Try to extract subject hash from URL if missing

test_twtxtparser.py

@@ -157,6 +157,14 @@ class ParseSubjectTest(unittest.TestCase):
         self.assertEqual(SubjectHash(start_pos=0, end_pos=31, hash=None, url="https://example.com/"),
                          _parse_subject("(  #<  https://example.com/ > ) more text", 0))
 
+    def test_only_url_hash_extracted_from_url(self):
+        self.assertEqual(SubjectHash(start_pos=0, end_pos=34, hash="hfp2hca", url="https://twtxt.net/twt/hfp2hca"),
+                         _parse_subject("(#<https://twtxt.net/twt/hfp2hca>)", 0))
+        self.assertEqual(SubjectHash(start_pos=0, end_pos=35, hash="hfp2hca", url="https://twtxt.net/conv/hfp2hca"),
+                         _parse_subject("(#<https://twtxt.net/conv/hfp2hca>)", 0))
+        self.assertEqual(SubjectHash(start_pos=0, end_pos=43, hash="5jqioeq", url="https://txt.sour.is/search?tag=5jqioeq"),
+                         _parse_subject("(#<https://txt.sour.is/search?tag=5jqioeq>)", 0))
+
 
 class ParseMarkdownLink(unittest.TestCase):
 

twtxtparser.py

@@ -23,6 +23,7 @@ Known problems and limitations:
 """
 
 import collections
+import re
 
 
 def _token_class(name, *fields):
@@ -106,6 +107,8 @@ def _parse_mention(text, start_pos):
     return __parse_text_and_url_in_brackets(text, start_pos, None, "@<", ">", Mention)
 
 
+_TWT_HASH_RE = re.compile(r"^[a-z0-9]{7}$")
+
 def _parse_subject(text, start_pos):
     data, closing_pos = _extract_between(text, start_pos, None, "(", ")")
     if data is None:
@@ -122,7 +125,14 @@ def _parse_subject(text, start_pos):
         # treat this as a subject to be safe…'
         return None
 
-    return SubjectHash(start_pos, closing_pos, subject_hash.hash, subject_hash.url)
+    hash = subject_hash.hash
+    if not hash:
+        # try to extract the hash from the last part of the URL
+        hash = subject_hash.url[-7:]
+        match = _TWT_HASH_RE.match(hash)
+        if not match or subject_hash.url[-8] not in ("/", "="):
+            hash = None
+    return SubjectHash(start_pos, closing_pos, hash, subject_hash.url)
 
 
 def _parse_subject_hash(text, start_pos, end_pos=None):