Initial commit

publiccode.yml

+publiccodeYmlVersion: '0.2'
+categories:
+  - metadata-validation
+  - data-quality-tools
+description:
+  de:
+    documentation: 'https://code.schleswig-holstein.de/opendata/dcat-catalog-check#readme'
+    features:
+      - Überprüft die Integrität von Distributionen in einem DCAT-Katalog
+      - Prüft, ob URLs aktiv sind und Dateien im richtigen Format vorliegen
+      - Führt spezifische Prüfungen für verschiedene Formate durch
+    genericName: DCAT Katalogüberprüfung
+    longDescription: |2
+      Der DCAT Catalog Check ist ein Python-Skript, das entwickelt wurde, um in einem DCAT-Katalog verzeichnete Distributionen (Dateien) zu überwachen und zu validieren. 
+      Es überprüft, ob die URLs von Distributionen aktiv sind, ob Dateien korrekt formatiert sind und ob die MIME-Typen mit den Metadaten übereinstimmen. 
+      Das Werkzeug unterstützt zahlreiche spezielle Formatvalidierungen wie GeoJSON, XML und PDF und protokolliert die Ergebnisse in einer Logdatei.
+      Außerdem wird die Validierung von Frictionless Data Packages und Frictionless Data Resources unterstützt.
+    shortDescription: |-
+      Python-Skript zur Validierung von Links und Formaten in einem DCAT-Katalog.
+  en:
+    features:
+      - Checks the integrity of distributions in a DCAT catalog
+      - Validates whether URLs are alive and files are correctly formatted
+      - Performs specific checks for various formats
+    genericName: DCAT catalog check
+    longDescription: |2
+      The DCAT Catalog Check is a Python script designed to monitor and validate links in a DCAT catalog. 
+      It ensures that resource URLs are active, files are properly formatted, and MIME types match the metadata. 
+      The tool supports various specialized format checks such as GeoJSON, XML, and PDF, logging the results for user review.
+      The validation of Frictionless Data Packages and Frictionless Data Resources is also supported.
+    shortDescription: |-
+      Python script to validate links and formats in a DCAT catalog.
+developmentStatus: stable
+intendedAudience:
+  scope:
+    - data-management
+legal:
+  license: EUPL
+  mainCopyrightOwner: Land Schleswig-Holstein
+  repoOwner: Land Schleswig-Holstein
+localisation:
+  availableLanguages:
+    - en
+  localisationReady: false
+maintenance:
+  contacts:
+    - email: opendata@lr.landsh.de
+      name: Open-Data-Leitstelle
+  type: internal
+name: DCAT Catalog Check
+platforms:
+  - linux
+  - windows
+releaseDate: '2024-12-19'
+softwareType: standalone/other
+softwareVersion: 1.0.0
+url: 'https://code.schleswig-holstein.de/opendata/dcat-catalog-check'
+usedBy:
+  - Open-Data-Portal Schleswig-Holstein

pyproject.toml

+[tool.poetry]
+
+name = "dcat-catalog-check"
+version = "1.0.0"
+description = "DCAT Catalog Check"
+authors = [
+  "Jesper Zedlitz <jesper.zedlitz@stk.landsh.de>",
+  "Thorge Petersen <petersen@rz.uni-kiel.de>",
+]
+license = "AGPL-3.0-or-later"
+readme = "README.md"
+
+[tool.poetry.dependencies]
+python = "^3.10"
+requests = "^2.32.3"
+python-magic = "^0.4.27"
+frictionless = "^5.18.0"
+geopandas = "^1.0.1"
+rdflib = "^7.1.1"
+pypdf = "^5.1.0"
+pillow = "^11.0.0"
+fiona = "^1.10.1"
+pyarrow = "^18.1.0"
+
+[tool.poetry.group.dev.dependencies]
+coverage = "^7.6.1"
+ruff = "^0.6.8"
+
+[build-system]
+requires = ["poetry-core"]
+build-backend = "poetry.core.masonry.api"

resources/file_types.json

+{
+  "7Z": [
+    "application/x-7z-compressed"
+  ],
+  "ARCINFO_COV": [
+    ""
+  ],
+  "ATOM": [
+    "text/xml", "application/xml"
+  ],
+  "CSV": [
+    "text/csv", "text/plain", "application/csv"
+  ],
+  "DOC": [
+    "application/msword"
+  ],
+  "DWG": [
+    "image/vnd.dwg"
+  ],
+  "GDB": [
+    "application/zip"
+  ],
+  "GEOJSON": [
+    "application/json"
+  ],
+  "GEOTIFF": [
+    "image/tiff"
+  ],
+  "GML": [
+    "text/xml", "application/xml"
+  ],
+  "GPKG": [
+    "application/zip"
+  ],
+  "GTFS": [
+    "application/zip"
+  ],
+  "HTML": [
+    "text/html"
+  ],
+  "JPEG": [
+    "image/jpeg"
+  ],
+  "JSON": [
+    "application/json"
+  ],
+  "JSON_LD": [
+    "application/json"
+  ],
+  "MP3": [
+    "audio/mpeg"
+  ],
+  "MPEG4": [
+    "video/mp4"
+  ],
+  "MXD": [
+    "application/CDFV2", "application/x-ole-storage"
+  ],
+  "ODS": [
+    "application/vnd.oasis.opendocument.spreadsheet"
+  ],
+  "PARQUET": [
+    "application/octet-stream"
+  ],
+  "PDF": [
+    "application/pdf"
+  ],
+  "PNG": [
+    "image/png"
+  ],
+  "QGS": [
+    "application/zip"
+  ],
+  "RDF": [
+    "text/xml", "application/xml"
+  ],
+  "SHP": [
+    "application/zip"
+  ],
+  "TIFF": [
+    "image/tiff"
+  ],
+  "TXT": [
+    "text/plain"
+  ],
+  "WFS_SRVC": [
+    "text/xml", "application/xml"
+  ],
+  "WMS_SRVC": [
+    "text/xml", "application/xml"
+  ],
+  "WMTS_SRVC": [
+    "text/xml", "application/xml"
+  ],
+  "XLS": [
+    "application/vnd.ms-excel", "application/CDFV2"
+  ],
+  "XLSX": [
+    "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+  ],
+  "XML": [
+    "text/xml", "application/xml"
+  ],
+  "XSD": [
+    "text/xml", "application/xml"
+  ],
+  "ZIP": [
+    "application/zip"
+  ]
+}

tests/data/bermuda.geojson

+{"type":"FeatureCollection","features": [{"type":"Feature","geometry":{"type":"Polygon","coordinates": [ [ [-64.8,32.3],[-65.5,18.3],[-80.3,25.2],[-64.8,32.3] ] ]},"properties":{ "name" : "Bermuda triangle"}}]}

tests/data/bermuda.gml

+<?xml version="1.0" encoding="utf-8" ?>
+<ogr:FeatureCollection
+     gml:id="aFeatureCollection"
+     xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+     xmlns:ogr="http://ogr.maptools.org/"
+     xmlns:gml="http://www.opengis.net/gml/3.2">
+  <gml:boundedBy><gml:Envelope srsName="urn:ogc:def:crs:EPSG::4326"><gml:lowerCorner>18.3 -80.3</gml:lowerCorner><gml:upperCorner>32.3 -64.8</gml:upperCorner></gml:Envelope></gml:boundedBy>
+                                                                                                                                                                
+  <ogr:featureMember>
+    <ogr:bermuda gml:id="bermuda.0">
+      <gml:boundedBy><gml:Envelope srsName="urn:ogc:def:crs:EPSG::4326"><gml:lowerCorner>18.3 -80.3</gml:lowerCorner><gml:upperCorner>32.3 -64.8</gml:upperCorner></gml:Envelope></gml:boundedBy>
+      <ogr:geometryProperty><gml:Polygon srsName="urn:ogc:def:crs:EPSG::4326" gml:id="bermuda.geom.0"><gml:exterior><gml:LinearRing><gml:posList>32.3 -64.8 18.3 -65.5 25.2 -80.3 32.3 -64.8</gml:posList></gml:LinearRing></gml:exterior></gml:Polygon></ogr:geometryProperty>
+      <ogr:name>Bermuda triangle</ogr:name>
+    </ogr:bermuda>
+  </ogr:featureMember>
+</ogr:FeatureCollection>

tests/data/correct.json

+{
+  "a": 1234
+}

tests/data/correct.xml

+<xml>
+  <data></data>
+</xml>

tests/data/incorrect.json

+{
+a: b
+}

tests/data/incorrect.xml

+<xml>