|
Basic implementation of EBSCO module (only basic search, no facets and so on).
|
|
I'm going to test this and perhaps add more stuff.
|
Same here, once I have the time to do it (test it, not enhance it).
Also be trying to look into Scopus API | SciVerse and do test drive of it (http://www.info.sciverse.com/scopus/scopus-in-detail/tools/api).
Combined with EBSCO's and Open Access records (only “driver” compliant ones, no restricted access ones.. sorry!) already harvested do make around 99% of our online resources needs.
Looking really, really good!
Filipe
|
|
I already added some settings etc, but noticed that there is no sys/EBSCO.php. Missing or not needed?
|
|
It appears that EBSCO.php is not needed -- the API access happens directly inside the SearchObject. However, as this becomes more complex, it may make sense to abstract some of that logic to its own class.
|
|
Hugo Agud (hagud@orex.es) continues in developing the EBSCO Intergration parser.
|
Great!!
Looking forward to test those improvments.
Please do keeps us updated.
Filipe
|
|
PS: o the top (1st) of Popular Issues: nice to know that is of a common interest to some.
|
|
Missing /web/sys/EBSCO.php added
|
|
I forgot to mention that the patch is working in default theme.
|
Hi,
Being fully 110% supportive if this solution, hoping to help develop similar ones (like mentioned SCOPUS one) I do have a small request: so sorry, but can any of you share a clue how one can get some results out of this (subpath / URL to call, whatever)?
Untared the file, made the diff in the layout.tpl of the default theme, have analyzed all the files, yet, nothing. Is there any entry one has to include in conf.ini in order to activate it?
I mean I have this back in a 1.2>1.3 version (now I have ported almost all of local theming (CSS and alike) and dev into 1.3 release) ... doing risky things with a single 31GB ./solr dir across both versions, swaping symbolic links to the DATA/storage drive where VuFind instances are actually stored, not in /user/local/...
Meanwhile a print-screen of the output would be nice to have an idea / confirm...
Many thanks for all of the attention the request may be given,
Filipe
|
Subpath is vufind_root/EBSCO/Search, you can see example output here: https://vufind-trunk.mzk.cz/EBSCO/Search?lookfor=test&mylang=en
No configuration is needed.
|
|
Attached NDL_ebsco_v1.diff is modified from original vufind-ebsco.tar.gz with some cleanup and added basic support for facets and more configuration.
|
@Vaclav Rosecky: many thanks for you prompt reply.
Yes, I guess it would be smothing like that, but when tried that URL construct I got (and still do): "Cannot Load Action"... :|
But it is in a brand new 1.3 install so still connecting the dots (cannot use the upgrade tool as I have too many mixed ups from trunck and local).
EDIT: Sorry was trying on this "new server" (1.3), instead of an halfway 1.1 to 1.2 instance where I have instaled it. There I get:
Fatal error: Call to undefined method SearchObject_EBSCO::initLimit() in /usr/local/vufind/web/sys/SearchObject/EBSCO.php on line 101
I guess I have to do some troubleshooting!
Meanwhile going to try Ere Maijala diff over original vufind-ebsco.tar.gz files.
Thanks,
Filipe
|
|
I can't seem to get any of this to work and the different versions of attachments, along with the patch file that doesn't actually patch but creates files from scratch, is confusing. Can we get this repackaged or at least a working proper patch based on 1.3 posted?
|
Hi!
If not in mistake, Ere's NDL_ebsco_v1.diff, after restoring the affected files from previous changes, did work for me… although I remember it was a different process because it is GIT diff file, instead of the subversion's patch we are used to, as you mention (I have done that sometime ago, so I don't remember how I did applied it). If mine was not so heavily modified (CSS stuff, social media, linking to non-standard VuFind features, etc.), I would share all my files from /web/services/EBSCO/, web/interface/themes/<theme>/EBSCO/, web/interface/themes/<theme>/layout.tpl, web/conf/EBSCO.ini (for which you need the EIT access data given by EBSCO’s) and the core, web/sys/SearchObject/EBSCO.php, but I guess that would give you more troubles than solutions due to those changes.
Can you send where it’s failing (activating “debug = true” in config.ini) or haven’t you applied any modifications to your code?
I guess I could share here the full /web/services/EBSCO/ files, enhanced with some code of mine, if not in mistake and the SMARTY ones (the ones that I have really modified), you could use the standard. I’ll try to construct both (also these last from Ere's NDL_ebsco_v1.diff (since they are new, just create them, take the “+” in the beginning of the lines [replace "\n+" for "\n"] and should be ok).
FB
PS: will update this reply when have them done.
|
(Russ, please do skip my last message; no need to do anything I mentioned there: with the pack I am sending hereby, you shall have EBSCO real-time search module up & running in less than 30 seconds, hopefully; nevertheless, before deploying the files into your server, please have a quick look at the included NOTES)
Please find attached the necessary files (EBSCO_EIT_RealTimeSearch_Filipe_Bento_20120810.zip); in some of them I kept my enhanced code [the part that you can use, for instances, code that calls standard features or new files that I have also included (AJAXified some features > for instance to call web/interface/themes/blueprint/AJAX/ebscosbox.tpl). Some of the remaing code generates data not been "consumed" by the standard EBSCO/*.tpls files, but having a quick look at it, you can easily identify which data you want also to display, link it to new searches, etc.. Even if my server was already public, only with the demo account you would be able to see some of these enhancements that have in mine's, because make it public could breach the terms of the contract with EBSCO.
For instances, I’m controlling the access to it in this <div> (basic one: it hides this option if the user is not logged in):
<div id="logoutOptions"{if !$user} class="hide"{/if}>
<strong>{$profile.firstname|escape} {$profile.lastname|escape}</strong> <a class="account" href="{$path}/MyResearch/Home">{translate text="Your Account"}</a> <a class="realtimesearch" href="{$path}/EBSCO/Home" onClick="getLightbox('EBSCO', 'Home', null, null,'{translate text="Real Time search at EBSCO Databases"}'); return false;">{translate text="Search in EBSCOs Databases (FT)"}</a>
<a class="logout" href="{$path}/MyResearch/Logout">{translate text="Log Out"}</a>
<p/>
</div>
in web/interface/themes/blueprint/Search/home.tpl
Files reverted to default (using Ere's diff file), files in web/interface/themes/blueprint/EBSCO/ [because mines are modified to our needs / non-standard features]
- list.tpl;
- list-none.tpl; (when search retrieves no results)
- list-list.tpl (the one that shows the results' list; if you want to add more data retrieved from EBSCO [data extracted, from the retrieved XML, by web/sys/SearchObject/EBSCO.php or add links to new searches, etc., it is this one that you should tweak).
Made for "blueprint" theme only. Don't know how the .tpls and CSS will behave in other themes... I’m sending a styles-ebsco.css that it's not my working one, because I have done some severe changes in mine to match the theme I’m using, an adapted one from blueprint. Instead I populated it with trunks' blueprint/css/styles.css: https://vufind.svn.sourceforge.net/svnroot/vufind/trunk/web/interface/themes/blueprint/css/styles.css contents (most of them not necessary… :| ). Also sending my web/interface/themes/blueprint/css/blueprint/css/blueprint/screen-ebsco.css that should work also out-of-the-box (have a logfile of every changes I’ve done to the several VuFind files, new I’ve created, and from a fast look, should be ok with a standard VuFind install).
To make it work for you > should only be necessary to input your EBSCO EIT access data:
orgid = University of ...
prof = sXXXXXXX.default.eit
pwd = YYYYYYY
in web/conf/EBSCO.ini
NOTE #1: I’m hardcoding the DBs to search in web/interface/themes/blueprint/EBSCO/searchbox.tpl
<select id="db" name="db">
<option value="a9h"{if $activedb == 'a9h'} selected{/if}>Academic Search Complete</option>
<option value="bth"{if $activedb == 'bth'} selected{/if}>Business Source Complete</option>
<option value="mnh"{if $activedb == 'mnh'} selected{/if}>MEDLINE with Full Text</option>
<option value="eric"{if $activedb == 'eric'} selected{/if}>ERIC: Education Resources Info. Center</option>
</select>
NOTE #1.1: EBSCO does not allow multi-db search at the same time with EIT; otherwise, we would be recreating EDS (in part) and so killing an important added-value flag to its market (and marketing);
NOTE #2: the only file that it's not new is web/interface/themes/blueprint/layout.tpl >> please make a backup of it first, just in case;
NOTE #3: You can check if your EBSCO EIT account details are working by testing with an URL (in this case, if you subscribe "Academic Search Complete": db=a9h):
http://eit.ebscohost.com/Services/SearchService.asmx/Search?prof=sXXXXXXX.default.eit+&pwd=YYYYYY&authType=profile&ipprof=&query=AN%3d69736890&db=a9h
replacing XXXXXX and YYYYY with your values, should give a XML file like:
<searchResponse xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema">
<Hits xmlns="http://epnet.com/webservices/SearchService/Response/2007/07/">1</Hits>
<Statistics xmlns="http://epnet.com/webservices/SearchService/Response/2007/07/">
<Statistic>
<Database>a9h</Database>
<Hits>1</Hits>
</Statistic>
</Statistics>
<SearchResults xmlns="http://epnet.com/webservices/SearchService/Response/2007/07/">
<records xmlns="">
<rec recordID="1">
<pdfLink/>
<plink>
http://search.ebscohost.com/login.aspx?direct=true&db=a9h&AN=69736890&site=ehost-live
</plink>
<header shortDbName="a9h" uiTerm="69736890" longDbName="Academic Search Complete" uiTag="AN">
(….)
NOTE #4: I’ve done this pack by going through the entire site backup and delete the files that aren’t related to this new feature or affected. If there is/are some file(s) missing, please let me know by sending the error displayed with “debug = true”. Also, I hope I’ve deleted all of our personal ESBCO’s account access data > if you detect any anywhere (I’m pretty sure there aren’t), please also let me know ASAP. Thank you!
NOTE #5: Haven't changed any of the original credits or added new ones (in the ones I've modified), so all the credits should go to Vaclav Rosecky and Ere Maijala;
NOTE #6: Still without access to my staging install so the zip instead of tar (suggestion: unzip and copy the files over to your Linux server using WinSCP, for instance);
NOTE #7 (and final): I never, ever used “I” for the Institutional work I’ve developed, but instead always “we” and remain anonymous (instead I use the name of the Services | Team); the “I” is because, as some of you are aware, “my” VuFind install is for the time being part of a wider project and associated study for my PhD. Soon, I hope to write “we”, meaning that has been approved to enter production at my home University (work in progress – final steps, I hope), even if still being in a long term LOA (Leave of Absence), writing down the thesis after conducting several Focus Groups representative of the several communities of University of Aveiro’s users, I hope.
The conceptual part and results obtained from the initial survey by questionnaire (online: 26 questions [most of them conditional] > 4.228 complete responses from a 14.000+ members University of Aveiro’s Academic Community) are already public and published in this book chapter:
Going Beyond the Bibliographic Catalog:
The Basis for a New Participatory Scientific Information Discovery and Sharing Model
Filipe Bento and Lídia Silva (University of Aveiro, Portugal)
IGI Global, 2013. ISBN 9781466619128, p. 1-38
http://www.igi-global.com/chapter/going-beyond-bibliographic-catalog/69262
part of the book I’ve been including in my signature (that as mentioned there, has a lot of VuFind related book chapters, including a Demian’s one ).
We (University of Aveiro's Libraries [supporters of this project / first final clients of it], my advisor and I) are also providing access to the full raw results of the questionnaire (“Open Data”) that can be used for further studies / cross data in new ways or to compare to other obtained data / studies, as long as it is for pure research proposes or academic usage and not to be released for commercial usage of any kind.
EDIT: Russ, could you please confirm that you were successful with the attached files? Many thanks!
|
Filipe, your package does indeed work. I just copied over the contents and moved the EBSCO template files to our own Blueprint-based template, made changes to our layout.tpl to include your modification, and put the relevant information into EBSCO.ini and the search is working well.
Thank you for repackaging it, much appreciated! As a bit of feature creep, putting the databases in the config file was a nice idea and would like to see that implemented at some point if possible, but having this working is fantastic.
Thanks again!
Demian: emails from this Jira instance still not coming through even though I'm watching the ticket... Just an FYI since I know you mentioned it before.
|
Russ,
I've been working to fix the lack of email notifications. I think I may have just made some progress. Please let me know if you receive a notification as a result of this latest comment!
|
|
I did get the email notification, great!
|
> Thanks again!
Russell, you're more than welcome! Thank you so much for confirming the package did indeed work; because I did not have an install where to check it first before attaching it, crossed my fingers, hoping that I haven’t exclude some relevant file or that the files that I’ve made for this purpose, using Ere's diff file or VuFind trunk’s CSS, were 100% flawless.
Regarding having the DBs codes and description in EBSCO.ini file (and not hardcode in the web/interface/themes/blueprint/EBSCO/searchbox.tpl), I’m with you 100%. When I have regained access to my VuFind staging server, I shall try to test patched code/files that places those values back in EBSCO.ini and read them as an array (not a single value as thought of, if not in mistake, in Vaclav Rosecky / Ere Maijala versions), having that entry EBSCO.ini as something as:
db = a9h:Academic Search Complete, bth:Business Source Complete, …
instead of the original single value / DB available to search (for instance "db = a9h").
All the best and hope to be back to this ticket ASAP to attach such a patch,
Filipe
|
|
Guys, I'm not sure if you're aware of this, but EBSCO has published some integration code on their wiki (http://edswiki.ebscohost.com/EBSCO_Discovery_Service_API_Sample_Applications), and they also have a VUFind demo running at http://vufinddemo.ebscohost.com/. I have the impression that this is using a new API that provides e.g. facet counts.
|
|
The wiki requires registration, and I haven't actually looked at it yet.
|
I have tried the official EBSCO integration and can confirm that the code they provide in the wiki is working, and its really using API2. One needs new account data to access the new API and use this official integration, the account data for access to the old API is not valid for the new one.
Our install looks exactly like the demo instance Ere pointed to, so having a look at this demo should provide a good start evaluating it. The captcha on the demo (every time you start a search) is fortunately not included in the code (or at least I have not found the point where to enable it yet - not that I would want that...).
|
Ere, Oliver (and all), hi!
Correct if I am wrong, but to have access to this new API (EDS API), doesn’t one have to subscribe to EDS service?
If it is similar to EIT (free for all of EBSCO’s customers, even they just subscribe a single EBSCO DB), one could use it to take ideas how to take the best out of EIT, like Oliver mentioned (well, if you are not an EDS subscriber), but I think we are pretty close to it (there isn’t much EIT can offer that we are not using already --- unfortunately no records’ counts in facets)…
Oliver, can you confirm that your institution subscribes EDS or you’ve requested a trial?
Many thanks,
Filipe
PS: otherwise, like I’ve mentioned before, if they offer EDS API for free or the same features in EIT API, what would be driving you to buy/subscribe EDS?
PPS: anyone watching this issue that is using MS-ForeFront (Online Security) is receiving the alert emails when new comments are entered? According to Demian, the ones sent to me are bouncing back with the compliments of MS-ForeFront (not being delivered). >> EDIT :: Status: SOLVED as of Wednesday, August 15, 2012 12:45 PM EST, by Microsoft Online Services Technical Support team (thanks, Demian for the diligences taken near MicroSoft!).
---------------
EDIT (5 seconds later :| ): Answering my own question > http://edswiki.ebscohost.com/EDS_API_Documentation :
"Getting Started
(...) In order to use the EDS API, you must be an EBSCO Discovery Service (EDS) Customer." (sic)
As expected, plain & pure "TANSTAAFL": There Ain't No Such Thing As A Free Lunch!
|
|
So essentially there may be a need for two separate VuFind modules -- one for the free but more limited API, and one for the full subscription API. That way, EDS users can take advantage of all the benefits, but other EBSCO subscribers can integrated database content into VuFind in other ways. I wonder if there's enough commonality between the APIs that this can be done as some kind of base-class/subclass arrangement, or if we're truly talking about two separate entities.
|
|
My bad for not realizing the need of EDS (not that I really understand the different licensing options EBSCO offers to begin with). I believe the EDS API is completely different from EIT.
|
We have just a trial license to use EDS API 1 and 2. EDS API 1 can be used with Ere's patch. EDS API 2 can be used with the code EDS provides on their Wiki. As I pointed out before, the account data for EDS API 1 is different from the account data to EDS API 2.
I'm, as Ere, believing the free EIT API is a different thing, but I did not dig into that deeper yet... Ebsco's policy here is a bit confusing...
|
I suppose what I wrote was not completely the truth: it seems I mixed EDS API with EIT API. EIT API is supposed to be what I considered to be EDS API 1. I never used the term EIT API before, though...
I hope that was not too confusing at all.
|
Just a quick remark worth doing: it is my believe that it’s really a great honor for VuFind to be elected by EBSCO as a showcase how Institutions can use their own local interface enriched with data provided in real time by EDS API services, instead of the standard one, in the clouds, EBSCO offers for accessing EDS (and this way also saving some precious $s, that you have to pay as an extra, if you want to include your own OPAC records, local institutional repository, publishing platform ones (OJS, for instance), etc.).
So, VuFind: Congratulations! (Villanova University's Falvey Memorial Library and all VuFind community, from developers to Institutions using VuFind, for betting in a solution that is growing every single day, either in features and in community members and so with higher odds of a sustainable/continuous development).
|
It is my opinion that we should split this ticket into EIT (this one) and a new one, for EDS API, so that this one can be closed, as, despite the fact the DBs and their description should be in EBSCO.ini (going to place my hands on it right away), it's pretty much ready to have its status changed to solved / close it.
What do you say?
Please bear in mind Demian’s initial description refers only to EIT API (EBSCOhost Integration Toolkit API): "EBSCO offers a free API for searching their databases > http://www.ebscohost.com/academic/ebscohost-integration-toolkit | http://eit.ebscohost.com/", free to any EBSCO’s customer and so much more universal / of more generalized interest I would say, not the EDS API (new) available only to EDS subscribers (even the EDS API wiki registration requires that, it seems to me).
Please don’t mix these two in the same pot, as their target audience is very different and reports from the ones who had analyzed this new one (EDS API) say it’s different enough from EIT one.
As mentioned, to have access to EDS API (EBSCO Discovery Service API) you have to subscribe to a complete new product / service, EDS: it’s EBSCO version of Ex Libris PrimoCentral, SS Summon, etc., and it costs so much or more than these last (I'm not allowed to disclosure prices here [and if I was, I wouldn't], because they depend from many factors, but I know how much it costs each of these solutions [even the several variations of Ex Libris Primo], no matter if you have a current active subscription to an EBSCO DB, you will have to pay for it as an extra and the value... well, if you know the ones from SS Summon or PrimoCentral, you have an idea. I know this because I belonged to the Working Group selected by b-on (our National Consortium, http://www.b-on.pt/) to evaluate possible solutions to replace MetaLib, and besides local demos, we received real/concrete proposals (that once again, I’m not allowed to disclosure, beside being for a nationwide usage, many Universities [and other HE institutions] and Hospitals).
In fact, locally (UA) we are implementing VuFind because we don’t have the money to buy/subscribe such solutions. We are aware that most of the resources, besides local ones, shall be OA ones only, but nevertheless, we always have this (EIT for EBSCO subscribed DBs) and b-on (external) for the most part of the remaining subscribed resources (ExLibris MetabLib+).
EDIT: also to avoid confusions, now that “we have” two different APIs from EBSCO (EIT and EDS) and since they seem to be pretty much different, perhaps this ticket’s title should be changed to Create module for EBSCO EIT API
|
I have opened a new ticket for EDS (VUFIND-636) and adjusted the title and description of this ticket to make it EIT-specific.
Before we can close this ticket, we need to make sure the code is 2.0-compatible. It might also be a good idea to rename some of the components so the code cleanly differentiates between EIT and EDS to prevent further confusion on this subject.
|
Thank you very much for clarifying that, Filipe! I aggree that we should split this ticket into two or at least rename it to avoid confusions for EBSCO-Users.
Just to sum it up (please correct me if I'm wrong in any point):
* EBSCO offers two APIs: EBSCO EIT API (API1) and EBSCO EDS API (API2)
* The two APIs are independent from each other and work completely different (they even need different account data)
* This ticket deals with EBSCO EIT API (API1)
* Access to EBSCO EDS API (API2) needs to be paid (which can be quite expensive)
* Access to EBSCO EIT API (API1) is free for any EBSCO customer - there are no additional costs for API usage
Can anyone confirm the last point, please? What data is supplied with API1 - only the metadata of the subscribed EBSCO databases? Are fulltexts included? Will API1 access continue working, since API2 has been established now?
|
Oliver, yes, 100% correct.
I don't see EBSCO discontinuing EIT API, as they still full support z39.50, for instance.
EDS is a different universe and won't affect the regular EBSCO DBs' subscriber.
I do not represent EBSCO in anyway, but have been working with them for ages, so in this department we are secure, EIT API is where to stay (and free).
Filipe
PS: yes, what you got in EBSCOhost interface search you get in EIT and even more. I’ll attach some screenshots of mine. It is just not possible to search across all the subscribed DBs at the same time, as EDS API allows for sure.
PPS (to all): please don’t get the wrong idea about EIT API – have a look at the package I’ve sent or the the URL with your EIT access data – quite powerfull – lots of data to extract if you choose the most complete format!
|
|
EBSCO EIT API in action (1 of X) -- Imagem 044.png
|
EBSCO EIT API in action -- linkage data extract allow several other actions, inside EBSCO or retrieve from other sources.
--> Imagem 045.png: when you click in the title of the article / eBook / book chapter, etc..
|
EBSCO EIT API in action: share in external services / communities.
--> Imagem 046.png: when you click on SJR in "Journal Name: Journal of Applied Physics [WCat][SJR]";
--> Imagem 047.png: when you click on "Get full text";
--> Imagem 048.png: when you click on "Access this article at the editor site";
--> Imagem 049.png: when you click on the "diigo" symbol / logo (all the information there was filled in automatically by the system).
PS: only the first article has the special right of showing authors affiliation and full abstract > see Imagem 051.png
|
|
Imagem 051.png: only the first article has the special right of showing authors affiliation and full abstract.
|
EBSCO EIT API in action -- linkage data extract allow several other actions, inside EBSCO or retrieve from other sources.
Imagem 052.png: should have attached this one instead, for the part when you click on "SJR" in front of the journal name... and it has a lot more.
Btw, didn't find a way to link directly to Thomson Reuters' Journal Citation Reports (formerly known as ISI JCR)... it's a shame. I think it their lost to not let subscribers of such service link to it in their own Webs. And I have no problem in saying this, as we are very good customers of Thomson Reuters (WoK [WoS], EndNote since the early versions – campus license, customized by me [the install], just to name a few).
|
|
Update: I've just spoken to Tom Preston from EBSCO, and he mentioned that EIT is being phased out in favor of a newer API. We probably shouldn't commit this patch as a result -- better to wait for the new one to be fully ready.
|
|
Julia Bauder has heard that the API will remain active for a while at least, and has provided this pull request: https://github.com/vufind-org/vufind/pull/142
|
Generated at Sat Apr 20 13:47:04 UTC 2024 using Jira 1001.0.0-SNAPSHOT#100250-rev:2b88e55752dc82be8616a67bc2b73a87c8e22b48.
EBSCO_EIT_RealTimeSearch_Filipe_Bento_20120810.zip
Imagem 044.png
NDL_ebsco_v1.diff