From 39b0f6e6b3d81a6b0851dc2c6f246fd30f633040 Mon Sep 17 00:00:00 2001
From: Cyprian Laskowski <cyp@trojina.si>
Date: Wed, 26 Feb 2020 23:38:35 +0100
Subject: [PATCH] IssueID #1097: added extra attributes etc for MWE

---
 examples/mwe_lexicon/mwe1.xml               |   3 -
 examples/mwe_lexicon/mwe_manual_example.xml | 156 ++++++++++++++++++++
 resources/schema/bilingual_dictionaries.xsd |   2 +-
 resources/schema/inventory.xsd              |   8 +
 4 files changed, 165 insertions(+), 4 deletions(-)
 delete mode 100644 examples/mwe_lexicon/mwe1.xml
 create mode 100644 examples/mwe_lexicon/mwe_manual_example.xml
diff --git a/examples/mwe_lexicon/mwe1.xml b/examples/mwe_lexicon/mwe1.xml
deleted file mode 100644
index 7cb7f64..0000000
--- a/examples/mwe_lexicon/mwe1.xml
+++ /dev/null
@@ -1,3 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<lexicon/>
-
diff --git a/examples/mwe_lexicon/mwe_manual_example.xml b/examples/mwe_lexicon/mwe_manual_example.xml
new file mode 100644
index 0000000..fdd5e5f
--- /dev/null
+++ b/examples/mwe_lexicon/mwe_manual_example.xml
@@ -0,0 +1,156 @@
+﻿<lexicon>
+    <entry>
+        <head>
+            <headword>
+              <lemma>kdo ni po kisli juhi priplaval</lemma>
+            </headword>
+            <lexicalUnit id="757847" type="MWE" origin="lbs_fe.1012" structure_id="1000504">
+              <component num="1">
+                <lexeme type="argument">kdo</lexeme>
+              </component>
+              <component num="2">
+                <lexeme lemma="biti" msd="Va-r3s-y" sloleks="">ni</lexeme>
+              </component>
+              <component num="3">
+                <lexeme lemma="po" msd="Sl" sloleks="">po</lexeme>
+              </component>
+              <component num="4">
+                <lexeme lemma="kisel" msd="Agpfsl" sloleks="">kisli</lexeme>
+              </component>
+              <component num="5">
+                <lexeme lemma="juha" msd="Ncfsl" sloleks="">juhi</lexeme>
+              </component>
+              <component num="6">
+                <lexeme lemma="priplavati" msd="Vmep-sm" sloleks="">priplaval</lexeme>
+              </component>
+            </lexicalUnit>
+            <measureList>
+                <measure type="raw_frequency" source="Gigafida 2.0">1</measure>
+                <!-- iz Gigafide -->
+            </measureList>
+            <relatedEntryList>
+                <relatedEntry origin="lbs_fe.2194">kdo ni po kisli župi priplaval</relatedEntry>
+            </relatedEntryList>
+        </head>
+        <body>
+            <senseList>
+                <sense id="317">
+                    <definitionList/>
+                    <exampleContainerList>
+                        <exampleContainer>
+                            <corpusExample exampleId="GF.111.11.11">
+                                <comp num="1">Župani</comp>
+                                <comp num="2">so</comp> kljub volitvam povečini na vodilna občinska mesta 
+                                <comp num="6">priplavali</comp>
+                                <comp num="3">po</comp>
+                                <comp num="4">kisli</comp>
+                                <comp num="5">juhi</comp> ali župi, kot jo nezmotljivo nemčimo na
+                                štajerskem območju. </corpusExample>
+                        </exampleContainer>
+                    </exampleContainerList>
+                </sense>
+            </senseList>
+        </body>
+    </entry>
+    <entry>
+        <head>
+            <headword>
+                <lemma>kdo ni po kisli župi priplaval</lemma>
+            </headword>
+            <lexicalUnit id="3243332" type="MWE" origin="lbs_fe.2194" structure_id="1000504">
+              <component num="1">
+                <lexeme type="argument">kdo</lexeme>
+              </component>
+              <component num="2">
+                <lexeme lemma="biti" msd="Va-r3s-y" sloleks="">ni</lexeme>
+              </component>
+              <component num="3">
+                <lexeme lemma="po" msd="Sl" sloleks="">po</lexeme>
+              </component>
+              <component num="4">
+                <lexeme lemma="kisel" msd="Agpfsl" sloleks="">kisli</lexeme>
+              </component>
+              <component num="5">
+                <lexeme lemma="župa" msd="Ncfsl" sloleks="">župi</lexeme>
+              </component>
+              <component num="6">
+                <lexeme lemma="priplavati" msd="Vmep-sm" sloleks="">priplaval</lexeme>
+              </component>
+            </lexicalUnit>
+            <measureList>
+                <measure type="raw_frequency" source="GF2.0">1</measure>
+                <!-- iz Gigafide -->
+            </measureList>
+            <relatedEntryList>
+                <relatedEntry origin="lbs_fe.1012">kdo ni po kisli juhi priplaval</relatedEntry>
+            </relatedEntryList>
+        </head>
+        <body>
+            <senseList>
+                <sense id="317">
+                    <definitionList/>
+                    <exampleContainerList>
+                        <exampleContainer>
+                            <corpusExample exampleId="GF.111.11.11" modified="false">
+                                <comp num="1">Župani</comp>
+                                <comp num="2">so</comp> kljub volitvam povečini na vodilna občinska mesta 
+                                <comp num="6">priplavali</comp>
+                                <comp num="3">po</comp>
+                                <comp num="4">kisli</comp> juhi ali 
+                                <comp num="5">župi</comp>, kot
+                                jo nezmotljivo nemčimo na štajerskem območju. </corpusExample>
+                        </exampleContainer>
+                    </exampleContainerList>
+                </sense>
+            </senseList>
+        </body>
+    </entry>
+    <entry>
+        <head>
+            <headword>
+                <lemma>kdo ni po juhi priplaval</lemma>
+            </headword>
+            <lexicalUnit id="45345435" type="MWE" origin="lbs_fe.1011" structure_id="1000503">
+                <!-- kdo ni po juhi priplaval -->
+                <component num="1">
+                  <lexeme type="argument">kdo</lexeme>
+                </component>
+                <component num="2">
+                  <lexeme lemma="biti" msd="Va-r3s-y" sloleks="">ni</lexeme>
+                </component>
+                <component num="3">
+                  <lexeme lemma="po" msd="Sl" sloleks="">po</lexeme>
+                </component>
+                <component num="4">
+                  <lexeme lemma="juha" msd="Ncfsl" sloleks="">juhi</lexeme>
+                </component>
+                <component num="5">
+                  <lexeme lemma="priplavati" msd="Vmep-sm" sloleks="">priplaval</lexeme>
+                </component>
+            </lexicalUnit>
+            <measureList>
+                <measure type="raw_frequency" source="GF2.0">24</measure>
+                <!-- iz Gigafide -->
+            </measureList>
+        </head>
+        <body>
+            <senseList>
+                <sense id="317">
+                    <definitionList/>
+                    <exampleContainerList>
+                        <exampleContainer>
+                            <corpusExample exampleId="GF.111.11.11">
+                                <comp num="1">Tip</comp> tudi 
+                                <comp num="2">ni</comp>
+                                <comp num="5">priplaval</comp>
+                                <comp num="3">po</comp>
+                                <comp num="4">juhi</comp> in bo hitro razumel, kaj vam gre po
+                                glavi.</corpusExample>
+                        </exampleContainer>
+                        <!-- ID v zgledih -->
+                    </exampleContainerList>
+                </sense>
+            </senseList>
+        </body>
+    </entry>
+</lexicon>
diff --git a/resources/schema/bilingual_dictionaries.xsd b/resources/schema/bilingual_dictionaries.xsd
index 641651d..6e483ab 100644
--- a/resources/schema/bilingual_dictionaries.xsd
+++ b/resources/schema/bilingual_dictionaries.xsd
@@ -14,7 +14,7 @@
               <xsd:element name="head">
                 <xsd:complexType>
                   <xsd:sequence>
-                    <xsd:element name="status" type="statusType"/>
+                    <xsd:element name="status" type="statusType" minOccurs="0"/>
 	            <xsd:element name="headword" type="headwordType"/>
 	            <xsd:element name="lexicalUnit" type="lexicalUnitType" minOccurs="0"/>
                     <xsd:element name="grammar" type="grammarType" minOccurs="0"/>
diff --git a/resources/schema/inventory.xsd b/resources/schema/inventory.xsd
index 9c7fc46..9cc84a5 100644
--- a/resources/schema/inventory.xsd
+++ b/resources/schema/inventory.xsd
@@ -60,8 +60,12 @@
   <xsd:complexType name="lexemeType">
     <xsd:simpleContent>
       <xsd:extension base="xsd:string">
+        <xsd:attribute name="type" type="xsd:string"/>
         <xsd:attribute name="lexical_unit_lexeme_id" type="xsd:int"/>
         <xsd:attribute name="sloleks" type="xsd:string"/>
+        <xsd:attribute name="kol" type="xsd:string"/>
+        <xsd:attribute name="lemma" type="xsd:string"/>
+        <xsd:attribute name="msd" type="xsd:string"/>
       </xsd:extension>
     </xsd:simpleContent>
   </xsd:complexType>
@@ -75,6 +79,8 @@
     </xsd:sequence>
     <xsd:attribute name="id" type="xsd:int" use="required"/>
     <xsd:attribute name="type" type="xsd:string" use="required"/>
+    <xsd:attribute name="structure_id" type="xsd:int"/>
+    <xsd:attribute name="origin" type="xsd:string"/>
   </xsd:complexType>
 
   <xsd:simpleType name="categoryType">
@@ -135,6 +141,7 @@
     <xsd:simpleContent>
       <xsd:extension base="xsd:string">
         <xsd:attribute name="lexical_unit_id" type="xsd:int"/>
+        <xsd:attribute name="origin" type="xsd:string"/>
       </xsd:extension>
     </xsd:simpleContent>
   </xsd:complexType>
@@ -230,6 +237,7 @@
     </xsd:sequence>
     <xsd:attribute name="corpus_id" type="xsd:int"/>
     <xsd:attribute name="example_id" type="xsd:int"/>
+    <xsd:attribute name="exampleId" type="xsd:string"/>
     <xsd:attribute name="modified" type="xsd:boolean"/>
     <xsd:attribute name="lexical_unit_id" type="xsd:int"/>
     <xsd:attribute name="audio" type="xsd:string"/>
