubMed entry.
The complete query is a conjunction of its terms.
You can for example write 'pmsearch eisen yeast' and get back (at this writing) 31
PubMed identifiers. You may also write these search terms using the PubMed qualifiers
to specify to what record in an entry a term should match. For example, writing
'pmsearch eisen[AU] yeast specifies that 'eisen' is an author. However, since the
PubMed way of writing things is hard to remember (for me at least), a second way of
qualifying terms is implemented. The general syntax in this case is:
<qualifier>=<term>
where <qualifier> is one of
author
abstract
journal
title
year
The previous example is then written as 'pmsearch author=eisen yeast'.
The journal titles can be hard to get right. Use either their full names (e.g.
'Molecular Biology and Evolution') or the official abbreviation ('Mol Biol Evol').
See the PubMed WWW site for details. In addition to the official abbreviations,
pmsearch recognizes some additional common acronyms, which is translated to a name
recognized by PubMed. Currently implemented are:
Key Journal
psb = Pac Symp Biocomput
jmb = J Mol Biol
ismb = Proc Int Conf Intell Syst Mol Biol
jme = J Mol Evol
nar = Nucleic Acids Res
mbe = Mol Biol Evol
embo = EMBO J
pnas = Proc Natl Acad Sci U S A
cabios = Comput Appl Biosci
jcb = J Comput Biol