Pick up migrations of some objects to case_utils.inherent_uuid

No effects were observed on Make-managed files.

Signed-off-by: Alex Nelson <[email protected]>

src/case_utils_extras.py

@@ -22,58 +22,18 @@
 date 2004-03-04) and did not have a linked superseding document.
 """
 
-import binascii
-import re
-import uuid
-from typing import Dict, Optional, Tuple
-
-from case_utils.namespace import NS_UCO_VOCABULARY, NS_XSD
-from rdflib import Literal, Namespace, URIRef
-
-
-L_MD5 = Literal("MD5", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SHA1 = Literal("SHA1", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SHA256 = Literal("SHA256", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SHA3_256 = Literal("SHA3-256", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SHA384 = Literal("SHA384", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SHA512 = Literal("SHA512", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-L_SSDEEP = Literal("SSDEEP", datatype=NS_UCO_VOCABULARY.HashNameVocab)
-
-# Key: hashMethod literal.
-# Value: Tuple.
-#   * Lowercase spelling
-# TODO - SHA3-256 is used by Digital Corpora, impacting CASE-Corpora, but the vocabulary string is not yet in UCO.
-# https://github.com/ucoProject/UCO/issues/526
-HASH_METHOD_CASTINGS: Dict[Literal, Tuple[str, Optional[int]]] = {
-    L_MD5: ("md5", 32),
-    L_SHA1: ("sha1", 40),
-    L_SHA256: ("sha256", 64),
-    L_SHA3_256: ("sha3-256", 64),
-    L_SHA384: ("sha384", 96),
-    L_SHA512: ("sha512", 128),
-    L_SSDEEP: ("ssdeep", None),
-}
-
-RX_UUID = re.compile(
-    "[0-9a-f]{8}-[0-9a-f]{4}-[0-5][0-9a-f]{3}-[0-9a-f]{4}-[0-9a-f]{12}$", re.IGNORECASE
+from case_utils.inherent_uuid import (
+    RX_UUID,
+    hash_method_value_uuid,
 )
+from rdflib import Literal, Namespace, URIRef
 
 
 def method_value_to_node(
     l_hash_method: Literal, l_hash_value: Literal, ns_kb: Namespace
 ) -> URIRef:
-    if l_hash_value.datatype != NS_XSD.hexBinary:
-        raise ValueError("Expected hexBinary datatype for l_hash_value.")
-    hash_value_str: str = binascii.hexlify(l_hash_value.toPython()).decode().lower()
-
-    hash_method_str = HASH_METHOD_CASTINGS[l_hash_method][0]
-
-    urn_template = "urn:hash::%s:%s"
-    urn_populated = urn_template % (hash_method_str, hash_value_str)
-
-    hash_uuid = str(uuid.uuid5(uuid.NAMESPACE_URL, urn_populated))
+    hash_uuid = str(hash_method_value_uuid(l_hash_method, l_hash_value))
     n_hash = ns_kb["Hash-" + hash_uuid]
-
     return n_hash
 
 

src/digital_corpora_supplement_ttl.py

@@ -29,11 +29,11 @@
     NS_UCO_VOCABULARY,
     NS_XSD,
 )
-from case_utils_extras import (
+from case_utils.inherent_uuid import (
     L_SHA256,
     L_SHA3_256,
-    method_value_to_node,
 )
+from case_utils_extras import method_value_to_node
 from rdflib import Graph, Literal, Namespace, URIRef
 
 NS_CASE_CORPORA = Namespace("http://example.org/ontology/case-corpora/")