Merge pull request #31 from Kurrawong/master

Convert Excel and update

scripts/_requirements.txt

@@ -0,0 +1,2 @@
+pyshacl
+httpx

scripts/add_org_rdf.py

@@ -0,0 +1,23 @@
+import glob
+from rdflib import Graph
+from rdflib.namespace import SDO
+
+for f in glob.glob("/Users/nick/Work/ga/ga-vocabs/vocabularies/*.ttl"):
+    g = Graph().parse(f)
+    addition = """
+        PREFIX sdo: <https://schema.org/>
+        PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
+        
+        
+        <https://linked.data.gov.au/org/ga>
+            a sdo:Organization ;
+            sdo:description "Geoscience Australia is Australia's pre-eminent public sector geoscience organisation. It is the nation's trusted advisor on the geology and geography of Australia. It applies science and technology to describe and understand the Earth for the benefit of Australia."@en ;
+            sdo:name "Geoscience Australia" ;
+            sdo:url "https://www.ga.gov.au"^^xsd:anyURI ;
+        .    
+        """
+    g += Graph().parse(data=addition, format="turtle")
+    g.bind("sdo", SDO)
+    g.serialize(destination=f, format="longturtle")
+
+    print(f"Done {f}")

scripts/process_043_excel.sh

@@ -0,0 +1,15 @@
+
+# /Users/nick/Library/Caches/pypoetry/virtualenvs/vocexcel-XgwzAsEM-py3.11/bin/python /Users/nick/Work/rdflib/VocExcel/vocexcel/__main__.py Borehole_Construction_Type_Boreholes.xlsx > Borehole_Construction_Type_Boreholes.ttl
+# echo "Borehole_Construction_Type_Boreholes"
+
+
+FILES=/Users/nick/Work/ga/ga-vocabs/excel/0.4.3/*.xlsx
+
+for f in $FILES
+do
+  echo "Processing $f"
+  fttl=${f%.*}.ttl
+
+  /Users/nick/Library/Caches/pypoetry/virtualenvs/vocexcel-XgwzAsEM-py3.11/bin/python /Users/nick/Work/rdflib/VocExcel/vocexcel/__main__.py $f > $fttl
+  echo "Completed $fttl"
+done

scripts/validate_vocabs.py

@@ -2,17 +2,20 @@
 from pyshacl import validate
 import httpx
 
-WARNINGS_INVALID = False # Allows warnings to flag as invalid when true
+WARNINGS_INVALID = False  # Allows warnings to flag as invalid when true
 SHOW_WARNINGS = True
 
+
 def main():
     # get the validator
-    r = httpx.get("https://w3id.org/profile/vocpub/validator/4.6", follow_redirects=True)
+    r = httpx.get("https://w3id.org/profile/vocpub/validator/4.7", follow_redirects=True)
     assert r.status_code == 200
+    with open("vocpub-4.7.ttl", "w") as f:
+        f.write(r.text)
 
     # for all vocabs...
-    warning_vocabs = {} # format {vocab_filename: warning_msg}
-    invalid_vocabs = {} # format {vocab_filename: error_msg}
+    warning_vocabs = {}  # format {vocab_filename: warning_msg}
+    invalid_vocabs = {}  # format {vocab_filename: error_msg}
     vocabs_dir = Path(__file__).parent.parent / "vocabularies"
     for f in vocabs_dir.glob("**/*"):
         # ...validate each file
@@ -45,9 +48,15 @@ def main():
             print(error)
             print("-----")
 
+    ret = True
     if WARNINGS_INVALID:
-        assert len(warning_vocabs.keys()) == 0, "Warning vocabs: {}".format(", ".join([str(x) for x in warning_vocabs.keys()]))
-    assert len(invalid_vocabs.keys()) == 0, "Invalid vocabs: {}".format(", ".join([str(x) for x in invalid_vocabs.keys()]))
+        if len(warning_vocabs.keys()) > 0:
+            print("Warning vocabs: {}".format(", ".join([str(x) for x in warning_vocabs.keys()])))
+            ret = False
+    if len(invalid_vocabs.keys()) > 0:
+        print("Invalid vocabs: {}".format(", ".join([str(x) for x in invalid_vocabs.keys()])))
+        ret = False
+    assert ret
 
 
 if __name__ == "__main__":