Fix issue with namespace rebinding

[aucampia]

Also add additional tests for namespace rebinding.

Note

Commits with @gjhiggins as author are taken from:

• Bind prefixes choices #1686 (comment)
• Bind prefixes choices #1686 (comment)
• https://gist.github.com/gjhiggins/b4b59bffc18aa080ce466b01eb0e5d70
• Fix issue with namespace rebinding #1800 (comment)

[aucampia]

pre-commit.ci autofix

[aucampia]

@gjhiggins Good spotting on this - I think we need to fix it for other stores also though, I will add an additional test to check this against other stores and then add fixes for that.

[ghost]

looks good, thank you

[ghost]

@gjhiggins Good spotting on this - I think we need to fix it for other stores also though, I will add an additional test to check this against other stores and then add fixes for that.

Maybe save you some time, here's a patch for berkeleydb.py

diff --git a/rdflib/plugins/stores/berkeleydb.py b/rdflib/plugins/stores/berkeleydb.py
index 9ca4380b..857b950d 100644
--- a/rdflib/plugins/stores/berkeleydb.py
+++ b/rdflib/plugins/stores/berkeleydb.py
@@ -470,10 +470,17 @@ class BerkeleyDB(Store):
         prefix = prefix.encode("utf-8")
         namespace = namespace.encode("utf-8")
         bound_prefix = self.__prefix.get(namespace)
-        if override and bound_prefix:
-            self.__namespace.delete(bound_prefix)
-        self.__prefix[namespace] = prefix
-        self.__namespace[prefix] = namespace
+        bound_namespace = self.__namespace.get(prefix)
+        if override:
+            if bound_prefix:
+                self.__namespace.delete(bound_prefix)
+            if bound_namespace:
+                self.__prefix.delete(bound_namespace)
+            self.__prefix[namespace] = prefix
+            self.__namespace[prefix] = namespace
+        else:
+            self.__prefix[bound_namespace or namespace] = bound_prefix or prefix
+            self.__namespace[bound_prefix or prefix] = bound_namespace or namespace
 
     def namespace(self, prefix):
         prefix = prefix.encode("utf-8")

[aucampia]

@nicholascar @gjhiggins I'm going to throw caution to the wind with this and fix some things which I would have liked to keep to separate PRs in this PR, but we really should try to keep these things less entangled in future so we can get changes in quicker.

[aucampia]

There are 4 stores doing about the same thing (rdflib.plugins.store.{memory.SimpleMemory,memory.Memory,memory.BerkeleyDB,}), I think the best option is to move the code for Store.{bind,prefix,namespace,namespaces} to a separate class.

[aucampia]

Will look at this further tomorrow, won't work that well with memory store, but there is also many cases where strings are checked for turthyness instead of is not None, will try fix them also.

[aucampia]

Actually, scratch the last comments, I will fix further issues and expand tests once this is in master. But beware, there are still at least one store with this issue and then this should be checking is None instead of turthyness, and it should also use a proper null coalescing method, and I'm not entirely sure of all the consequences, but I will figure it out once this and #1686 is merged.

[aucampia]

Suggested change

	if bound_prefix:
	if bound_prefix is not None:

[aucampia]

Suggested change

	if bound_namespace:
	if bound_namespace is not None:

[aucampia]

This will not do the right thing if bound_namespace is ""

[aucampia]

This will not do the right thing if bound_prefix is ""

[aucampia]

Suggested change

	if bound_prefix:
	if bound_prefix is not None:

[aucampia]

Suggested change

	if bound_namespace:
	if bound_namespace is not None:

[aucampia]

This will not do the right thing if bound_namespace is ""

[aucampia]

This will not do the right thing if bound_prefix is ""

[aucampia]

Should use null coalesing, something like this:

_AnyT = TypeVar("_AnyT")


def _coalesce(*args: Optional[_AnyT]) -> Optional[_AnyT]:
    for arg in args:
        if arg is not None:
            return arg
    return None

[aucampia]

@gjhiggins I will redo this on master.