From 0517d2c5a0481c62bce23a6f15495e0ceec59bb5 Mon Sep 17 00:00:00 2001
From: MohammadYusif <MohammadYusif@users.noreply.github.com>
Date: Sun, 7 Jun 2026 20:57:11 +0300
Subject: [PATCH] fix: ignore URL fragments in markdown links when parsing tags
 (#52)

get_tags removed wikilinks before tag parsing but left standard markdown
inline links intact, so a URL fragment such as
'.../features.html#hot-module-replacement' was matched by the tag regex
and surfaced as a spurious tag ('hot-').

Strip the link URLs (keeping the link text) before tag extraction, using
the existing INLINE_LINK_AFTER_HTML_PROC_REGEX, mirroring how wikilinks
are already handled. Genuine tags, including a '#tag' written in a link's
display text, are still captured.
---
 obsidiantools/md_utils.py          |  8 ++++++++
 tests/general/tags_url-fragment.md |  7 +++++++
 tests/test_md_utils.py             | 16 ++++++++++++++++
 3 files changed, 31 insertions(+)
 create mode 100644 tests/general/tags_url-fragment.md

diff --git a/obsidiantools/md_utils.py b/obsidiantools/md_utils.py
index 50a6c3a..058932e 100644
--- a/obsidiantools/md_utils.py
+++ b/obsidiantools/md_utils.py
@@ -244,6 +244,8 @@ def get_tags(filepath: Path, *, show_nested: bool = False) -> list[str]:
         str_transform_func=_transform_md_file_string_for_tag_parsing)
     # remove wikilinks so that '#' headers are not caught:
     src_txt = _remove_wikilinks_from_source_text(src_txt)
+    # remove md link URLs so that '#' fragments in them are not caught:
+    src_txt = _remove_md_link_urls_from_source_text(src_txt)
     tags = _get_tags_from_source_text(src_txt, show_nested=show_nested)
     return tags
 
@@ -447,6 +449,12 @@ def _remove_wikilinks_from_source_text(src_txt: str) -> str:
     return re.sub(WIKILINK_REGEX, '', src_txt)
 
 
+def _remove_md_link_urls_from_source_text(src_txt: str) -> str:
+    # replace '[text](<url>)' with its 'text' so that any '#' fragment
+    # in the URL is not mistaken for a tag, while keeping the link text:
+    return re.sub(INLINE_LINK_AFTER_HTML_PROC_REGEX, r'\1', src_txt)
+
+
 def _transform_md_file_string_for_tag_parsing(txt: str) -> str:
     return txt.replace('\\#', '')
 
diff --git a/tests/general/tags_url-fragment.md b/tests/general/tags_url-fragment.md
new file mode 100644
index 0000000..05a741d
--- /dev/null
+++ b/tests/general/tags_url-fragment.md
@@ -0,0 +1,7 @@
+# Tags vs URL fragments
+
+thanks to [Vite's HMR feature](https://vitejs.dev/guide/features.html#hot-module-replacement).
+
+A real tag: #python
+
+Another link [docs](https://example.com/page#section-two) followed by #data text.
diff --git a/tests/test_md_utils.py b/tests/test_md_utils.py
index 4a3ed1f..f87290b 100644
--- a/tests/test_md_utils.py
+++ b/tests/test_md_utils.py
@@ -260,6 +260,22 @@ def test_sussudio_tags_with_nesting_shown():
     assert actual_tags == expected_tags
 
 
+def test_url_fragments_in_md_links_not_parsed_as_tags():
+    # the '#' fragment of a markdown link URL must not be read as a tag,
+    # while genuine tags in the note are still captured (issue #52):
+    actual_tags = get_tags(
+        Path('.') / 'tests/general/tags_url-fragment.md')
+    expected_tags = ['python', 'data']
+    assert actual_tags == expected_tags
+
+
+def test_url_fragments_in_md_links_not_parsed_as_tags_with_nesting():
+    actual_tags = get_tags(
+        Path('.') / 'tests/general/tags_url-fragment.md', show_nested=True)
+    expected_tags = ['python', 'data']
+    assert actual_tags == expected_tags
+
+
 def test_embedded_files_alias_scaling():
     actual_embedded_images = get_embedded_files(
         Path('.') / 'tests/general/embedded-images_in-table.md')