Fix arxiv stamp removal regex and add tests

2 files changed, 23 insertions, 9 deletions

diff --git a/paper2remarkable/providers/arxiv.py b/paper2remarkable/providers/arxiv.py
index e022658..1fd1795 100644
--- a/paper2remarkable/providers/arxiv.py
+++ b/paper2remarkable/providers/arxiv.py
@@ -19,6 +19,10 @@ from ..log import Logger
 
 logger = Logger()
 
+DEARXIV_TEXT_REGEX = (
+    b"arXiv:\d{4}\.\d{4,5}v\d+\s+\[[\w\-]+\.\w+\]\s+\d{1,2}\s\w{3}\s\d{4}"
+)
+
 
 class ArxivInformer(Informer):
     pass
@@ -73,11 +77,7 @@ class Arxiv(Provider):
         with open(uncompress_file, "rb") as fid:
             data = fid.read()
             # Remove the text element
-            data = re.sub(
-                b"\(arXiv:\d{4}\.\d{4,5}v\d+\s+\[\w+\.\w+\]\s+\d{1,2}\s\w{3}\s\d{4}\)Tj",
-                b"()Tj",
-                data,
-            )
+            data = re.sub(b"\(" + DEARXIV_TEXT_REGEX + b"\)Tj", b"()Tj", data)
             # Remove the URL element
             data = re.sub(
                 b"<<\\n\/URI \(http://arxiv\.org/abs/\d{4}\.\d{4,5}v\d+\)\\n\/S /URI\\n>>\\n",
diff --git a/tests/test_providers.py b/tests/test_providers.py
index 143fc78..1479967 100644
--- a/tests/test_providers.py
+++ b/tests/test_providers.py
@@ -5,11 +5,12 @@ __author__ = "G.J.J. van den Burg"
 
 """Tests"""
 
-import unittest
-import tempfile
 import hashlib
-import shutil
 import os
+import re
+import shutil
+import tempfile
+import unittest
 
 from paper2remarkable.providers import (
     ACM,
@@ -20,8 +21,9 @@ from paper2remarkable.providers import (
     PubMed,
     Springer,
 )
+from paper2remarkable.providers.arxiv import DEARXIV_TEXT_REGEX
 
-VERBOSE = True
+VERBOSE = False
 
 
 def md5sum(filename):
@@ -35,6 +37,18 @@ def md5sum(filename):
     return hasher.hexdigest()
 
 
+class TestArxiv(unittest.TestCase):
+    def test_text_regex_1(self):
+        key = b"arXiv:1908.03213v1 [astro.HE] 8 Aug 2019"
+        m = re.fullmatch(DEARXIV_TEXT_REGEX, key)
+        self.assertIsNotNone(m)
+
+    def test_text_regex_2(self):
+        key = b"arXiv:1908.03213v1 [astro-ph.HE] 8 Aug 2019"
+        m = re.fullmatch(DEARXIV_TEXT_REGEX, key)
+        self.assertIsNotNone(m)
+
+
 class TestProviders(unittest.TestCase):
     @classmethod
     def setUpClass(cls):