Expand wildcard attributes in IfcCsv export (#3424)

Exporting with a "*" attribute (or "Pset_Name.*") emitted a literal "*"
column filled with the null placeholder instead of expanding: the
get_wildcard_attributes helper had no callers after a refactor dropped the
expansion step.

Wire the expansion back in: before the per-element loop, export() now calls
expand_wildcard_attributes, which passes through non-wildcard queries
unchanged, expands a bare "*" to every direct attribute of the exported
elements (new get_element_attributes helper), and expands "Pset_Name.*" /
"Qto_Name.*" via the existing get_wildcard_attributes helper. Expanded
columns that duplicate an existing one (e.g. GlobalId or an explicit Name)
are skipped.

Verified headless: exporting an IfcWall with ["Name","*"] now yields real
attribute columns with values (W1, MyWall, T1, ...) instead of a literal
"*", "Pset_WallCommon.*" expands to that pset's properties, and an explicit
named-attribute export is unchanged.

Generated with the assistance of an AI coding tool.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/ifccsv/ifccsv.py

@@ -105,6 +105,9 @@ class IfcCsv:
             attributes.insert(0, "GlobalId")
             headers.insert(0, "GlobalId")
 
+        elements = list(elements)
+        attributes, headers = self.expand_wildcard_attributes(attributes, headers, elements)
+
         for element in elements:
             result = []
 
@@ -384,6 +387,44 @@ class IfcCsv:
         self.dataframe = pd.DataFrame(self.results, columns=self.headers)
         return self.dataframe
 
+    def expand_wildcard_attributes(self, attributes, headers, elements):
+        """Expand any attribute containing ``*`` into concrete attribute names.
+
+        A bare ``*`` expands to every direct attribute of the exported elements.
+        A ``Pset_Name.*`` (or ``Qto_Name.*``) query expands to every property of
+        that named property set / quantity set, using ``get_wildcard_attributes``.
+        Attributes without a ``*`` are passed through unchanged so explicit
+        named queries keep their position and any custom header.
+        """
+        expanded_attributes = []
+        expanded_headers = []
+        for attribute, header in zip(attributes, headers):
+            if "*" not in attribute:
+                expanded_attributes.append(attribute)
+                expanded_headers.append(header)
+                continue
+            if attribute == "*":
+                expansions = self.get_element_attributes(elements)
+            else:
+                expansions = self.get_wildcard_attributes(attribute)
+            for expanded in expansions:
+                if expanded in expanded_attributes:
+                    continue  # Avoid duplicate columns (e.g. GlobalId).
+                expanded_attributes.append(expanded)
+                # Expanded columns fall back to the attribute name as header.
+                expanded_headers.append(None)
+        return expanded_attributes, expanded_headers
+
+    def get_element_attributes(self, elements):
+        results = []
+        for element in elements:
+            for name in element.get_info(recursive=False):
+                if name in ("id", "type"):
+                    continue
+                if name not in results:
+                    results.append(name)
+        return results
+
     def get_wildcard_attributes(self, attribute):
         results = set()
         pset_qto_name = attribute.split(".", 1)[0]