Merge pull request #621 from FAIRmat-NFDI/sibling-inheritance

Author: lukaspie

.github/workflows/plugin_test.yaml

@@ -36,10 +36,10 @@ jobs:
             branch: main
             tests_to_run: tests/.
           - plugin: pynxtools-raman
-            branch: main
+            branch: sibling-inheritance
             tests_to_run: tests/.
           - plugin: pynxtools-spm
-            branch: main
+            branch: field-inheritance
             tests_to_run: tests/.
           - plugin: pynxtools-xps
             branch: main

src/pynxtools/data/NXtest.nxdl.xml

@@ -123,6 +123,7 @@
         </group>
         <group name="identified_calibration" type="NXcalibration" optional="true">
             <field name="identifier_1"/>
+            <field name="identifier_2" optional="True"/>
         </group>
         <group name="named_collection" type="NXcollection" optional="true"/>
     </group>

src/pynxtools/dataconverter/helpers.py

@@ -24,7 +24,7 @@
 from datetime import datetime, timezone
 from enum import Enum
 from functools import lru_cache
-from typing import Any, Callable, List, Optional, Tuple, Union, Sequence
+from typing import Any, Callable, List, Optional, Tuple, Union, Sequence, cast
 
 import h5py
 import lxml.etree as ET
@@ -67,6 +67,7 @@ class ValidationProblem(Enum):
     NXdataMissingAxisData = 19
     NXdataAxisMismatch = 20
     KeyToBeRemoved = 21
+    InvalidConceptForNonVariadic = 22
 
 
 class Collector:
@@ -150,6 +151,12 @@ def _log(self, path: str, log_type: ValidationProblem, value: Optional[Any], *ar
             )
         elif log_type == ValidationProblem.KeyToBeRemoved:
             logger.warning(f"The attribute {path} will not be written.")
+        elif log_type == ValidationProblem.InvalidConceptForNonVariadic:
+            value = cast(Any, value)
+            log_text = f"Given {value.type} name '{path}' conflicts with the non-variadic name '{value}'"
+            if value.type == "group":
+                log_text += f", which should be of type {value.nx_class}."
+            logger.warning(log_text)
 
     def collect_and_log(
         self,

src/pynxtools/dataconverter/nexus_tree.py

@@ -41,6 +41,7 @@
     is_variadic,
     is_appdef,
     remove_namespace_from_tag,
+    NEXUS_TO_PYTHON_DATA_TYPES,
 )
 from pynxtools.definitions.dev_tools.utils.nxdl_utils import (
     get_nx_namefit,
@@ -214,19 +215,6 @@ def __init__(
         self.is_a = []
         self.parent_of = []
 
-    def _construct_inheritance_chain_from_parent(self):
-        """
-        Builds the inheritance chain of the current node based on the parent node.
-        """
-        if self.parent is None:
-            return
-        for xml_elem in self.parent.inheritance:
-            elem = xml_elem.find(
-                f"nx:{self.type}/[@name='{self.name}']", namespaces=namespaces
-            )
-            if elem is not None:
-                self.inheritance.append(elem)
-
     def get_path(self) -> str:
         """
         Gets the path of the current node based on the node name.
@@ -356,6 +344,7 @@ def get_all_direct_children_names(
         Returns:
             Set[str]: A set of children names.
         """
+
         if depth is not None and (not isinstance(depth, int) or depth < 0):
             raise ValueError("Depth must be a positive integer or None")
 
@@ -601,6 +590,7 @@ def add_node_from(self, xml_elem: ET._Element) -> Optional["NexusNode"]:
                 type=tag,
                 optionality=default_optionality,
                 nxdl_base=xml_elem.base,
+                inheritance=[xml_elem],
             )
         elif tag == "group":
             name = xml_elem.attrib.get("name")
@@ -684,6 +674,19 @@ def __init__(self, **data) -> None:
         self._construct_inheritance_chain_from_parent()
         self._set_optionality()
 
+    def _construct_inheritance_chain_from_parent(self):
+        """
+        Builds the inheritance chain of the current node based on the parent node.
+        """
+        if self.parent is None:
+            return
+        for xml_elem in self.parent.inheritance:
+            elem = xml_elem.find(
+                f"nx:{self.type}/[@name='{self.name}']", namespaces=namespaces
+            )
+            if elem is not None:
+                self.inheritance.append(elem)
+
 
 class NexusGroup(NexusNode):
     """
@@ -864,6 +867,145 @@ class NexusEntity(NexusNode):
     open_enum: bool = False
     shape: Optional[Tuple[Optional[int], ...]] = None
 
+    def _check_compatibility_with(self, xml_elem: ET._Element) -> bool:
+        """Check compatibility of this node with an XML element from the (possible) inheritance"""
+
+        def _check_name_fit(xml_elem: ET._Element) -> bool:
+            elem_name = xml_elem.attrib.get("name")
+            name_any = is_name_type(xml_elem, "any")
+            name_partial = is_name_type(xml_elem, "partial")
+
+            if get_nx_namefit(self.name, elem_name, name_any, name_partial) < 0:
+                return False
+            return True
+
+        def _check_type_fit(xml_elem: ET._Element) -> bool:
+            elem_type = xml_elem.attrib.get("type")
+            if elem_type:
+                if not set(NEXUS_TO_PYTHON_DATA_TYPES[self.dtype]).issubset(
+                    NEXUS_TO_PYTHON_DATA_TYPES[elem_type]
+                ):
+                    return False
+            return True
+
+        def _check_units_fit(xml_elem: ET._Element) -> bool:
+            elem_units = xml_elem.attrib.get("units")
+            if elem_units and elem_units != "NX_ANY":
+                if elem_units != self.unit:
+                    if not elem_units == "NX_TRANSFORMATION" and self.unit in [
+                        "NX_LENGTH",
+                        "NX_ANGLE",
+                        "NX_UNITLESS",
+                    ]:
+                        return False
+            return True
+
+        def _check_enum_fit(xml_elem: ET._Element) -> bool:
+            elem_enum = xml_elem.find(f"nx:enumeration", namespaces=namespaces)
+            if elem_enum is not None:
+                if self.items is None:
+                    # Case where inherited entity is enumerated, but current node isn't
+                    return False
+                elem_enum_open = elem_enum.attrib.get("open", "false")
+
+                if elem_enum_open == "true":
+                    return True
+
+                elem_enum_items = []
+                for items in elem_enum.findall(f"nx:item", namespaces=namespaces):
+                    value = items.attrib["value"]
+                    if value[0] == "[" and value[-1] == "]":
+                        import ast
+
+                        try:
+                            elem_enum_items.append(ast.literal_eval(value))
+                        except (ValueError, SyntaxError):
+                            raise Exception(
+                                f"Error parsing enumeration item in the provided NXDL: {value}"
+                            )
+                    else:
+                        elem_enum_items.append(value)
+
+                def convert_to_hashable(item):
+                    """Convert lists to tuples for hashable types, leave non-list items as they are."""
+                    if isinstance(item, list):
+                        return tuple(item)  # Convert sublists to tuples
+                    return item  # Non-list items remain as they are
+
+                set_items = {convert_to_hashable(sublist) for sublist in self.items}
+                set_elem_enum_items = {
+                    convert_to_hashable(sublist) for sublist in elem_enum_items
+                }
+
+                if not set(set_items).issubset(set_elem_enum_items):
+                    if self.name == "definition":
+                        pass
+                    else:
+                        # TODO: should we be this strict here? Or can appdefs define additional terms?
+                        pass
+            return True
+
+        def _check_dimensions_fit(xml_elem: ET._Element) -> bool:
+            if not self.shape:
+                return True
+            elem_dimensions = xml_elem.find(f"nx:dimensions", namespaces=namespaces)
+            if elem_dimensions is not None:
+                rank = elem_dimensions.attrib.get("rank")
+                if rank is not None and not isinstance(rank, int):
+                    try:
+                        int(rank)
+                    except ValueError:
+                        # TODO: Handling of symbols
+                        return True
+                elem_dim = elem_dimensions.findall("nx:dim", namespaces=namespaces)
+                elem_dimension_rank = rank if rank is not None else len(rank)
+                dims: List[Optional[int]] = [None] * int(rank)
+
+                for dim in elem_dim:
+                    idx = int(dim.attrib["index"])
+                    if value := dim.attrib.get("value", None):
+                        # If not, this is probably an old dim element with ref.
+                        try:
+                            value = int(value)
+                            dims[idx] = value
+                        except ValueError:
+                            # TODO: Handling of symbols
+                            pass
+                elem_shape = tuple(dims)
+
+                if elem_shape:
+                    if elem_shape != self.shape:
+                        return False
+
+            return True
+
+        check_functions = [
+            _check_name_fit,
+            _check_type_fit,
+            _check_units_fit,
+            _check_enum_fit,
+            # TODO: check if any inheritance is wrongfully assigned without dim checks
+            # _check_dimensions_fit,
+        ]
+
+        for func in check_functions:
+            if not func(xml_elem):
+                return False
+        return True
+
+    def _construct_inheritance_chain_from_parent(self):
+        """
+        Builds the inheritance chain of the current node based on the parent node.
+        """
+        if self.parent is None:
+            return
+        for xml_elem in self.parent.inheritance:
+            subelems = xml_elem.findall(f"nx:{self.type}", namespaces=namespaces)
+            if subelems is not None:
+                for elem in subelems:
+                    if self._check_compatibility_with(elem):
+                        self.inheritance.append(elem)
+
     def _set_type(self):
         """
         Sets the dtype of the current entity based on the values in the inheritance chain.
@@ -950,7 +1092,13 @@ def _set_shape(self):
 
     def __init__(self, **data) -> None:
         super().__init__(**data)
+        self._set_unit()
+        self._set_type()
+        self._set_items_and_enum_type()
+        self._set_optionality()
+        self._set_shape()
         self._construct_inheritance_chain_from_parent()
+        # Set all parameters again based on the acquired inheritance
         self._set_unit()
         self._set_type()
         self._set_items_and_enum_type()

src/pynxtools/dataconverter/validation.py

@@ -155,6 +155,31 @@ def best_namefit_of(name: str, nodes: Iterable[NexusNode]) -> Optional[NexusNode
     for node in nodes:
         if not node.variadic:
             if instance_name == node.name:
+                if concept_name and concept_name != node.name:
+                    inherited_names = [
+                        name
+                        if (name := elem.attrib.get("name")) is not None
+                        else type_attr[2:].upper()
+                        for elem in node.inheritance
+                        if (name := elem.attrib.get("name")) is not None
+                        or (type_attr := elem.attrib.get("type"))
+                        and len(type_attr) > 2
+                    ]
+                    if concept_name not in inherited_names:
+                        if node.type == "group":
+                            if concept_name != node.nx_class[2:].upper():
+                                collector.collect_and_log(
+                                    concept_name,
+                                    ValidationProblem.InvalidConceptForNonVariadic,
+                                    node,
+                                )
+                        else:
+                            collector.collect_and_log(
+                                concept_name,
+                                ValidationProblem.InvalidConceptForNonVariadic,
+                                node,
+                            )
+                        return None
                 return node
         else:
             if concept_name and concept_name == node.name:
@@ -194,16 +219,32 @@ def validate_dict_against(
     """
 
     def get_variations_of(node: NexusNode, keys: Mapping[str, Any]) -> List[str]:
+        variations = []
+
+        prefix = f"{'@' if node.type == 'attribute' else ''}"
         if not node.variadic:
-            if f"{'@' if node.type == 'attribute' else ''}{node.name}" in keys:
-                return [node.name]
+            if f"{prefix}{node.name}" in keys:
+                variations += [node.name]
             elif (
                 hasattr(node, "nx_class")
                 and f"{convert_nexus_to_caps(node.nx_class)}[{node.name}]" in keys
             ):
-                return [f"{convert_nexus_to_caps(node.nx_class)}[{node.name}]"]
-
-        variations = []
+                variations += [f"{convert_nexus_to_caps(node.nx_class)}[{node.name}]"]
+
+            # Also add all variations like CONCEPT[node.name] for inherited concepts
+            inherited_names = []
+            for elem in node.inheritance:
+                inherited_name = elem.attrib.get("name")
+                if not inherited_name:
+                    inherited_name = elem.attrib.get("type")[2:].upper()
+                if inherited_name.startswith("NX"):
+                    inherited_name = inherited_name[2:].upper()
+                inherited_names += [inherited_name]
+            for name in set(inherited_names):
+                if f"{prefix}{name}[{prefix}{node.name}]" in keys:
+                    variations += [f"{prefix}{name}[{prefix}{node.name}]"]
+
+            return variations
 
         for key in keys:
             concept_name, instance_name = split_class_and_name_of(key)