A Proquest platform API

We subscribe to a number of databases via Proquest.

I wanted an API for having my software execute fielded searches against a Proquest a database — specifically Dissertations and Theses in my current use ase — and get back structured machine-interpretable results.

I had vaguely remembered hearing about such an API, but was having trouble finding any info about it.

It turns out, while you’ll have trouble finding any documentation about it, or even any evidence it exists on the web, and you’ll have trouble getting information about it from Proquest support too — such an api does exist.  Hooray.

You may occasionally see it called the “XML Gateway” in some Proquest documentation materials (although Proquest support doesn’t neccesarily know this term). And it was probably intended for and used by federated search products — which makes me realize, oh yeah, if I have any database that’s used by a federated search product, then it’s probably got some kind of API.

And it’s an SRU endpoint.

(Proquest may also support z39.50, but at least some Proquest docs suggest they recommend you transition to the “XML Gateway” instead of z39.50, and I personally find it easier to work with then z39.50).

Here’s an example query:

http://fedsearch.proquest.com/search/sru/pqdtft?operation=searchRetrieve&version=1.2&maximumRecords=30&startRecord=1&query=title%3D%22global%20warming%22%20AND%20author%3DCastet

For me, coming from an IP address recognized as ‘on campus’ for our general Proquest access, no additional authentication is required to use this API. I’m not sure if we at some point prior had them activate the “XML Gateway” for us, likely for a federated search product, or if it’s just this way for everyone.

The path component after “/sru”, “pqdtft” is the database code for Proquest Dissertations and Theses. I’m not sure where you find a list of these database codes in general; if you’ve made a succesful API request to that endpoint, there will be a <diagnosticMessage> element near the end of the response listing all database codes you have access to (but without corresponding full English names, you kind of have to guess).

The value of the ‘query’ parameter is a valid CQL query, as usual for SRU. It can be a bit tricky figuring out how to express what you want in CQL, but the CQL standard docs are decent, if you spend a bit of time with them to learn CQL.

Unfortunately, there seems to be no SRU “explain” response available from Proquest to tell you what fields/operators are available. But guessing often works, “title”, “author”, and “date” are all available — I’m not sure exactly how ‘date’ works, need to experiment more — although doing things like `date > 1990 AND date <= 2010` appears initially to work.

The CQL query param above un-escaped is:

title="global warming" AND author=Castet

Responses seem to be in MARCXML, and that seems to be the only option.

It looks like you can tell if a full text is available (on Proquest platform) for a given item, based on whether there’s an 856 field with second indicator set to “0” — that will be a URL to full text. I think. It looks like.

Did I mention if there are docs for any of this, I don’t have them?

So, there you go, a Proquest search API!

I also posted this to the code4lib listserv, and got some more useful details and hints from Andrew Anderson.

Oh, and if you want to link to a document you found this way, one way that seems to work is to take the Proquest document ID from the marc 001 field in the response, and construct a URL like `http://search.proquest.com/pqdtft/docview/$DOCID$`.  Seems to work. Linking to fulltext if it’s available otherwise a citation page.  Note the `pqdtft` code in the URL, again meaning ‘Proquest Dissertations and Theses’ — the same db I was searching to find to the doc id.

About these ads
This entry was posted in General. Bookmark the permalink.

One Response to A Proquest platform API

  1. jrochkind says:

    Specifically asking Proquest support for the “Federated-Search.docx” document, as suggested by Andrew Anderson, did get me that document.

    It mostly doesn’t tell us anything we hadn’t already figured out above, but has a few useful expansions and new details.

    One of them is that apparently there are special database codes for searching certain pre-selected groupings of multiple databases at once, like “all_subscribed”, and “subject_arts”. I don’t need that for my present use case, but it is potentially crucial for others; I did try “all_subscribed” and it appeared to work. Neat!

    It also includes the list of search fields it says are supported. And confirms that included MARC 856’s will have their second indicator set appropriately, such that “0” is always full text, and “1” never is.

    Probably worth asking for this document if you plan to develop against the SRU api.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s