Importing an External non-OJS Journal

[thiessenp]

Hi,

I'm new to OJS and still ramping up. I've read through the online documentation. I'm very impressed with the amount of detail in the OJS docs for an OS project. I'll be developing plugins etc. for the next couple of years for my University. My first task is to import an external Journal thats housed on a completely different system. I noticed in the Support discussion a similar post and the response to create a mapping tool which in the case of the post was from Dublin Core to the OJS format.

Would anyone have any examples or could share any code for mapping an external format to the OJS format? (also any tips, hard learned lessons would be really appreciated)

Finally, as a new developer to OJS I'd like to get developing as quickly as possible. Beyond the OJS Technical Document (OJSTechnicalReference.pdf) and OJS in an Hour (OJSinanHour.pdf) would anyone have any other reading to suggest?

Thanks in advance,
-peter

[asmecher]

Hi Peter,

Thanks for the feedback!

I tend to hack together custom scripts to get data into OJS's import format, since data typically comes from a number of sources, e.g. CSV files, flat-files, etc, and a single-format text processing solution often isn't flexible enough. If you've got data in an XML format, I'd suggest using XSL to transform the data; I've done this successfully to convert OJS 1.x import data into OJS 2.x's import format. (See http://pkp.sfu.ca/cvs/cvsweb.cgi/ojs2/plugins/importexport/native/ojs1-to-ojs2.xsl for this file's CVS entry.)

Regards,
Alec Smecher
Public Knowledge Project Team

[thiessenp]

Hi Alec,

thanks for the reply. The XSLT example is quite helpful. I'm sure I'll have questions in the near future

Thanks again
-peter

[thiessenp]

Hi Alec,

forgot to ask, could I see an example of one of your hacked together custom scripts? Would you have any that convert CVS to OJS XML?

Thanks,
-peter

[asmecher]

Hi Peter,

Unfortunately, these are far too task-specific and full of specific cases to be of any general use, but as a general pattern, they usually look something like:

Code: Select all

<?php

// Escape a string that will be used in XML.
function fix($string) {
    return htmlspecialchars($string, ENT_NOQUOTES);
}

// Load article data for the export
function loadArticles ($filename) {
    // Add code here to load the input file into a structure that looks like:
    $data = array(
        array(
            'title' => 'First Article Title',
            'abstract' => 'First Article Abstract',
            'authors' => array(
                'firstName' => 'Joe',
                'lastName' => 'Public'
            )
        ),
        array(
            'title' => 'Second Article Title',
            'abstract' => 'Second Article Abstract',
            'authors' => array(
                'firstName' => 'Jane',
                'lastName' => 'Doe'
            )
        )
    );

    return $data;
}

$articles = loadArticles('input.csv');

$issueTitle = 'Issue Title';
$volume = 1;
$number = 1;
$year = 1;

$out = fopen('output.xml', 'w');
if (!$out) die ('Cannot open output file.');

fwrite($out, "<?xml version=\"1.0\" ?>\n");
fwrite($out, "<issue published=\"true\">\n");
fwrite($out, "\t<title>" . fix($issueTitle) . "</title>\n");
fwrite($out, "\t<volume>" . $volume  . "</volume>\n");
fwrite($out, "\t<number>" . $number . "</number>\n");
fwrite($out, "\t<year>" . $year . "</year>\n");

// For this sample script, put all articles in the same section
fwrite($out, "\t<section>\n");
fwrite($out, "\t<section>\n");
fwrite ($out, "\t\t<title>Articles</title>\n");

foreach ($articles as $article) {
    fwrite ($out, "\t\t<article>\n");
    fwrite ($out, "\t\t\t<title>" . fix($article['title']) . "</title>\n");
    fwrite ($out, "\t\t\t<abstract>" . fix($article['abstract']) . "</abstract>\n");

    foreach ($article['authors'] as $author) {
        // ...etc...
    }

    // ...etc...

    fwrite ($out, "\t\t</article>\n");
}

fwrite($out, "\t</section>\n");
fwrite($out, "</issue>\n");

fclose($out);

?>

If you want to get data from CSV (I assume you mean CSV, not CVS), you can use the fgetcsv function e.g. in your loadArticles function.

Regards,
Alec Smecher
Public Knowledge Project Team

[thiessenp]

very helpful - thanks!

-peter

[Owen]

What about using phpMyAdmin to import?

(I think I asked this as part of another thread but didn't get an answer).

[asmecher]

Hi Owen,

phpMyAdmin will only import batches of SQL statements or, possibly, something like CSV-formatted data that corresponds exactly to a table layout; I suspect none of these will be helpful for bringing in back-issues unless e.g. you're moving an OJS installation from one server to another.

Rergards,
Alec Smecher
Public Knowledge Project Team

[Owen]

Keeping to the existing schema for a particular table would be no problem. It's just whether you've got to fiddle with other tables in the process: is there only 1 table involved in journal archive imports?

[asmecher]

Hi Owen,

Doing an issue import will probably affect 5-10 tables when all is said and done.

Regards,
Alec Smecher
Public Knowledge Project Team