Fix URL encoded filenames

CHANGELOG.rst

@@ -3,6 +3,16 @@
 Changelog
 =========
 
+2.4.1 (2021-12-23)
+------------------
+
+- Fix URL decoding of filenames.
+  `#104 <https://github.com/hynek/doc2dash/pull/104>`_
+
+
+----
+
+
 2.4.0 (2021-11-16)
 ------------------
 

src/doc2dash/parsers/utils.py

@@ -3,6 +3,7 @@
 import errno
 import logging
 import os
+import urllib
 
 from collections import defaultdict
 
@@ -108,7 +109,9 @@ def patch_anchors(parser, show_progressbar):
 
     def patch_files(files):
         for fname, entries in files:
-            full_path = os.path.join(parser.doc_path, fname)
+            full_path = os.path.join(
+                parser.doc_path, urllib.parse.unquote(fname)
+            )
             with codecs.open(full_path, mode="r", encoding="utf-8") as fp:
                 soup = BeautifulSoup(fp, "html.parser")
                 for entry in entries:

tests/parsers/test_utils.py

@@ -45,6 +45,14 @@ def entries():
     ]
 
 
+@pytest.fixture
+def entries_url_format():
+    return [
+        ParserEntry(name="foo", type="Method", path="foo%20bar.html#foo"),
+        ParserEntry(name="qux", type="Class", path="foo%20bar.html"),
+    ]
+
+
 class TestPatchTOCAnchors:
     @pytest.mark.parametrize("progressbar", [True, False])
     def test_with_empty_db(self, progressbar):
@@ -71,6 +79,23 @@ def test_single_entry(self, monkeypatch, tmpdir, entries):
             TOCEntry(name="foo", type="Method", anchor="foo")
         ] == parser._patched_entries
 
+    def test_single_entry_url_format(
+        self, monkeypatch, tmpdir, entries_url_format
+    ):
+    """
+    URL-encoded filenames are decoded before being added.
+    """
+        foo = tmpdir.mkdir("foo")
+        foo.join("foo bar.html").write("docs!")
+        parser = FakeParser(doc_path=str(foo))
+        toc = patch_anchors(parser, show_progressbar=False)
+        for e in entries_url_format:
+            toc.send(e)
+        toc.close()
+        assert [
+            TOCEntry(name="foo", type="Method", anchor="foo")
+        ] == parser._patched_entries
+
     def test_complains(self, entries, tmpdir):
         """
         If patching fails, a debug message is logged.