source: other-projects/nightly-tasks/diffcol/trunk/model-collect/Enhanced-PDF/etc/collect.cfg@ 27976

Last change on this file since 27976 was 27976, checked in by ak19, 11 years ago

Updating Enhanced-PDF collection now that extra_meta is sorted and the images generated from a pdf are sorted in doc.xml's gsdlassocfile meta section

File size: 2.5 KB
Line 
1creator
2maintainer
3public true
4
5buildtype mgpp
6infodbtype gdbm
7
8indexes text dc.Title,ex.dc.Title,ex.Title ex.Source
9defaultindex text
10
11levels document section
12
13indexoptions accentfold casefold stem
14
15defaultlevel document
16
17# import options, needed for diffcol testing
18# especially in conjunction with -sort flag to ArchivesInfPlugin
19OIDtype hash_on_full_filename
20sortmeta OID
21
22plugin ZIPPlugin
23plugin GreenstoneXMLPlugin
24plugin TextPlugin
25plugin HTMLPlugin
26plugin EmailPlugin
27plugin PDFPlugin -process_exp notext.*\.pdf -convert_to pagedimg_jpg
28plugin PDFPlugin -use_sections -convert_to html
29plugin RTFPlugin
30plugin WordPlugin
31plugin PostScriptPlugin
32plugin PowerPointPlugin
33plugin ExcelPlugin
34plugin ImagePlugin
35plugin ISISPlugin
36plugin NulPlugin
37plugin EmbeddedMetadataPlugin
38plugin MetadataXMLPlugin
39plugin ArchivesInfPlugin -sort
40plugin DirectoryPlugin
41
42classify List -metadata dc.Title;ex.Title -partition_type_within_level approximate_size
43classify List -metadata ex.Source -partition_type_within_level approximate_size
44
45format VList "<td valign=\"top\">[link][icon][/link]</td>
46<td valign=\"top\">[ex.srclink]{Or}{[ex.thumbicon],[ex.srcicon]}[ex./srclink]</td>
47<td valign=\"top\">[highlight]
48{Or}{[dc.Title],[exp.Title],[ex.Title],Untitled}
49[/highlight]{If}{[ex.Source],<br><i>([ex.Source])</i>}</td>"
50
51format HList "[link][highlight][ex.Title][/highlight][/link]"
52
53format DocumentHeading "{Or}{[parent(Top):Title],[Title],untitled}<br>"
54
55format DocumentText "{If}{[NoText] eq \"1\",[srcicon],[Text]}"
56
57format DocumentButtons "Detach|Highlight"
58
59format SearchTypes "plain,form"
60
61format SearchVList "<td valign=\"top\">[link][icon][/link]</td>
62<td valign=\"top\">{If}{[ex.FileFormat] eq \'PDF\', <a href=\\"_httpcollection_/index/assoc/[archivedir]/[srclinkFile]#search=&quot;_queryterms_&quot;\\">{Or}{[ex.thumbicon],[ex.srcicon]}</a>,
63[ex.srclink]{Or}{[ex.thumbicon],[ex.srcicon]}[ex./srclink]}</td>
64<td valign=\"top\">[highlight]
65{Or}{[dc.Title],[ex.Title],Untitled}
66[/highlight]{If}{[ex.Source],<br><i>([ex.Source])</i>}</td>"
67
68collectionmeta collectionname [l=en] "Enhanced-PDF"
69collectionmeta .text [l=en] "_labeltext_"
70collectionmeta .dc.Title,ex.dc.Title,Title [l=en] "_labelTitle_"
71collectionmeta .Source [l=en] "_labelSource_"
72collectionmeta .document [l=en] "_textdocument_"
73collectionmeta .section [l=en] "_textsection_"
74collectionmeta .document:text [l=en] "_labeltext_"
75collectionmeta .document:dc.Title,Title,ex.dc.Title [l=en] "_labelTitle_"
76collectionmeta .document:Source [l=en] "_labelSource_"
77collectionmeta .document:dc.Title,Title [l=en] "_labelTitle_"
Note: See TracBrowser for help on using the repository browser.