Empty URLs have been incorrectly resolved in the downloaded turtle files

[isotes]

It seems that empty URLs are in some cases resolved to the encompassing rdf file leading to incorrect (and at least for Münster) malformed entries in the turtle files.

Example from Münster_(Westfalen).rdf:

<https://opendata.stadt-muenster.de/dataset/sporthallen-und-sportst%C3%A4tten-standorte/resource/96e271af-7e05-4c2e-9406-17a3535e88a2>
        a                    cat:Distribution ;
        dcterms:description  "" ;
        dcterms:format       "wms" ;
        dcterms:issued       "2019-07-01T17:33:24+02:00"^^xsd:date ;
        dcterms:modified     "2019-07-18T11:21:22+02:00"^^xsd:date ;
        dcterms:title        "Sporthallen und Sportstätten - Standorte - WMS-GetMap" ;
        cat:accessURL        <https://opendata.stadt-muenster.de/dataset/sporthallen-und-sportst%C3%A4tten-standorte/resource/96e271af-7e05-4c2e-9406-17a3535e88a2> ;
        cat:byteSize         "" ;
        cat:downloadURL      <file:///home/lisa/repos/crawling/target/Münster_(Westfalen).rdf> ;
        cat:mediaType        "" ;
        foaf:page            "https://opendata.stadt-muenster.de/dataset/sporthallen-und-sportst%C3%A4tten-standorte/resource/96e271af-7e05-4c2e-9406-17a3535e88a2" .

The cat:downloadURL <file:///home/lisa/repos/crawling/target/Münster_(Westfalen).rdf> ; is incorrect and malformed ('ü').

Looking at the catalog entry on the website it should be empty: <dcat:downloadURL rdf:resource=""/>

Grepping for 'home/lisa' in catalog/toLoad leads to 2948 results for various fields (at least dcat:accessURL, dcat:downloadURL, and vcard:hasURL). I did not check if the reason is always an empty URL in the original data.

[Aklakan]

So there are multiple problems:

• The RDF of https://opendata.stadt-muenster.de/catalog.xml is invalid as it contains

<dcat:downloadURL rdf:resource="<div class="field field-name-field-link-remote-file field-type-file field-label-hidden"><div class="field-items"><div class="field-item even">https://www.stadt-muenster.de/ows/mapserv706/odgruenserv?REQUEST=GetFeature&SERVICE=WFS&VERSION=1.0.0&TYPENAME=ms:Gruenflaechen&OUTPUTFORMAT=CSV&EXCEPTIONS=XML&MAXFEATURES=70000</div></div></div>"/>

• Sometimes there are also empty IRIs i.e. <> - we need to ensure to resolve relative IRIs against the URL from where it was retrieved
• sparql-integrate might not set the base url correctly when downloading RDF from the web - needs to be checked