More robust url parsing

datagraph/parse.py

@@ -1,17 +1,23 @@
 import json
-import urllib
+import urllib.parse
 
 import schemeld
 
 
-def determine_concepts_internal(json, context, outputs):
+def parse_url(url: str | urllib.parse.ParseResult) -> urllib.parse.ParseResult:
+    if isinstance(url, urllib.parse.ParseResult):
+        return url
+    return urllib.parse.urlparse(url)
+
+
+def determine_concepts_internal(json: dict | list, context, outputs: list[schemeld.Concept]) -> None:
     if isinstance(json, list):
         for m in json:
             determine_concepts_internal(m, context, outputs)
         return
 
     assert isinstance(json, dict), type(json)
-    context = urllib.parse.urlparse(json.get('@context', context))
+    context = parse_url(json.get('@context', context))
     assert context.netloc == 'schema.org'
 
     if '@graph' in json:
@@ -20,13 +26,13 @@ def determine_concepts_internal(json, context, outputs):
         outputs.append(schemeld.Concept(context, json))
 
 
-def determine_concepts(json):
+def determine_concepts(json: dict | list) -> list[schemeld.Concept]:
     concepts = []
     determine_concepts_internal(json, '', concepts)
     return concepts
 
 
-def determine_concepts_in_soup(soup):
+def determine_concepts_in_soup(soup) -> list[schemeld.Concept]:
     # TODO: Check type
     ld_json_elements = soup.find_all('script', type='application/ld+json')
     concepts = []

datagraph/schemeld.py

@@ -5,7 +5,7 @@ from typing import Any
 STRICT_VALIDATION = True
 
 Key = int | str | urllib.parse.ParseResult
-Context = str  # TODO
+Context = urllib.parse.ParseResult  # TODO
 
 
 def canonical_keys(base_key: Key, context: Context | None) -> list[Any]: