Re: [Pybliographer-general] new dev branch: zope & pubmed?

[Daniel C. <dan...@ya...>]

I asked:
> In the meantime have you considered adding a live query to the
> Pubmed/Entrez query engine?  You can get probably get medline results
back
> to read into Pyblio without too much work, as described here:
>
>   http://www.ncbi.nlm.nih.gov/entrez/query/static/linking.html

Frederic responded:
  Have you seen a way _not_ to get an HTML output for a query ? Because a
  module that doesn't rely on the HTML formatting would be much more
  robust...

According to the docs (I searched 'em up and down too :), they want you to
get the html first, which seems wasteful especially because you can get
the citation format if you want within the html.  _But_ they still accept
the old format (old as of 3-6 months ago) queries and will return just
citations, as described here:

  http://www.ncbi.nlm.nih.gov/entrez/utils/qmap_help.html

I poked through the source from our old DubMed project and found the old
format:

http://www.ncbi.nlm.nih.gov/entrez/utils/qmap.cgi?db=m&form=4&term=cancer[text]&dispmax=3&html=no&dopt=l

Some of the values we can tweak in this old format include:

html=no      the magic bit :)

form=4  AND  term=someterm[somefield]
  does a normal search; the term value can grow to handle lots of fields
  and boolean logic.

form=6  AND  uid=someuid  AND dopt=m
  looks for medline 'neighbors' (aka related articles) for the given
  article uid

dopt=q
  returns only a hit count for your given search.


According to Matt's code (in java, if you're interested see
http://oldap.med.yale.edu/cgi-bin/cvsweb.cgi/dubmed/URLBuilder.java?rev=1.2&content-type=text/x-cvsweb-markup
) the dopt values can be combined, aka, dopt=mq => neighbors, count only.

I have no idea if it's worth implementing this old style query.  They
don't seem to mind but they could theoretically always turn off the
converter.

If you are interested in trying it out the new way, attached is some
skeleton code for getting back medline cites from a keyword term search
taken from argv.  This is actually my first bit of python code... I'm
almost done with _Learning Python_ and I'm very impressed with how
straightforward this language is. :)  Anyway my apologies in advance if it
seems poorly organized.


  -Dan