Attach characteristics and ddi reader bug fix

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "ipumspy"
-version = "0.8.0"
+version = "0.8.1"
 description = "A collection of tools for working with IPUMS data"
 authors = ["Kevin H. Wilson <kevin_wilson@brown.edu>",
            "Renae Rodgers <rodge103@umn.edu>"]

src/ipumspy/api/extract.py

@@ -681,7 +681,7 @@ def build(self) -> Dict[str, Any]:
             }
 
         # XXX shoehorn fix until server-side bug is fixed
-        if self.collection == "meps" or self.collection == "dhs":
+        if self.collection in ["meps", "dhs", "mtus", "ahtus"]:
             for variable in built["variables"].keys():
                 built["variables"][variable].pop("attachedCharacteristics")
 

src/ipumspy/ddi.py

@@ -231,19 +231,33 @@ def read(cls, elt: Element, ddi_namespace: str) -> FileDescription:
             rts = [rectype.attrib["rectype"] for rectype in file_rectypes]
         except KeyError:
             rts = []
-        # rectype get rectype id var and rectype key var
+        # (old ddi ordering of record types): rectype get rectype id var and rectype key var
         # these should be the same across record types for all collections
-        # so we should be fine to just grab the first appearance of recidvar and keyvar
+        # so we should be fine to just grab the last appearance of recidvar and keyvar
+        # (the first appearance now is the top level record type, which does not have
+        # the keyvar attribute - so now grabbing from the last record type in the hierarchy)
         try:
             rectype_idvar = elt.find("./ddi:fileStrc/ddi:recGrp", namespaces).attrib[
                 "recidvar"
             ]
             rectype_keyvar = elt.find("./ddi:fileStrc/ddi:recGrp", namespaces).attrib[
                 "keyvar"
             ]
+        # (new ddi ordering of rectypes): if we get a key error, the first appearance now is
+        # the top level record type, which does not have the keyvar attribute -
+        # so now grabbing from the last record type in the hierarchy
+        except KeyError:
+            rectype_idvar = elt.findall("./ddi:fileStrc/ddi:recGrp", namespaces)[
+                -1
+            ].attrib["recidvar"]
+            rectype_keyvar = elt.findall("./ddi:fileStrc/ddi:recGrp", namespaces)[
+                -1
+            ].attrib["keyvar"]
+        # if we get an Attribute Error, this is a rectangular file
         except AttributeError:
             rectype_idvar = ""
             rectype_keyvar = ""
+
         return cls(
             filename=elt.find("./ddi:fileName", namespaces).text,
             description=elt.find("./ddi:fileCont", namespaces).text,