Add header altRecordID support

This (and all the METS elements) could be made a bit more general to
save some code, but leaving that as a future enhancement for now.

metsrw/__init__.py

@@ -6,7 +6,7 @@
 
 from .exceptions import MetsError, ParseError
 from .fsentry import FSEntry
-from .metadata import Agent, AMDSec, SubSection, MDRef, MDWrap
+from .metadata import Agent, AltRecordID, AMDSec, SubSection, MDRef, MDWrap
 from .mets import METSDocument
 from .utils import (
     NAMESPACES,
@@ -47,6 +47,7 @@
 
 __all__ = [
     "Agent",
+    "AltRecordID",
     "AMDSec",
     "AM_PNTR_SCT_PATH",
     "AM_SCT_PATH",

metsrw/metadata.py

@@ -98,6 +98,58 @@ def serialize(self, now=None):
         return el
 
 
+class AltRecordID(object):
+    """
+    An object representing an alternative record identifier in the METS document
+    (alternatives to the OBJID).
+
+    This is ordinarily created by :class:`metsrw.mets.METSDocument` instances and
+    does not have to be instantiated directly.
+
+    :param str id: Optional unique identifer for the identifier.
+    :param str type: Optional identifer type, e.g. 'Accession number'.
+    """
+
+    ALT_RECORD_ID_TAG = etree.QName(utils.NAMESPACES[u"mets"], u"altRecordID")
+
+    def __init__(self, alt_record_id, **kwargs):
+        self.text = alt_record_id
+        # We use kwargs here to avoid shadowing builtins (id and type).
+        self.id = kwargs.get("id", None)
+        self.type = kwargs.get("type", None)
+
+    @classmethod
+    def parse(cls, element):
+        """
+        Create a new AltRecordID by parsing root.
+
+        :param element: Element to be parsed into an AltRecordID.
+        :raises exceptions.ParseError: If element is not a valid altRecordID.
+        """
+        if element.tag != cls.ALT_RECORD_ID_TAG:
+            raise exceptions.ParseError(
+                u"AltRecordID got unexpected tag {}; expected {}".format(
+                    element.tag, cls.ALT_RECORD_ID_TAG
+                )
+            )
+
+        return cls(element.text, id=element.get(u"ID"), type=element.get(u"TYPE"))
+
+    def serialize(self):
+        attrs = {}
+
+        if self.id:
+            attrs[u"ID"] = self.id
+
+        if self.type:
+            attrs[u"TYPE"] = self.type
+
+        element = etree.Element(self.ALT_RECORD_ID_TAG, **attrs)
+        element.text = self.text
+
+        return element
+
+
 class Agent(object):
     """
     An object representing an agent with a relationship to the METS record.

metsrw/mets.py

@@ -34,6 +34,7 @@ def __init__(self):
         # can be inferred via their #children attribute
         self.createdate = None
         self.objid = None
+        self.alternate_ids = []
         self._root_elements = []
         self._all_files = None
         self._iter = None
@@ -192,10 +193,10 @@ def _mets_header(self, now):
             header_attrs[u"LASTMODDATE"] = now
 
         header_element = etree.Element(header_tag, **header_attrs)
-
         for agent in self.agents:
-            agent_element = agent.serialize()
-            header_element.append(agent_element)
+            header_element.append(agent.serialize())
+        for alternate_id in self.alternate_ids:
+            header_element.append(alternate_id.serialize())
 
         return header_element
 
@@ -519,11 +520,20 @@ def _parse_header(self, tree):
         self.createdate = createdate
 
         if header is not None:
-            agent_elements = header.findall(u"mets:agent", namespaces=utils.NAMESPACES)
+            agent_elements = header.findall(
+                metadata.Agent.AGENT_TAG, namespaces=utils.NAMESPACES
+            )
             for agent_element in agent_elements:
                 agent = metadata.Agent.parse(agent_element)
                 self.agents.append(agent)
 
+            alternate_ids = header.findall(
+                metadata.AltRecordID.ALT_RECORD_ID_TAG, namespaces=utils.NAMESPACES
+            )
+            for alternate_id_element in alternate_ids:
+                alternate_id = metadata.AltRecordID.parse(alternate_id_element)
+                self.alternate_ids.append(alternate_id)
+
     def _validate(self):
         raise NotImplementedError()
 

tests/test_metadata.py

@@ -72,6 +72,33 @@ def test_serialize_with_other_role(self):
         assert element.get("OTHERROLE") == agent.role
 
 
+class TestAltRecordId(TestCase):
+    def test_parse_exception_on_wrong_tag(self):
+        element = etree.Element("test")
+        with pytest.raises(metsrw.ParseError):
+            metsrw.AltRecordID.parse(element)
+
+    def test_parse(self):
+        element = etree.Element(
+            metsrw.AltRecordID.ALT_RECORD_ID_TAG, ID="543", TYPE="Test"
+        )
+        element.text = "a-unique-id"
+        alt_record_id = metsrw.AltRecordID.parse(element)
+
+        assert alt_record_id.text == element.text
+        assert alt_record_id.id == element.get("ID")
+        assert alt_record_id.type == element.get("TYPE")
+
+    def test_serialize(self):
+        alt_record_id = metsrw.AltRecordID("12345", id="1", type="Accession Id")
+        element = alt_record_id.serialize()
+
+        assert element.get("ID") == alt_record_id.id
+        assert element.get("TYPE") == alt_record_id.type
+
+        assert element.text == alt_record_id.text
+
+
 class TestAMDSec(TestCase):
     """ Test AMDSec class. """
 

tests/test_mets.py

@@ -223,6 +223,36 @@ def test_parse_header_with_agent(self):
         assert mets.agents[0].name == u"39461beb-22eb-4942-88af-848cfc3462b2"
         assert mets.agents[0].notes[0] == u"Archivematica dashboard UUID"
 
+    def test_mets_header_with_alt_record_id(self):
+        mets = metsrw.METSDocument()
+        alt_record_id = metsrw.AltRecordID(
+            "39461beb-22eb-4942-88af-848cfc3462b2", type="Accession ID"
+        )
+        mets.alternate_ids.append(alt_record_id)
+
+        header_element = mets._mets_header("2014-07-16T22:52:02.480108")
+        alt_record_id_element = header_element.find(
+            "{http://www.loc.gov/METS/}altRecordID"
+        )
+
+        assert alt_record_id_element.get("TYPE") == alt_record_id.type
+        assert alt_record_id_element.text == alt_record_id.text
+
+    def test_parse_header_with_alt_record_id(self):
+        mets = metsrw.METSDocument.fromstring(
+            b"""<?xml version='1.0' encoding='ASCII'?>
+<mets xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.loc.gov/METS/" xsi:schemaLocation="http://www.loc.gov/METS/ http://www.loc.gov/standards/mets/version18/mets.xsd">
+    <metsHdr CREATEDATE="2015-12-16T22:38:48">
+        <altRecordID TYPE="Accession Id">39461beb-22eb-4942-88af-848cfc3462b2</altRecordID>
+    </metsHdr>
+    <structMap ID="structMap_1" LABEL="Archivematica default" TYPE="physical"/>
+</mets>"""
+        )
+
+        assert len(mets.alternate_ids) == 1
+        assert mets.alternate_ids[0].type == u"Accession Id"
+        assert mets.alternate_ids[0].text == u"39461beb-22eb-4942-88af-848cfc3462b2"
+
     def test_fromfile_invalid_xlink_href(self):
         """Test that ``fromfile`` raises ``ParseError`` if an xlink:href value
         in the source METS contains an unparseable URL.