Skip to content

Commit

Permalink
Updated tests for cmdi teiHeader
Browse files Browse the repository at this point in the history 
And new tei merge test case
  • Loading branch information
@PrinsINT
PrinsINT committed Jun 19, 2024
1 parent 4ad80de commit 912e257
Show file tree
Hide file tree
Showing 11 changed files with 505 additions and 373 deletions.
2 changes: 1 addition & 1 deletion server/src/test/kotlin/org/ivdnt/galahad/port/Util.kt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -311,7 +311,7 @@ class TestResult(
) {

fun ignoreDate(): TestResult {
val date = Regex("\\d{4}-\\d{2}-\\d{2} \\d{2}:\\d{2}:\\d{2}")
val date = Regex("\\d{4}-\\d{2}-\\d{2}")

actual = date.replace(actual, "__DATE_IGNORED_BY_TEST__")
expected = date.replace(expected, "__DATE_IGNORED_BY_TEST__")
Expand Down
135 changes: 83 additions & 52 deletions server/src/test/resources/all-formats/output/from-Conllu-to-TeiP5.tei.xml
View file
Original file line number Diff line number Diff line change
@@ -1,57 +1,88 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:tei="http://www.tei-c.org/ns/1.0" xmlns:xs="http://www.w3.org/2001/XMLSchema">
<teiHeader>
<fileDesc>
<titleStmt>
<title>input</title>
</titleStmt>
<publicationStmt>
<p>
<date>2024-03-13 14:23:35</date>
<idno type="sourceID">What if source id is undefined?</idno>
<idno type="pid">What if no pid is defined?</idno>
</p>
</publicationStmt>
<sourceDesc>
<listBibl xml:id="inlMetadata">
<bibl>
<interpGrp type="questions">
<interp>Welke inl data?</interp>
<interp>wat als er geen is?</interp>
</interpGrp>
</bibl>
</listBibl>
<listBibl xml:id="galahadMetadata">
<bibl>
<interpGrp type="model">
<interp>pie-tdn-all</interp>
</interpGrp>
<interpGrp type="tagset">
<interp>TDN-Core</interp>
</interpGrp>
<interpGrp type="corpus_name">
<interp>testCorpus</interp>
</interpGrp>
<interpGrp type="corpus_uuid">
<interp>059197e6-a1c9-45b3-9428-9fd28e1e8e32</interp>
</interpGrp>
<interpGrp type="date">
<interp>2024-03-13 14:23:35</interp>
</interpGrp>
<interpGrp type="user_id">
<interp>testUser</interp>
</interpGrp>
<interpGrp type="pid">
<interp>d9527977-7a1f-4f93-b19b-5c26c93afea6</interp>
</interpGrp>
<interpGrp type="title">
<interp>input</interp>
</interpGrp>
</bibl>
</listBibl>
</sourceDesc>
</fileDesc>
</teiHeader>
<teiHeader>
<fileDesc>
<titleStmt>
<title>input</title>
<respStmt>
<resp>linguistic annotation by GaLAHaD (https://portal.clarin.ivdnt.org/galahad)</resp>
<orgName xml:lang="nl">Instituut voor de Nederlandse Taal</orgName>
<orgName xml:lang="en">Dutch Language Institute</orgName>
</respStmt>
<respStmt>
<resp>exported as tei-p5 by GaLAHaD (https://portal.clarin.ivdnt.org/galahad)</resp>
<orgName xml:lang="nl">Instituut voor de Nederlandse Taal</orgName>
<orgName xml:lang="en">Dutch Language Institute</orgName>
</respStmt>
</titleStmt>
<publicationStmt>
<publisher>!Needs to be filled in!</publisher>
<idno type="sourceID">input</idno>
<idno type="GaLAHaDPersistentIdentifier">__UUID_IGNORED_BY_TEST___tei</idno>
</publicationStmt>
<notesStmt>
<note resp="GaLAHaD" type="corpusName">testCorpus</note>
<note resp="GaLAHaD" type="sourceCollection">source name</note>
<note resp="GaLAHaD" type="sourceCollectionURL">http://source.url</note>
</notesStmt>
<sourceDesc>
<ab>
<idno type="sourceID">input</idno>
</ab>
<ab type="date">
<date from="0" to="0"/>
</ab>
</sourceDesc>
</fileDesc>
<encodingDesc>
<appInfo resp="GaLAHaD">
<application ident="pie-tdn-all" version="1.0.1" xml:id="pie-tdn-all">
<label>POS-tagger and lemmatiser</label>
<ptr target="https://github.com/INL/galahad-taggers-dockerized/tree/1.0.2/pie/TDN-ALL"/>
</application>
</appInfo>
<editorialDecl resp="GaLAHaD">
<interpretation xml:id="A0001">
<ab subtype="POS-tagging_lemmatisation" type="linguisticAnnotation">
<interpGrp type="annotationStyle">
<interp>inline</interp>
</interpGrp>
<interpGrp type="Documentation">
<interp/>
</interpGrp>
<interpGrp type="annotationSet">
<interp>TDN-Core</interp>
</interpGrp>
<interpGrp type="annotationDescription">
<interp>The file was automatically annotated within the platform GaLAHaD, which is a central hub for enriching historical Dutch.</interp>
</interpGrp>
<interpGrp type="annotationFormat">
<interp>TEI xml</interp>
</interpGrp>
</ab>
<ab subtype="POS-tagging_lemmatisationProvenance1" type="linguisticAnnotation">
<interpGrp type="annotationMode">
<interp>automatically annotated</interp>
</interpGrp>
<interpGrp type="processor">
<interp sameAs="#pie-tdn-all"/>
</interpGrp>
<date from="__DATE_IGNORED_BY_TEST__" to="__DATE_IGNORED_BY_TEST__"/>
</ab>
</interpretation>
</editorialDecl>
</encodingDesc>
<profileDesc>
<langUsage>
<language ident="nl">
Dutch
<interpGrp type="dominantLanguage">
<interp>true</interp>
</interpGrp>
</language>
</langUsage>
</profileDesc>
</teiHeader>
<text>
<body>
<div>
Expand Down
135 changes: 83 additions & 52 deletions server/src/test/resources/all-formats/output/from-Folia-to-TeiP5.tei.xml
View file
Original file line number Diff line number Diff line change
@@ -1,57 +1,88 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:tei="http://www.tei-c.org/ns/1.0" xmlns:xs="http://www.w3.org/2001/XMLSchema">
<teiHeader>
<fileDesc>
<titleStmt>
<title>input.folia</title>
</titleStmt>
<publicationStmt>
<p>
<date>2024-03-13 13:47:56</date>
<idno type="sourceID">What if source id is undefined?</idno>
<idno type="pid">What if no pid is defined?</idno>
</p>
</publicationStmt>
<sourceDesc>
<listBibl xml:id="inlMetadata">
<bibl>
<interpGrp type="questions">
<interp>Welke inl data?</interp>
<interp>wat als er geen is?</interp>
</interpGrp>
</bibl>
</listBibl>
<listBibl xml:id="galahadMetadata">
<bibl>
<interpGrp type="model">
<interp>pie-tdn-all</interp>
</interpGrp>
<interpGrp type="tagset">
<interp>TDN-Core</interp>
</interpGrp>
<interpGrp type="corpus_name">
<interp>testCorpus</interp>
</interpGrp>
<interpGrp type="corpus_uuid">
<interp>3308d415-15b9-4f10-bd10-ddbe605caf91</interp>
</interpGrp>
<interpGrp type="date">
<interp>2024-03-13 13:47:56</interp>
</interpGrp>
<interpGrp type="user_id">
<interp>testUser</interp>
</interpGrp>
<interpGrp type="pid">
<interp>17a41382-9c06-46be-b5e5-7fc86b4dc1bb</interp>
</interpGrp>
<interpGrp type="title">
<interp>input.folia</interp>
</interpGrp>
</bibl>
</listBibl>
</sourceDesc>
</fileDesc>
</teiHeader>
<teiHeader>
<fileDesc>
<titleStmt>
<title>input.folia</title>
<respStmt>
<resp>linguistic annotation by GaLAHaD (https://portal.clarin.ivdnt.org/galahad)</resp>
<orgName xml:lang="nl">Instituut voor de Nederlandse Taal</orgName>
<orgName xml:lang="en">Dutch Language Institute</orgName>
</respStmt>
<respStmt>
<resp>exported as tei-p5 by GaLAHaD (https://portal.clarin.ivdnt.org/galahad)</resp>
<orgName xml:lang="nl">Instituut voor de Nederlandse Taal</orgName>
<orgName xml:lang="en">Dutch Language Institute</orgName>
</respStmt>
</titleStmt>
<publicationStmt>
<publisher>!Needs to be filled in!</publisher>
<idno type="sourceID">input.folia</idno>
<idno type="GaLAHaDPersistentIdentifier">__UUID_IGNORED_BY_TEST___tei</idno>
</publicationStmt>
<notesStmt>
<note resp="GaLAHaD" type="corpusName">testCorpus</note>
<note resp="GaLAHaD" type="sourceCollection">source name</note>
<note resp="GaLAHaD" type="sourceCollectionURL">http://source.url</note>
</notesStmt>
<sourceDesc>
<ab>
<idno type="sourceID">input.folia</idno>
</ab>
<ab type="date">
<date from="0" to="0"/>
</ab>
</sourceDesc>
</fileDesc>
<encodingDesc>
<appInfo resp="GaLAHaD">
<application ident="pie-tdn-all" version="1.0.1" xml:id="pie-tdn-all">
<label>POS-tagger and lemmatiser</label>
<ptr target="https://github.com/INL/galahad-taggers-dockerized/tree/1.0.2/pie/TDN-ALL"/>
</application>
</appInfo>
<editorialDecl resp="GaLAHaD">
<interpretation xml:id="A0001">
<ab subtype="POS-tagging_lemmatisation" type="linguisticAnnotation">
<interpGrp type="annotationStyle">
<interp>inline</interp>
</interpGrp>
<interpGrp type="Documentation">
<interp/>
</interpGrp>
<interpGrp type="annotationSet">
<interp>TDN-Core</interp>
</interpGrp>
<interpGrp type="annotationDescription">
<interp>The file was automatically annotated within the platform GaLAHaD, which is a central hub for enriching historical Dutch.</interp>
</interpGrp>
<interpGrp type="annotationFormat">
<interp>TEI xml</interp>
</interpGrp>
</ab>
<ab subtype="POS-tagging_lemmatisationProvenance1" type="linguisticAnnotation">
<interpGrp type="annotationMode">
<interp>automatically annotated</interp>
</interpGrp>
<interpGrp type="processor">
<interp sameAs="#pie-tdn-all"/>
</interpGrp>
<date from="__DATE_IGNORED_BY_TEST__" to="__DATE_IGNORED_BY_TEST__"/>
</ab>
</interpretation>
</editorialDecl>
</encodingDesc>
<profileDesc>
<langUsage>
<language ident="nl">
Dutch
<interpGrp type="dominantLanguage">
<interp>true</interp>
</interpGrp>
</language>
</langUsage>
</profileDesc>
</teiHeader>
<text>
<body>
<div>
Expand Down
Loading

0 comments on commit 912e257

Please sign in to comment.