proiel 1.0.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/LICENSE +19 -0
- data/README.md +99 -0
- data/bin/console +6 -0
- data/bin/setup +5 -0
- data/lib/proiel/annotation_schema.rb +127 -0
- data/lib/proiel/citations.rb +84 -0
- data/lib/proiel/div.rb +133 -0
- data/lib/proiel/positional_tag.rb +127 -0
- data/lib/proiel/proiel_xml/proiel-1.0/proiel-1.0.xsd +172 -0
- data/lib/proiel/proiel_xml/proiel-1.0/teilite.xsd +7387 -0
- data/lib/proiel/proiel_xml/proiel-1.0/xml.xsd +287 -0
- data/lib/proiel/proiel_xml/proiel-2.0/proiel-2.0.xsd +185 -0
- data/lib/proiel/proiel_xml/reader.rb +237 -0
- data/lib/proiel/proiel_xml/schema.rb +81 -0
- data/lib/proiel/proiel_xml/validator.rb +177 -0
- data/lib/proiel/sentence.rb +191 -0
- data/lib/proiel/source.rb +114 -0
- data/lib/proiel/statistics.rb +41 -0
- data/lib/proiel/token.rb +407 -0
- data/lib/proiel/tokenization.rb +90 -0
- data/lib/proiel/treebank.rb +214 -0
- data/lib/proiel/treebank_object.rb +21 -0
- data/lib/proiel/version.rb +9 -0
- data/lib/proiel.rb +28 -0
- metadata +210 -0
|
@@ -0,0 +1,172 @@
|
|
|
1
|
+
<?xml version="1.0"?>
|
|
2
|
+
|
|
3
|
+
<!-- PROIEL XML format version 1.0 -->
|
|
4
|
+
|
|
5
|
+
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
|
|
6
|
+
xmlns:tei="http://www.tei-c.org/ns/1.0">
|
|
7
|
+
<xs:import namespace="http://www.w3.org/XML/1998/namespace"
|
|
8
|
+
schemaLocation="http://www.w3.org/2001/xml.xsd"/>
|
|
9
|
+
<xs:import namespace="http://www.tei-c.org/ns/1.0"
|
|
10
|
+
schemaLocation="http://www.tei-c.org/release/xml/tei/custom/schema/xsd/teilite.xsd"/>
|
|
11
|
+
|
|
12
|
+
<xs:annotation>
|
|
13
|
+
<xs:documentation xml:lang="en">PROIEL XML format version 1.0</xs:documentation>
|
|
14
|
+
</xs:annotation>
|
|
15
|
+
|
|
16
|
+
<xs:complexType name="Source">
|
|
17
|
+
<xs:sequence>
|
|
18
|
+
<xs:element name="title" minOccurs="1" maxOccurs="1" type="xs:string"/>
|
|
19
|
+
<xs:element name="author" minOccurs="0" maxOccurs="1" type="xs:string"/>
|
|
20
|
+
<xs:element name="edition" minOccurs="0" maxOccurs="1" type="xs:string"/>
|
|
21
|
+
<xs:element name="citation-part" minOccurs="1" maxOccurs="1" type="xs:string"/>
|
|
22
|
+
<xs:element name="tei-header" minOccurs="1" maxOccurs="1" type="TeiHeader"/>
|
|
23
|
+
<xs:element name="div" minOccurs="1" maxOccurs="unbounded" type="Div"/>
|
|
24
|
+
</xs:sequence>
|
|
25
|
+
<xs:attribute name="id" type="xs:string" use="required"/>
|
|
26
|
+
<xs:attribute name="language" type="xs:string" use="required"/>
|
|
27
|
+
</xs:complexType>
|
|
28
|
+
|
|
29
|
+
<xs:complexType name="TeiHeader">
|
|
30
|
+
<xs:sequence>
|
|
31
|
+
<xs:element ref="tei:teiHeader" minOccurs="0" maxOccurs="1"/>
|
|
32
|
+
</xs:sequence>
|
|
33
|
+
</xs:complexType>
|
|
34
|
+
|
|
35
|
+
<xs:complexType name="Div">
|
|
36
|
+
<xs:sequence>
|
|
37
|
+
<xs:element name="title" minOccurs="1" maxOccurs="1" type="xs:string"/>
|
|
38
|
+
<xs:element name="sentence" minOccurs="1" maxOccurs="unbounded" type="Sentence"/>
|
|
39
|
+
</xs:sequence>
|
|
40
|
+
<xs:attribute name="presentation-before" type="xs:string" use="optional"/>
|
|
41
|
+
<xs:attribute name="presentation-after" type="xs:string" use="optional"/>
|
|
42
|
+
</xs:complexType>
|
|
43
|
+
|
|
44
|
+
<xs:complexType name="Sentence">
|
|
45
|
+
<xs:sequence>
|
|
46
|
+
<xs:element name="token" minOccurs="1" maxOccurs="unbounded" type="Token"/>
|
|
47
|
+
</xs:sequence>
|
|
48
|
+
<xs:attribute name="id" type="xs:nonNegativeInteger" use="optional"/>
|
|
49
|
+
<xs:attribute name="presentation-before" type="xs:string" use="optional"/>
|
|
50
|
+
<xs:attribute name="presentation-after" type="xs:string" use="optional"/>
|
|
51
|
+
<xs:attribute name="status" type="SentenceStatus" use="optional"/>
|
|
52
|
+
</xs:complexType>
|
|
53
|
+
|
|
54
|
+
<xs:simpleType name="SentenceStatus">
|
|
55
|
+
<xs:restriction base="xs:string">
|
|
56
|
+
<xs:enumeration value="annotated"/>
|
|
57
|
+
<xs:enumeration value="reviewed"/>
|
|
58
|
+
<xs:enumeration value="unannotated"/>
|
|
59
|
+
</xs:restriction>
|
|
60
|
+
</xs:simpleType>
|
|
61
|
+
|
|
62
|
+
<xs:complexType name="Token">
|
|
63
|
+
<xs:sequence>
|
|
64
|
+
<xs:element name="slash" minOccurs="0" maxOccurs="unbounded" type="Slash"/>
|
|
65
|
+
</xs:sequence>
|
|
66
|
+
|
|
67
|
+
<xs:attribute name="id" type="xs:nonNegativeInteger" use="optional"/>
|
|
68
|
+
<xs:attribute name="lemma" type="xs:string" use="optional"/>
|
|
69
|
+
<xs:attribute name="part-of-speech" type="xs:string" use="optional"/>
|
|
70
|
+
<xs:attribute name="morphology" type="xs:string" use="optional"/>
|
|
71
|
+
<xs:attribute name="citation-part" type="xs:string" use="optional"/>
|
|
72
|
+
<xs:attribute name="relation" type="xs:string" use="optional"/>
|
|
73
|
+
<xs:attribute name="head-id" type="xs:nonNegativeInteger" use="optional"/>
|
|
74
|
+
<xs:attribute name="information-status" type="xs:string" use="optional"/>
|
|
75
|
+
<xs:attribute name="antecedent-id" type="xs:nonNegativeInteger" use="optional"/>
|
|
76
|
+
<xs:attribute name="contrast-group" type="xs:string" use="optional"/>
|
|
77
|
+
<xs:attribute name="foreign-ids" type="xs:string" use="optional"/>
|
|
78
|
+
|
|
79
|
+
<!-- XSD does not allow us to constrain the use of these attributes properly so they are marked optional even though their occurrence depends on the value of empty-token-sort and form. -->
|
|
80
|
+
<xs:attribute name="empty-token-sort" type="EmptyTokenSort" use="optional"/>
|
|
81
|
+
<xs:attribute name="form" type="xs:string" use="optional"/>
|
|
82
|
+
<xs:attribute name="presentation-before" type="xs:string" use="optional"/>
|
|
83
|
+
<xs:attribute name="presentation-after" type="xs:string" use="optional"/>
|
|
84
|
+
</xs:complexType>
|
|
85
|
+
|
|
86
|
+
<xs:simpleType name="EmptyTokenSort">
|
|
87
|
+
<xs:restriction base="xs:string">
|
|
88
|
+
<xs:enumeration value="P"/>
|
|
89
|
+
<xs:enumeration value="C"/>
|
|
90
|
+
<xs:enumeration value="V"/>
|
|
91
|
+
</xs:restriction>
|
|
92
|
+
</xs:simpleType>
|
|
93
|
+
|
|
94
|
+
<xs:complexType name="Slash">
|
|
95
|
+
<xs:attribute name="target-id" type="xs:nonNegativeInteger" use="required"/>
|
|
96
|
+
<xs:attribute name="relation" type="xs:string" use="required"/>
|
|
97
|
+
</xs:complexType>
|
|
98
|
+
|
|
99
|
+
<xs:complexType name="PartOfSpeechValue">
|
|
100
|
+
<xs:attribute name="tag" type="xs:string" use="required"/>
|
|
101
|
+
<xs:attribute name="summary" type="xs:string" use="required"/>
|
|
102
|
+
</xs:complexType>
|
|
103
|
+
|
|
104
|
+
<xs:complexType name="PartsOfSpeech">
|
|
105
|
+
<xs:sequence>
|
|
106
|
+
<xs:element name='value' minOccurs="1" maxOccurs="unbounded" type='PartOfSpeechValue'/>
|
|
107
|
+
</xs:sequence>
|
|
108
|
+
</xs:complexType>
|
|
109
|
+
|
|
110
|
+
<xs:complexType name="InformationStatusValue">
|
|
111
|
+
<xs:attribute name="tag" type="xs:string" use="required"/>
|
|
112
|
+
<xs:attribute name="summary" type="xs:string" use="required"/>
|
|
113
|
+
</xs:complexType>
|
|
114
|
+
|
|
115
|
+
<xs:complexType name="InformationStatuses">
|
|
116
|
+
<xs:sequence>
|
|
117
|
+
<xs:element name='value' minOccurs="1" maxOccurs="unbounded" type='InformationStatusValue'/>
|
|
118
|
+
</xs:sequence>
|
|
119
|
+
</xs:complexType>
|
|
120
|
+
|
|
121
|
+
<xs:complexType name="RelationValue">
|
|
122
|
+
<xs:attribute name="tag" type="xs:string" use="required"/>
|
|
123
|
+
<xs:attribute name="summary" type="xs:string" use="required"/>
|
|
124
|
+
<xs:attribute name="primary" type="xs:boolean" use="required"/>
|
|
125
|
+
<xs:attribute name="secondary" type="xs:boolean" use="required"/>
|
|
126
|
+
</xs:complexType>
|
|
127
|
+
|
|
128
|
+
<xs:complexType name="Relations">
|
|
129
|
+
<xs:sequence>
|
|
130
|
+
<xs:element name='value' minOccurs="1" maxOccurs="unbounded" type='RelationValue'/>
|
|
131
|
+
</xs:sequence>
|
|
132
|
+
</xs:complexType>
|
|
133
|
+
|
|
134
|
+
<xs:complexType name="MorphologyValue">
|
|
135
|
+
<xs:attribute name="tag" type="xs:string" use="required"/>
|
|
136
|
+
<xs:attribute name="summary" type="xs:string" use="required"/>
|
|
137
|
+
</xs:complexType>
|
|
138
|
+
|
|
139
|
+
<xs:complexType name="MorphologyField">
|
|
140
|
+
<xs:sequence>
|
|
141
|
+
<xs:element name='value' minOccurs="1" maxOccurs="unbounded" type='MorphologyValue'/>
|
|
142
|
+
</xs:sequence>
|
|
143
|
+
|
|
144
|
+
<xs:attribute name="tag" type="xs:string" use="required"/>
|
|
145
|
+
</xs:complexType>
|
|
146
|
+
|
|
147
|
+
<xs:complexType name="Morphology">
|
|
148
|
+
<xs:sequence>
|
|
149
|
+
<xs:element name='field' minOccurs="1" maxOccurs="unbounded" type='MorphologyField'/>
|
|
150
|
+
</xs:sequence>
|
|
151
|
+
</xs:complexType>
|
|
152
|
+
|
|
153
|
+
<xs:complexType name="Annotation">
|
|
154
|
+
<xs:sequence>
|
|
155
|
+
<xs:element name='relations' minOccurs="1" maxOccurs="1" type='Relations'/>
|
|
156
|
+
<xs:element name='parts-of-speech' minOccurs="1" maxOccurs="1" type='PartsOfSpeech'/>
|
|
157
|
+
<xs:element name='morphology' minOccurs="1" maxOccurs="1" type='Morphology'/>
|
|
158
|
+
<xs:element name='information-statuses' minOccurs="1" maxOccurs="1" type='InformationStatuses'/>
|
|
159
|
+
</xs:sequence>
|
|
160
|
+
</xs:complexType>
|
|
161
|
+
|
|
162
|
+
<xs:complexType name="Proiel">
|
|
163
|
+
<xs:sequence>
|
|
164
|
+
<xs:element name='annotation' minOccurs="0" maxOccurs="1" type='Annotation'/>
|
|
165
|
+
<xs:element name='source' minOccurs="1" maxOccurs="unbounded" type='Source'/>
|
|
166
|
+
</xs:sequence>
|
|
167
|
+
<xs:attribute name='export-time' type="xs:dateTime" use="optional"/>
|
|
168
|
+
<xs:attribute name="schema-version" type="xs:decimal" use="required" fixed="1.0"/>
|
|
169
|
+
</xs:complexType>
|
|
170
|
+
|
|
171
|
+
<xs:element name='proiel' type='Proiel'/>
|
|
172
|
+
</xs:schema>
|