server/src/tests/Libraries/Mappers/CocoonTextRdfMapperTest.php
changeset 152 dd6b3adde73b
parent 143 023b6d467566
child 504 4ab820b387da
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/server/src/tests/Libraries/Mappers/CocoonTextRdfMapperTest.php	Wed Apr 13 13:38:40 2016 +0200
@@ -0,0 +1,291 @@
+<?php
+
+use CorpusParole\Libraries\Mappers\CocoonTextRdfMapper;
+use CorpusParole\Libraries\CocoonUtils;
+
+use EasyRdf\Graph;
+use EasyRdf\Resource;
+use EasyRdf\Literal;
+
+use Illuminate\Foundation\Testing\WithoutMiddleware;
+use Illuminate\Foundation\Testing\DatabaseMigrations;
+use Illuminate\Foundation\Testing\DatabaseTransactions;
+
+class CocoonTextRdfMapperTest extends TestCase
+{
+
+    const TEST_INPUT_DOCS = [
+        'BASE' => <<<EOT
+        @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+        @prefix owl: <http://www.w3.org/2002/07/owl#> .
+        @prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+        @prefix fn: <http://www.w3.org/2005/xpath-functions#> .
+        @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
+        @prefix sesame: <http://www.openrdf.org/schema/sesame#> .
+        @prefix v: <http://rdf.data-vocabulary.org/#> .
+        <http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-UVE_MOCIKA> a <http://crdo.risc.cnrs.fr/schemas/Resource> ;
+            <http://purl.org/dc/elements/1.1/description> "Voilà pourquoi le bernard-l'hermite, aujourd'hui, se cache dans les coquilles vides qu'il trouve, alors que le crabe de cocotier n'a pas honte de se promener tout nu."@fr ;
+            <http://purl.org/dc/elements/1.1/format> "text/xml"^^<http://purl.org/dc/terms/IMT> ;
+            <http://purl.org/dc/elements/1.1/type> "primary_text"^^<http://www.language-archives.org/OLAC/1.1/linguistic-type> , <http://purl.org/dc/dcmitype/Text> , "narrative"^^<http://www.language-archives.org/OLAC/1.1/discourse-type> ;
+            <http://purl.org/dc/elements/1.1/subject> <http://lexvo.org/id/iso639-3/uve> ;
+            <http://purl.org/dc/elements/1.1/language> <http://lexvo.org/id/iso639-3/uve> ;
+            <http://www.language-archives.org/OLAC/1.1/depositor> <http://viaf.org/viaf/56614135> ;
+            <http://purl.org/dc/elements/1.1/publisher> <http://viaf.org/viaf/154919513> ;
+            <http://purl.org/dc/elements/1.1/rights> "Copyright (c) Moyse-Faurie, Claire" ;
+            <http://purl.org/dc/elements/1.1/title> "The two hermit crabs and the coconut crab"@en ;
+            <http://purl.org/dc/terms/isFormatOf> <http://cocoon.huma-num.fr/exist/crdo/moyse-faurie/uve/crdo-UVE_MOCIKA.xml> , <http://cocoon.huma-num.fr/exist/crdo/moyse-faurie/uve/crdo-UVE_MOCIKA.xhtml> ;
+            <http://purl.org/dc/terms/accessRights> "Freely available for non-commercial use" ;
+            <http://purl.org/dc/terms/license> <http://creativecommons.org/licenses/by-nc-sa/2.5/> ;
+            <http://purl.org/dc/terms/isPartOf> <http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-COLLECTION_LANGUESDEFRANCE> , <http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-COLLECTION_LACITO> ;
+            <http://purl.org/dc/terms/spatial> "NC"^^<http://purl.org/dc/terms/ISO3166> , "New Caledonia, Ohnyat (Ouvéa)" ;
+            <http://purl.org/dc/terms/available> "2011-02-05"^^<http://purl.org/dc/terms/W3CDTF> ;
+            <http://purl.org/dc/terms/issued> "2011-02-05T23:22:23+01:00"^^<http://purl.org/dc/terms/W3CDTF> ;
+            <http://www.language-archives.org/OLAC/1.1/speaker> "Idakote, Félicien" ;
+            <http://purl.org/dc/terms/requires> <http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-UVE_MOCIKA_SOUND> ;
+            <http://purl.org/dc/terms/alternative> "Les deux bernard-l'hermite et le crabe de cocotier"@fr ;
+            <http://www.language-archives.org/OLAC/1.1/researcher> "Moyse-Faurie, Claire" ;
+            <http://purl.org/dc/terms/modified> "2002-02-20"^^<http://purl.org/dc/terms/W3CDTF> ;
+            <http://purl.org/dc/terms/conformsTo> <http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-dtd_archive> .
+EOT
+    ];
+
+    const TEST_INPUT_ID = "crdo-UVE_MOCIKA_SOUND";
+    const TEST_CHO_URI = "http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-UVE_MOCIKA_SOUND";
+
+
+    private $inputGraphes = [];
+    private $resGraphes = [];
+    private $mappers = [];
+
+    function __construct(string $name = null) {
+        parent::__construct($name);
+    }
+
+    public function setUp() {
+        parent::setUp();
+
+        $this->testGraphUri = config('corpusparole.corpus_doc_id_base_uri').CocoonTextRdfMapperTest::TEST_INPUT_ID;
+        foreach(CocoonTextRdfMapperTest::TEST_INPUT_DOCS as $key => $inputDoc) {
+            $this->inputGraphes[$key] = new Graph("http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-UVE_MOCIKA", $inputDoc);
+            $this->mappers[$key] = new CocoonTextRdfMapper($this->inputGraphes[$key]);
+            $this->mappers[$key]->mapGraph();
+            $this->resGraphes[$key] = $this->mappers[$key]->getOutputGraphes()["http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-UVE_MOCIKA_SOUND"];
+        }
+    }
+
+    public function tearDown() {
+        parent::tearDown();
+    }
+
+    /**
+     * Just test that the construct and setup are ok
+     *
+     * @return void
+     */
+    public function testInit()
+    {
+        $this->assertTrue(true);
+    }
+
+
+    /**
+     * Test that the returned graph has the same uri that the original.
+     *
+     * @return void
+     */
+    public function testUri() {
+
+        //echo $this->resGraphes['BASE']->serialise('turtle');
+        //echo var_export($this->resGraphes['BASE']->toRdfPhp());
+
+        $this->assertNotNull($this->resGraphes['BASE']);
+        $this->assertEquals($this->testGraphUri, $this->resGraphes['BASE']->getUri(), "URIS must be translated");
+    }
+
+    /**
+     * Test that the return graph has one ore:Aggregation resource
+     *
+     * @return void
+     */
+    public function testAggregationResource() {
+
+        $resources = $this->resGraphes['BASE']->allOfType('ore:Aggregation');
+
+        $this->assertCount(1, $resources, "Must found only one resources of type ore:Aggregation");
+        $this->assertEquals($this->testGraphUri,$resources[0]->getUri());
+    }
+
+    /**
+     * Test providedCHO identifier.
+     *
+     * @return void
+     */
+    public function testProvidedChoId() {
+        $resources = $this->resGraphes['BASE']->allOfType('edm:ProvidedCHO');
+
+        $this->assertCount(1, $resources, "Must found only one resources of type edm:ProvidedCHO");
+        $providedCHO = $resources[0];
+
+        $identifier = $providedCHO->getLiteral('<http://purl.org/dc/elements/1.1/identifier>');
+
+        $this->assertInstanceOf('EasyRdf\Literal', $identifier, "identifier value must be a literal");
+        $this->assertEquals(config('corpusparole.corpus_id_scheme') . CocoonTextRdfMapperTest::TEST_INPUT_ID, $identifier->getValue(), "Value must be equals to ".config('corpusparole.corpus_id_scheme').CocoonTextRdfMapperTest::TEST_INPUT_ID);
+
+    }
+
+    public function testRightsWebResourceIdentity() {
+        $properties = [
+            "http://purl.org/dc/elements/1.1/rights",
+            'http://purl.org/dc/terms/license',
+        ];
+
+        $webResourceList = $this->resGraphes['BASE']->all('edm:WebResource', '^rdf:type');
+        $sourceNode = $this->inputGraphes['BASE']->get('http://crdo.risc.cnrs.fr/schemas/Resource', '^rdf:type');
+
+        foreach ($webResourceList as $webResource) {
+            foreach ($properties as $prop) {
+                $outputValuesStr = [];
+                foreach($webResource->all($this->resGraphes['BASE']->resource($prop)) as $outputValue) {
+                    array_push($outputValuesStr, strval($outputValue));
+                }
+                $this->assertNotEmpty($outputValuesStr, "we must found some values to test $prop");
+                foreach ($sourceNode->all($this->inputGraphes['BASE']->resource($prop)) as $value) {
+                    $this->assertContains(strval($value), $outputValuesStr, "$prop not found in output graph");
+                }
+            }
+        }
+    }
+
+    public function testRightsWebResourceTrim() {
+        $properties = [
+            'http://purl.org/dc/terms/accessRights',
+        ];
+
+        $webResourceList = $this->resGraphes['BASE']->all('edm:WebResource', '^rdf:type');
+        $sourceNode = $this->inputGraphes['BASE']->get('http://crdo.risc.cnrs.fr/schemas/Resource', '^rdf:type');
+
+        foreach ($webResourceList as $webResource) {
+            foreach ($properties as $prop) {
+                $outputValuesStr = [];
+                foreach($webResource->all($this->resGraphes['BASE']->resource($prop)) as $outputValue) {
+                    array_push($outputValuesStr, strval($outputValue));
+                }
+                $this->assertNotEmpty($outputValuesStr, "we must found some values to test $prop");
+                foreach ($sourceNode->all($this->inputGraphes['BASE']->resource($prop)) as $value) {
+                    $this->assertContains(trim(strval($value)), $outputValuesStr, "$prop not found in output graph");
+                }
+            }
+        }
+    }
+
+
+    /**
+     * Test providedCHO uri.
+     *
+     * @return void
+     */
+    public function testProvidedChoURI() {
+        $resources = $this->resGraphes['BASE']->allOfType('edm:ProvidedCHO');
+
+        $this->assertCount(1, $resources, "Must found only one resources of type edm:ProvidedCHO");
+        $this->assertEquals(CocoonTextRdfMapperTest::TEST_CHO_URI,$resources[0]->getUri());
+    }
+
+
+
+    /**
+     * Test that the returned graph does not have a http://purl.org/dc/dcmitype/Sound type
+     *
+     * @return void
+     */
+    public function testType() {
+        //"primary_text"^^<http://www.language-archives.org/OLAC/1.1/linguistic-type> , <http://purl.org/dc/dcmitype/Text> , "narrative"^^<http://www.language-archives.org/OLAC/1.1/discourse-type> ;
+        $providedCHO = $this->resGraphes['BASE']->get('edm:ProvidedCHO', '^rdf:type');
+
+        $this->assertNotNull($providedCHO);
+
+        $this->assertcount(3, $providedCHO->all($this->resGraphes['BASE']->resource('http://purl.org/dc/elements/1.1/type')), 'but should find 2 literals');
+
+        $dcTypeList = $providedCHO->all($this->resGraphes['BASE']->resource('http://purl.org/dc/elements/1.1/type'));
+        foreach ($dcTypeList as $dcType) {
+            $this->assertThat(
+                $dcType,
+                $this->logicalXor(
+                    $this->isInstanceOf(EasyRdf\Literal::class),
+                    $this->isInstanceOf(EasyRdf\Resource::class)
+                )
+            );
+        }
+    }
+
+     /**
+      * Test that the web resources
+      *
+      * @return void
+      */
+    public function testWebResources() {
+
+        $resources = $this->resGraphes['BASE']->allOfType('edm:WebResource');
+
+        $this->assertCount(2, $resources, "Must found three webresource");
+
+        $aggregation = $this->resGraphes['BASE']->resource($this->testGraphUri);
+
+        foreach ($resources as $wres) {
+            $mimetypes = $wres->all($this->resGraphes['BASE']->resource('http://purl.org/dc/elements/1.1/format'));
+            $this->assertCount(1, $mimetypes, "Must find one mimetype.");
+            $mimetype = $mimetypes[0];
+            $this->assertInstanceOf("EasyRdf\Literal", $mimetype, "mimetype must be literal");
+            $this->assertEquals("dc:IMT",$mimetype->getDatatype());
+        }
+    }
+
+     /**
+      * Test that the web resources has good date
+      *
+      * @return void
+      */
+    public function testWebResourcesDate() {
+
+        $resources = $this->resGraphes['BASE']->allOfType('edm:WebResource');
+
+        foreach ($resources as $wres) {
+            $this->assertFalse($wres->hasProperty("http://purl.org/dc/terms/available"),"web resource must not have http://purl.org/dc/terms/available");
+            $this->assertFalse($wres->hasProperty("http://purl.org/dc/terms/modified"),"web resource must not have http://purl.org/dc/terms/modified");
+            $this->assertTrue($wres->hasProperty("http://purl.org/dc/terms/issued"), "Must have http://purl.org/dc/terms/issued");
+
+            $issued = $wres->getLiteral('dc:issued');
+            //<http://purl.org/dc/terms/issued> "2014-12-05T15:00:19+01:00"^^<http://purl.org/dc/terms/W3CDTF> ;
+            $this->assertInstanceOf('EasyRdf\Literal', $issued, "issued value must be a literal");
+            $this->assertEquals('http://purl.org/dc/terms/W3CDTF', $issued->getDatatypeUri(), "issued datatype uri must be a http://purl.org/dc/terms/W3CDTF");
+            $this->assertEquals('2011-02-05T23:22:23+01:00', $issued->getValue(), "Value must be 2014-12-05T15:00:19+01:00");
+
+        }
+
+    }
+
+
+    /**
+     * Test that the web resources as conformsTo property
+     *
+     * @return void
+     */
+   public function testWebResourcesConformsTo() {
+
+       $resources = $this->resGraphes['BASE']->allOfType('edm:WebResource');
+       foreach ($resources as $wres) {
+
+           $mimetypes = $wres->all($this->resGraphes['BASE']->resource('http://purl.org/dc/elements/1.1/format'));
+           $mimetype = $mimetypes[0]->getValue();
+           if($mimetype !== 'application/xml') {
+               continue;
+           }
+           $conformsTo = $wres->get('<http://purl.org/dc/terms/conformsTo>');
+           $this->assertNotNull($conformsTo, "Must have a conformsTo node");
+           $this->assertInstanceOf('EasyRdf\Resource', $conformsTo, "Must be a resource");
+           $this->assertEquals("http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-dtd_archive", $conformsTo->getUri(), "Must be http://purl.org/poi/crdo.vjf.cnrs.fr/crdo-dtd_archive");
+       }
+   }
+
+
+}