cleanup format_subject

aidencullo · Feb 1, 2024 · c256acb · c256acb
1 parent 7ac48f9
commit c256acb
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 13 deletions.
diff --git a/src/sanitize.py b/src/sanitize.py
@@ -15,23 +15,22 @@ def format_email(raw_email):
 
 def format_date(date_str):
     parser.parse(date_str)
-    # print(f'(\'{date_str}\',\'{date_datetime}\'),')
     return parser.parse(date_str)
 
 
 def format_subject(subject_str):
-    subject = decode_header(subject_str)
-    text_fragments = []
-    for data, _ in subject:
-        try:
-            text_fragments.append(data.decode())
-        except (UnicodeDecodeError, AttributeError):
-            if isinstance(data, bytes):
-                data = data.decode('unicode_escape')
-            text_fragments.append(data)
-    subject = ' '.join(text_fragments)
+    subject = decode_header(subject_str)[0][0]
+    if isinstance(subject, bytes):
+        subject = decode_bytes(subject)
     return subject
 
 
+def decode_bytes(subject):
+    try:
+        return subject.decode()
+    except UnicodeDecodeError:
+        return subject.decode('unicode_escape')
+
+
 def format_uid(uid_str):
     return int(uid_str)
diff --git a/tests/sanitize_test.py b/tests/sanitize_test.py
@@ -43,8 +43,9 @@ class TestSanitize:
         ('Time for your appointment!','Time for your appointment!'),
         ('Added today: new Permanent \'Software Developer\' Jobs | Rochester Jobs & Vacancies...','Added today: new Permanent \'Software Developer\' Jobs | Rochester Jobs & Vacancies...'),
         ('[GitHub] Your Dependabot alerts for the week of Jan 23 - Jan 30','[GitHub] Your Dependabot alerts for the week of Jan 23 - Jan 30'),
-        ('=?utf-8?B?SGVsbG8sIGdvbGRlbiBnb29kbmVzcyDinKg=?= ','Hello, golden goodness ✨  '),
-        ('=?utf-8?B?SW50ZWxsaVNlYXJjaOKEoiBBbGVydCBmb3VuZCAxMzUgbmV3IGpv?==?utf-8?B?YnMsIGJhc2VkIG9uIHlvdXIgcHJvZmlsZSA=?=','IntelliSearch™ Alert found 135 new jobs, based on your profile '),
+        ('=?utf-8?B?SGVsbG8sIGdvbGRlbiBnb29kbmVzcyDinKg=?= ','Hello, golden goodness ✨'),
+        ('=?utf-8?B?SW50ZWxsaVNlYXJjaOKEoiBBbGVydCBmb3VuZCAxMzUgbmV3IGpv?==?utf-8?B?YnMsIGJhc2VkIG9uIHlvdXIgcHJvZmlsZSA=?=','IntelliSearch™ \
+Alert found 135 new jobs, based on your profile '),
         ('Instantly get 3 free months by referring a friend','Instantly get 3 free months by referring a friend'),
         ('=?utf-8?Q?=C3=9ALTIMA=20PREVENTA=20PARA=20CAMDEN?=','ÚLTIMA PREVENTA PARA CAMDEN'),
         ('=?UTF-8?q?4-Unit_Vacation_Rental_in_New_York_=F0=9F=8C=87_26%_Initial_Ren?==?UTF-8?q?tal_Yield_|_Managed_by_Airbnb_Superhost?=','4-Unit Vacation Rental in New York 🌇 26% Initial Rental Yield | Managed by Airbnb Superhost'),