Merge pull request #18 from bis-med-it/flatten_hierarchy

Flatten hierarchy

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pysdmx"
-version = "1.0.0-beta-7"
+version = "1.0.0-beta-8"
 description = "Your opinionated Python SDMX library"
 authors = [
     "Xavier Sosnovsky <xavier.sosnovsky@bis.org>",

src/pysdmx/__init__.py

@@ -1,3 +1,3 @@
 """Your opinionated Python SDMX library."""
 
-__version__ = "1.0.0-beta-7"
+__version__ = "1.0.0-beta-8"

src/pysdmx/model/code.py

@@ -257,6 +257,34 @@ def by_id(self, id: str) -> Sequence[HierarchicalCode]:
         """
         return self.__by_id(id, self.codes)
 
+    def __get_codes(
+        self, codes: Sequence[HierarchicalCode]
+    ) -> Sequence[HierarchicalCode]:
+        out = []
+        for code in codes:
+            out.append(code)
+            if code.codes:
+                out.extend(self.__get_codes(code.codes))
+        return out
+
+    def all_codes(self) -> Sequence[HierarchicalCode]:
+        """Get all the codes in the hierarchy as a flat list.
+
+        This is useful for validation purposes. The sequence behaves
+        as a set, i.e. even if a code is attached to multiple nodes,
+        it will be available only once in the returned sequence.
+
+        Returns:
+            A flat list of the codes present in the hierarchy.
+        """
+        out = []
+        # We need to do this below because a hierarchical code is not
+        # (yet?) hashable.
+        for c in self.__get_codes(self.codes):
+            if c not in out:
+                out.append(c)
+        return out
+
 
 class HierarchyAssociation(Struct, frozen=True, omit_defaults=True):
     """Links a hierarchy to a component withing the context of a dataflow."""

src/pysdmx/util/__init__.py

@@ -7,6 +7,8 @@
 
 from pysdmx.errors import NotFound
 
+NF = "Not found"
+
 
 class Reference(Struct, frozen=True):
     """The coordinates of an SDMX maintainable artefact.
@@ -58,9 +60,7 @@ def parse_urn(urn: str) -> Reference:
         )
     else:
         raise NotFound(
-            404,
-            "Not found",
-            f"{urn} does not match {maintainable_urn_pattern}",
+            404, NF, f"{urn} does not match {maintainable_urn_pattern}"
         )
 
 
@@ -76,9 +76,7 @@ def parse_item_urn(urn: str) -> ItemReference:
             item_id=m.group(5),
         )
     else:
-        raise NotFound(
-            404, "Not found", f"{urn} does not match {item_urn_pattern}."
-        )
+        raise NotFound(404, NF, f"{urn} does not match {item_urn_pattern}.")
 
 
 def find_by_urn(artefacts: Sequence[Any], urn: str) -> Any:
@@ -95,7 +93,7 @@ def find_by_urn(artefacts: Sequence[Any], urn: str) -> Any:
         urns = [f"{a.agency}:{a.id}({a.version})" for a in artefacts]
         raise NotFound(
             404,
-            "Not found",
+            NF,
             (
                 f"Could not find an artefact matching the following URN: "
                 f"{urn}. The artefacts received were: {urns}."

tests/model/test_hierarchy.py

@@ -169,3 +169,26 @@ def test_codes_by_id_diff_names(id, name, agency):
     m = list(m)
     assert grandchild1 in m
     assert grandchild3 in m
+
+
+def test_all_codes(id, name, agency):
+    grandchild1 = HierarchicalCode("grandchild1", "grandchild 1")
+    grandchild2 = HierarchicalCode("grandchild2", "grandchild 2")
+    grandchild3 = HierarchicalCode("grandchild3", "grandchild 3")
+    child1 = HierarchicalCode(
+        "child1", "child 1", codes=[grandchild1, grandchild2]
+    )
+    child2 = HierarchicalCode(
+        "child2", "child 2", codes=[grandchild1, grandchild2, grandchild3]
+    )
+    parent1 = HierarchicalCode("parent1", "parent 1")
+    parent2 = HierarchicalCode(
+        "parent2",
+        "parent 2",
+        codes=[child1, child2],
+    )
+
+    h = Hierarchy(id, name, agency, codes=[parent1, parent2])
+
+    m = h.all_codes()
+    assert len(m) == 7