Programmatically create Biblio entry from Pubmed PMID

Actually, with a tiny bit of restructuring of the functions "biblio_pm_biblio_import()" and "_biblio_pm_import()" this could be done quite easily. We would just need to move the file reading code from "_biblio_pm_import() to biblio_pm_biblio_import() and pass a list of pmid's to _biblio_pm_import() rather than the file name. Then you could directly call _biblio_pm_import() with your (comma separated) list of PMID's obtained from else where and they would be saved a biblio nodes.

Here are the modifications I would propose...

function biblio_pm_biblio_import($file, $terms = array(), $batch = FALSE, $session_id = NULL, $save = TRUE, $string = FALSE) {
  $nids = array();
  $dups = array();
  $pmids = file($file->uri, FILE_IGNORE_NEW_LINES | FILE_SKIP_EMPTY_LINES);
  if (empty($pmids)) {
    drupal_set_message(t("Could not open PubMed ID file"), 'error');
    return;
  }

  list($nids, $dups) = _biblio_pm_import( $pmids, $terms, $batch, $session_id);

  return array($nids, $dups);
}

function _biblio_pm_import( $pmids, $terms, $batch, $session_id) {
  module_load_include('php', 'biblio_pm', 'EntrezClient');
  module_load_include('php', 'biblio_pm', 'EntrezPubmedArticle');
  $retmax = 100;
  $resmax = count($pmids);
  $start = 0;
  $Eclient = new BiblioEntrezClient;
  $Eclient->post($pmids);
  $Eclient->setReturnMax($retmax);
....

I thought it made sense to restructure the code in as described, so I went ahead and pushed the changes as described above. The one exception being that I renamed the _biblio_pm_import function to biblio_pm_import_ids, and it takes an array of PMID's as the first argument.

I'm assuming you are calling "biblio_pm_import_ids()" with an array of PMID's?

The duplicate detection is done in "biblio_pm_import_ids()" prior to saving the node and it is only saved if it's not a duplicate. What is the problem you are seeing?

The way I deal with setting the author UID can be found in the biblio_import() function in the biblio.import.export.inc file. Using the "nids" returned by the import function I just set them all in one go, because even if you set the uid on the node before it is saved it gets reset by the node module to whichever uid is saving the node. Here is the relevant snipet...

  $user = $results['user'];
  $userid = $results['userid'];

  if (user_access('administer biblio') && count($nids) && $user->uid != $userid) {

    db_update('node')
      ->fields(array('uid' => $results['userid']))
      ->condition('nid', $nids, 'IN')
      ->execute();

    db_update('node_revision')
      ->fields(array('uid' => $results['userid']))
      ->condition('nid', $nids, 'IN')
      ->execute();
  }

WRT that modified module you are proposing, maybe that functionality can just be built into the existing biblio_pm.module.

Ron.

subscribe

What are the chances that this function/feature will be incorporated in upcoming builds in the near future? I am supporting an academic website which relies on the PubMed Integration module's functions (with some custom coding) to pull back and store the PubMed Data. However, since the PubMed Integration module does not exist on D7 and we are planning to upgrade to D7 in the fall, this would be a highly desirable feature.

In addition, the PubMed Integration Module's developers are noting (on this post: http://drupal.org/node/1128878) that this feature may eliminate the need for their module altogether, so it would be good to know when/if this is in the works so we can plan accordingly.

Thanks.

@sbrow126 the pubmed integration of the main biblio module for Drupal 7 works pretty well. I use it extensively on two D7 sites (with a bit of custom code). The directions in this issues are really good and it means I'll be able to get rid of some of my custom code.

@sbrow126 (#13) Sorry, maybe I'm a bit thick, but which features of the pubmed integration module are you referring to? If it's the "function biblio_pm_import_ids" bit, that's is already in the biblio_pm module.

Thank you. My concern was being able to submit a single PMID that the user would input and retrieve all the necessary Biblio node data from just that. The issue I was having was that we used the PubMed Integration module as a means of facilitating this originally. However, given that there was no updates to the Integration since 2009, my concern was how to get that functionality through D7.

I had to add some custom code to the module itself to get it to work correctly (there was a $global user callin missing so the original function pubmed_integration_import_data and pubmed_integration_load_citation_by_pmid was actually not working properly). There were a few other custom tweaks I had to make to get the PubMed link on a citation to work as well.

The main issue is that I still needed a single PMID to be passed and retrieve a single entry, and as a result the batch process that biblio now uses for the import would still need to be modified, hence my original request. That said, if the current best approach is to modify biblio_pm_import_ids on D7, then that's a simpler solution that upgrading the PubMed Integration module.

The site I am working on allows user to input a single PMID through another module and then using that entry, i created a custom function which would call the citation import process and create a biblio node/citation from it (after validation checks, etc.). So it sounds like on D7 biblio_pm_import_id is the way to go?

Just to be clear, you want to (programmically) pass a PMID and get a node object returned? You don't want the node saved to the database at this point?

This would still be possible, but I would have to write a little helper function to do it.

I've added a helper function to the module which should achieve what you want

You can now do this...

  $node = biblio_pm_fetch_pmid($pmid)

and it will return a "biblio" node object or an empty node object if the pmid is not found.

http://drupalcode.org/project/biblio.git/commit/b9a3788

I've added a helper function to the module which should achieve what you want

You can now do this...

  $node = biblio_pm_fetch_pmid($pmid)

and it will return a "biblio" node object or an empty node object if the pmid is not found.

http://drupalcode.org/project/biblio.git/commit/b9a3788

Just on this you had one or two brackets missing

$pmids = array();
include_once "sites/all/modules/biblio/includes/biblio.import.export.inc";
$esearch_url = "http://eutils.ncbi.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&retmax=99&reldate=30&term=pulmonary+hypertension";
$esearch_load = simplexml_load_file($esearch_url);
if (!is_object($esearch_load)) {
  watchdog("Failed to get retrieve Pubmed entries");
}
else {
  foreach($esearch_load->xpath("//Id") as $pmid) {
    $pmids[]= (string)$pmid;
  }
}
$nids = biblio_pm_import_ids($pmids, array(), FALSE,NULL);
if(isset($nids[0][0])) {

  db_update('node')
    ->fields(array('uid' => 1))
    ->condition('nid', $nids[0], 'IN')
    ->execute();
  db_update('node_revision')
    ->fields(array('uid' => 1))
    ->condition('nid', $nids[0], 'IN')
    ->execute();
    }

Further on this: a lot of hosting companies quite sensibly prevent execute(); being stored on a database, so the RUles solution has limited applicability. Module I guess.

If you need further help, please re-open and provide details.

Would it be also possible to programmatically fetch a node from the DOI of a publication? Similar to

biblio_pm_fetch_pmid($pmid), something like

biblio_doi_fetch_doi($doi)?

It would be a new issue to add such a function.

May I add a new issue request?

Any user can use the "Create a new issue" link on the module issues page. Be aware that this module gets little maintenance, mostly bug fixes. Of course, you could write the patch yourself.