Skip to content
GitLab
Select Git revision
  • nettoyage
  • master default
  • AlvisNLP-SB
  • testTermExtraction
Find file BlameHistoryPermalink
main-annotation.plan 2.21 KiB
1 
<?xml version="1.0" encoding="UTF-8"?>
2 
<alvisnlp-plan id="main">
3
4
5 
	<!--
6 
  	///////////////////////////////////////////////////////////////////////////////////////
7 
  	//
8 
  	//  I. Load corpus
9 
  	//
10 
  	///////////////////////////////////////////////////////////////////////////////////////
11 
  	-->
12 
	<!-- Load and segment html pages -->
13 
	<!-- Assign every page with corresponding names of each corpora -->
14
15 
	<load-corpus>
16 
		<d2kab class="XMLReader">
17 
			<html/>
18 
			<xslTransform>resources/segmentation/html2alvisnlp.xslt</xslTransform>
19 
			<sourcePath>resources/corpus/test_d2kab/</sourcePath>
20 
			<constantSectionFeatures>corpus=test d2kab</constantSectionFeatures>
21 
		</d2kab>
22
23 
		<vespa class="XMLReader">
24 
			<html/>
25 
			<xslTransform>resources/segmentation/html2alvisnlp.xslt</xslTransform>
26 
			<sourcePath>resources/corpus/test_vespa/</sourcePath>
27 
			<constantSectionFeatures>corpus=test vespa</constantSectionFeatures>
28 
		</vespa>
29
30 
		<alea class="XMLReader">
31 
			<html/>
32 
			<xslTransform>resources/segmentation/html2alvisnlp.xslt</xslTransform>
33 
			<sourcePath>resources/corpus/test_alea/</sourcePath>
34 
			<constantSectionFeatures>corpus=test alea</constantSectionFeatures>
35 
		</alea>
36 
	</load-corpus>
37
38
39
40 
	<!--
41 
  	///////////////////////////////////////////////////////////////////////////////////////
42 
  	//
43 
  	//  II. Preprocess
44 
  	//
45 
  	///////////////////////////////////////////////////////////////////////////////////////
46 
  	-->
47 
	<!-- Segment documents to phrases and words -->
48 
	<!-- Plus, correct some POS-tags -->
49
50 
	<preprocess-text href="modules/preprocess.plan"/>
51
52
53
54 
	<!--
55 
  	///////////////////////////////////////////////////////////////////////////////////////
56 
  	//
57 
  	//  III. Project concepts
58 
  	//
59 
  	///////////////////////////////////////////////////////////////////////////////////////
60 
  	-->
61 
	<!-- (1)Project RDF-->
62 
	<!-- (2)Project TOMAP-->
63
64 
	<project-concepts href="modules/projectors.plan"/>
65
66
67 
	<!--
68 
  	///////////////////////////////////////////////////////////////////////////////////////
69 
  	//
70 
  	//  IV. Find themes
7172737475767778798081
// /////////////////////////////////////////////////////////////////////////////////////// --> <!-- Segment bulletins thematically --> <find-themes href="modules/patterns/find-themes.plan"/> </alvisnlp-plan>