Coerce malformed <email> in aggregate report metadata to None (#740)

xmltodict turns stray angle brackets in <email> (e.g.
"<bad-xml@bad-xml.net>") into a nested dict, which then flows through
parse_aggregate_report_xml as the org_email value. Parsing succeeds, but
Elasticsearch / OpenSearch reject the document at index time because the
org_email mapping is text — observed as document_parsing_exception /
mapper_parsing_exception with a "{#text=..., bad-xml=null}" preview.

When report_metadata["email"] comes back as a dict, log it at debug and
discard. The rest of the report still ingests with org_email=None
instead of failing the whole document downstream.

Co-authored-by: Sean Whalen <seanthegeek@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

parsedmarc/__init__.py

@@ -729,6 +729,12 @@ def parse_aggregate_report_xml(
 
         report = xmltodict.parse(xml)["feedback"]
         report_metadata = report["report_metadata"]
+        if isinstance(report_metadata.get("email"), dict):
+            logger.debug(
+                "Discarding malformed <email> in report_metadata: %r",
+                report_metadata["email"],
+            )
+            report_metadata["email"] = None
         schema = "draft"
         if "version" in report:
             schema = report["version"]