Uploaded On |
May 4, 2012, 5:29 p.m. |
Uploaded By |
matts |
Status |
Production (approved on May 4, 2012, 5:29 p.m. by matts)
|
<map>
<entry>
<string>plugin_version</string>
<string>3</string>
</entry>
<entry>
<string>plugin_config_props</string>
<list>
<org.lockss.daemon.ConfigParamDescr>
<key>base_url</key>
<displayName>Base URL</displayName>
<description>Usually of the form http://<journal-name>.com/</description>
<type>3</type>
<size>40</size>
<definitional>true</definitional>
<defaultOnly>false</defaultOnly>
</org.lockss.daemon.ConfigParamDescr>
</list>
</entry>
<entry>
<string>au_name</string>
<string>"Virginia Tech ETD's - Early unsorted"</string>
</entry>
<entry>
<string>au_start_url</string>
<string>"%slockss/manifest.html", base_url</string>
</entry>
<entry>
<string>au_def_pause_time</string>
<long>6000</long>
</entry>
<entry>
<string>au_def_new_content_crawl</string>
<long>31536000000</long>
</entry>
<entry>
<string>plugin_notes</string>
<string>Harvest unrestricted ETD's published prior to the year 2000 which do not follow any sane or logical naming structure. Essentially grabs anything that doesn't match the 2000-present naming convention and puts it into one giant AU separate from the other VT ETD's. No new content *should* show up here.</string>
</entry>
<entry>
<string>plugin_name</string>
<string>VT ETD's early unsorted</string>
</entry>
<entry>
<string>plugin_identifier</string>
<string>edu.vt.library.thesesearly</string>
</entry>
<entry>
<string>au_crawlrules</string>
<list>
<string>4,"^%s", base_url</string>
<string>1,"%slockss/manifest.html", base_url</string>
<string>1,"^%swithheld/?$", base_url</string>
<string>1,"%sbrowse/by_author/all.html", base_url</string>
<string>2,"/\?"</string>
<string>2,"%savailable/etd-[0-9]+-[0-9]+/.*", base_url</string>
<string>2,"%swithheld/etd-[0-9]+-[0-9]+/.*", base_url</string>
<string>1,"%savailable/.*", base_url</string>
<string>1,"%swithheld/.*", base_url</string>
</list>
</entry>
</map>