Citeseer.ist.psu.edu

DEVONagent Tips
1

Hello,

if I do the following search in the DA-Browser:

citeseer.ist.psu.edu/cs?cs=1&q=a … m=0&af=Any

I get 8 hits. The above query is from your Citeseer.plist, only agentQuery and agentNumber are replaced.

If I do a search with “amorphous silicon hydrogen” and select your search engin Science->Citeseer I now get 0 Pages. In the Log window is no entry.

What went wrong?

Stefan

PS. If you use “amorphous silicon hydrogen” and select your search engin Science->SciSeek you get many results but every result don’t contains any of the above search words.

2

That page presents the results of a search for a phrase. So, although the word “hydrogen” doesn’t appear in any of the found items, a phrase search picked up items that contained the phrase “amorphous silicon hydrogenated”.

But the searches you made with a DEVONagent search set was different. DA was told to find pages that contain the words “amorphous” AND “silicon” AND “hydrogen”. See the difference?

You could have instead written your search query as “amorphous silicon (hydrogen OR hydrogenated)” and the results would have included those on your examples page. But a phrase search could have found still more pages that the modified DA query still wouldn’t find, such as pages containing the first two terms plus “hydrogenates” and “hydrogenating” for example.

So you could have written a more general DA query using a wildcard to “extend” the term “hydrogen”. so that “hydrogen”, “hydrogenate”, “hydrogenates”, “hydrogenating” and any other terms using the stem “hydrogen” would have been found.

You will find that the query operators and syntax in DA allow very powerful and precise query formulation.

3

Hello,

thanks for your replay.
Sorry for my bad explanation. I mean, if I use

amorphous silicon hydrogenated

which is the same as

amorphous and silicon and hydrogenated

I don’t get any result from DA for the search engine Citeseer.
If I use

amorphous silicon hydrogenated

in Citeseer.ist.psu.edu I get 8 hits.

Test results of the semitransparent amorphous silicon… - Fernandez Ferrando Josa (Correct)
7 July 1998 Test results of the semitransparent amorphous silicon sensors for the link system of CMS
1998 Test results of the semitransparent amorphous silicon sensors for the link system of CMS M.G.
studied here consist of (see Fig. 1) a layer of hydrogenated amorphous silicon of 1m thickness sandwiched
cmsdoc.cern.ch/documents/98/note98_050.ps.Z

Stefan

4

OK – I hadn’t looked at the Citeseer search page. You won’t get any results from a “standard” DA search set that queries the URL page itself, because the query isn’t fed to Citeseer’s search engine.

To do that, one needs to write a special plugin that will communicate with Citeseer’s search engine. There’s an introduction to writing DA plugins in the user documentation.

DEVONagent already includes a number of plugins, one of which is for Citeseer.

So I created a new Search Set and checked the Citeseer plugin for it. Result: no results for a search for the string "amorphous silicon hydrogenated ". What happened? I’m not sure, but in checking a few pages on the Citeseer site it appears they have been doing some housekeeping, resulting in redirection of previously linked pages. Perhaps the supplied Citeseer plugin will need to be updated as a result.

Frustrated by my attempt to find literature pertaining to “amorphous silicon hydrogenated” I then checked the plugin for Google on that new search set. Just for the heck of it, I instructed DEVONagent to follow links one level down.

Now, although the search takes a while on my satellite Internet connection, DEVONagent has downloaded to my computer a treasure trove of literature on the topic. I’ve got not just the 8 results found by Citeseeker directly, but 479 documents ranging from abstracts to papers, articles, patents and books. DEVONagent has prepared a summary of each document, a list of “topic” terms and will display a graphical depiction of term usage for each “topic”. So if I’m interested in chemical or physical properties, techniques for deposition of thin films and so on I can explore this collection of references with some assistance.

If I wish I can dump that collection to my Archive, or to a DEVONthink Pro database. Or I can instruct that search set to run periodically and inform me just of new literature that has appeared since the last search run, perhaps by sending me an email listing the new results.

Since I had told DEVONagent’s search set to ask Google for items related to my query, why didn’t I just ask Google in the first place? Here’s why: Google will provide a list of 754,000 page references (at the moment I looked), not the pages themselves. DEVONagent has selected 479 of those, downloaded them to my computer and provided a summary of each document together with some analysis to the words used in the collection.

5

Yes you are right, I can use google and DA supplies many results but google is not the on and only search engine. I think therefor different search engins plugins exists in DA.

But my main problem is, that I get no response from DA There is no window (or error log …) which logs the communication between DA and the web server.
How can I decide that if I get “0 Pages” result in DA that there is no result or only the search engine plugin don’t working. See the thread about an eBay plugin.

Besides, I have try to create a search plugin, as described in the manual, for the Journal of Applied Physics the same as above, no respons from DA. The URL string in the plugin works if you use it directly in a browser. How can I write a plugin for a server if there is no debug window?

Stefan

6

Our plugin is OK given that Citeseer hasn’t changed their website. What I’m afraid of is that they block DEVONagent from accessing the site. But a message indicating that no data has been received would be nice, I agree.

7

Actually (thanks to Christian) our plugin is not OK because a configuration error crept in. You can fix that yourself if you want until we ship an update by creating a new file Citeseer.plist in your “home folder > Library > Application Support > DEVONagent > Plugins” with these contents:


<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
	<key>DateEnd</key>
	<string>",</string>
	<key>DateStart</key>
	<string>year = "</string>
	<key>Description</key>
	<string>Searches the CiteSeer library. Citeseer is a scientific literature digital library and search engine that focuses primarily on the literature in computer and information science.</string>
	<key>EngineNextUrl</key>
	<string>http://citeseer.ist.psu.edu/cs?qb=dbnum%3D1%2Cstart%3D_agentOffset_%2Cao%3DCitations%2Cam%3D40%2Caf%3DAny%2Cqtype%3Ddocument:&amp;q=_agentQuery_</string>
	<key>EngineUrl</key>
	<string>http://citeseer.ist.psu.edu/cs?cs=1&amp;q=_agentQuery_&amp;submit=Documents&amp;co=Citations&amp;cm=50&amp;cf=Any&amp;ao=Citations&amp;am=_agentNumber_&amp;af=Any</string>
	<key>Identifier</key>
	<string>com.devon-technologies.Citeseer.plugin</string>
	<key>Info</key>
	<string>Citeseer (Science) Plugin ©2004-2007 DEVONtechnologies &amp; Ed Vawter (QD Information Services, LLC)</string>
	<key>LinksEnd</key>
	<string>Try your query at:</string>
	<key>LinksNotMatching</key>
	<array>
		<string>*/correct/*</string>
		<string>*citeseer.ist.psu.edu/cs*</string>
		<string>*citeseer.ist.psu.edu/rd/*</string>
	</array>
	<key>LinksStart</key>
	<string>documents found.</string>
	<key>Name</key>
	<string>Citeseer</string>
	<key>OffsetPerPage</key>
	<integer>20</integer>
	<key>Operators</key>
	<integer>123</integer>
	<key>OperatorsDictionary</key>
	<dict>
		<key>NEAR</key>
		<string>w</string>
	</dict>
	<key>ResultsPerPage</key>
	<integer>20</integer>
	<key>Start</key>
	<integer>0</integer>
	<key>TextEnd</key>
	<string>Cited by:</string>
	<key>TextStart</key>
	<string>Abstract:</string>
	<key>Version</key>
	<string>1.3</string>
</dict>
</plist>
8

Thank you for your answer.
I have copy your code as you suggested but DA don’t use the new plugin. But I am at work and only had a MacBook without XCode installed. I will try it again at home.

Is it possible to get in a new version of DA a “debug window” which logs exactly what DA sends and what DA gets back?

with best regards

Stefan

9

The next release will also log failed requests to search engines (see “Log” tab of search windows), especially after communication failure or if no data is returned.