fix bmo parser for new format / lints

actual_imap_poll/parsers.py

@@ -58,6 +58,18 @@ class TransactionParser(ABC):
             raise TransactionParsingFailed("No content of message found")
         return content
 
+    @staticmethod
+    def strip_html(msg: EmailMessage) -> str:
+        body = msg.get_body(preferencelist=("html", "plain"))
+        if body is None:
+            raise TransactionParsingFailed("No HTML body of message found")
+        content = body.get_content()
+        if content is None:
+            raise TransactionParsingFailed("No content of message found")
+
+        soup = BeautifulSoup(content, "html.parser")
+        return soup.get_text()
+
 
 class TransactionParsingFailed(Exception):
     pass
@@ -136,8 +148,8 @@ class MBNAParser(TransactionParser):
 
 class BMOParser(TransactionParser):
     EXTRACT_RE = re.compile(
-        r"We want to let you know that a (withdrawal|deposit) of\s+\$([0-9,]+\.\d{2})\s+has been made (?:to|from) your account ending\s+in\s+(\d{3})",  # noqa: E501
-        flags=re.MULTILINE,
+        r"There was a (withdrawal|deposit).*Amount:\s*\$([0-9,]+\.\d{2}).*Account:\s*Ending in ([0-9]+)",  # noqa: E501
+        flags=re.DOTALL,
     )
 
     def __init__(self, account_map: dict[int, UUID]):
@@ -147,9 +159,8 @@ class BMOParser(TransactionParser):
         return msg["From"] == "bmoalerts@bmo.com"
 
     def extract(self, msg: EmailMessage) -> Optional[Transaction]:
-        content = self.get_content(msg)
-        soup = BeautifulSoup(content, "html.parser")
-        matches = self.EXTRACT_RE.search(soup.get_text())
+        content = self.strip_html(msg)
+        matches = self.EXTRACT_RE.search(content)
         if matches is None:
             raise TransactionParsingFailed("No matches for extraction RE")