Smarter extraction of filenames from NZB-subject

sabnzbd/nzbstuff.py

@@ -85,11 +85,11 @@ import sabnzbd.cfg as cfg
 import sabnzbd.nzbparser
 from sabnzbd.downloader import Server
 from sabnzbd.database import HistoryDB
-from sabnzbd.deobfuscate_filenames import *
+from sabnzbd.deobfuscate_filenames import is_probably_obfuscated
 
 # Name patterns
-SUBJECT_FN_MATCHER = re.compile(r'"([^"]*)"')
+RE_SUBJECT_FILENAME_QUOTES = re.compile(r'"([^"]*)"')  # In the subject, we expect the filename within double quotes
-RE_NORMAL_NAME = re.compile(r"\.\w{1,5}$")  # Test reasonably sized extension at the end
+RE_SUBJECT_BASIC_FILENAME = re.compile(r"([\w\-+()'\s.,]*\.\w{2,4})")  # Otherwise something that looks like a filename
 RE_RAR = re.compile(r"(\.rar|\.r\d\d|\.s\d\d|\.t\d\d|\.u\d\d|\.v\d\d)$", re.I)
 RE_PROPER = re.compile(r"(^|[\. _-])(PROPER|REAL|REPACK)([\. _-]|$)")
 
@@ -326,7 +326,7 @@ class NzbFile(TryList):
         self.date: datetime.datetime = date
         self.subject: str = subject
         self.type: Optional[str] = None
-        self.filename: str = name_extractor(subject)
+        self.filename: str = sanitize_filename(name_extractor(subject))
         self.filename_checked = False
         self.filepath: Optional[str] = None
 
@@ -2123,10 +2123,20 @@ def scan_password(name: str) -> Tuple[str, Optional[str]]:
 def name_extractor(subject: str) -> str:
     """ Try to extract a file name from a subject line, return `subject` if in doubt """
     result = subject
-    for name in re.findall(SUBJECT_FN_MATCHER, subject):
+    # Filename nicely wrapped in quotes
+    for name in re.findall(RE_SUBJECT_FILENAME_QUOTES, subject):
         name = name.strip(' "')
-        if name and RE_NORMAL_NAME.search(name):
+        if name:
             result = name
+
+    # Found nothing? Try a basic filename-like search
+    if result == subject:
+        for name in re.findall(RE_SUBJECT_BASIC_FILENAME, subject):
+            name = name.strip()
+            if name:
+                result = name
+
+    # Return the subject
     return result
 
 

tests/test_nzbstuff.py

@@ -95,3 +95,41 @@ class TestNZBStuffHelpers:
 
         for file_name, clean_file_name in file_names.items():
             assert nzbstuff.create_work_name(file_name) == clean_file_name
+
+    @pytest.mark.parametrize(
+        "subject, filename",
+        [
+            ('Great stuff (001/143) - "Filename.txt" yEnc (1/1)', "Filename.txt"),
+            (
+                '"910a284f98ebf57f6a531cd96da48838.vol01-03.par2" yEnc (1/3)',
+                "910a284f98ebf57f6a531cd96da48838.vol01-03.par2",
+            ),
+            ('Subject-KrzpfTest [02/30] - ""KrzpfTest.part.nzb"" yEnc', "KrzpfTest.part.nzb"),
+            (
+                '[PRiVATE]-[WtFnZb]-[Supertje-_S03E11-12_-blabla_+_blabla_WEBDL-480p.mkv]-[4/12] - "" yEnc 9786 (1/1366)',
+                "Supertje-_S03E11-12_-blabla_+_blabla_WEBDL-480p.mkv",
+            ),
+            (
+                '[N3wZ] MAlXD245333\\::[PRiVATE]-[WtFnZb]-[Show.S04E04.720p.AMZN.WEBRip.x264-GalaxyTV.mkv]-[1/2] - "" yEnc  293197257 (1/573)',
+                "Show.S04E04.720p.AMZN.WEBRip.x264-GalaxyTV.mkv",
+            ),
+            (
+                'reftestnzb bf1664007a71 [1/6] - "20b9152c-57eb-4d02-9586-66e30b8e3ac2" yEnc (1/22) 15728640',
+                "20b9152c-57eb-4d02-9586-66e30b8e3ac2",
+            ),
+            (
+                "Re: REQ Author Child's The Book-Thanks much - Child, Lee - Author - The Book.epub (1/1)",
+                "REQ Author Child's The Book-Thanks much - Child, Lee - Author - The Book.epub",
+            ),
+            ('63258-0[001/101] - "63258-2.0" yEnc (1/250) (1/250)', "63258-2.0"),
+            (
+                "Singer - A Album (2005) - [04/25] - 02 Sweetest Somebody (I Know).flac",
+                "- 02 Sweetest Somebody (I Know).flac",
+            ),
+            ("<>random!>", "<>random!>"),
+            ("nZb]-[Supertje-_S03E11-12_", "nZb]-[Supertje-_S03E11-12_"),
+            ("Bla [Now it's done.exe]", "Now it's done.exe"),
+        ],
+    )
+    def test_name_extractor(self, subject, filename):
+        assert nzbstuff.name_extractor(subject) == filename