tags stashed on env

and using arguments, which normalizes spaces

src/sphinx_tags/__init__.py

@@ -3,6 +3,7 @@
 """
 import os
 import re
+from collections import defaultdict
 from fnmatch import fnmatch
 from pathlib import Path
 from typing import List
@@ -36,16 +37,18 @@ class TagLinks(SphinxDirective):
     separator = ","
 
     def run(self):
-        # Undo splitting args by whitespace, and use our own separator (to support tags with spaces)
-        tags = " ".join(self.arguments).split(self.separator)
-        tags = [t.strip() for t in tags]
+        tagline = " ".join(self.arguments).split(self.separator)
+        tags = [tag.strip() for tag in tagline]
 
         tag_dir = Path(self.env.app.srcdir) / self.env.app.config.tags_output_dir
         result = nodes.paragraph()
         result["classes"] = ["tags"]
         result += nodes.inline(text=f"{self.env.app.config.tags_intro_text} ")
         count = 0
 
+        current_doc_dir = Path(self.env.doc2path(self.env.docname)).parent
+        relative_tag_dir = Path(os.path.relpath(tag_dir, current_doc_dir))
+
         for tag in tags:
             count += 1
             # We want the link to be the path to the _tags folder, relative to
@@ -56,10 +59,8 @@ def run(self):
             #  - subfolder
             #   |
             #    - current_doc_path
-            current_doc_dir = Path(self.env.doc2path(self.env.docname)).parent
-            relative_tag_dir = Path(os.path.relpath(tag_dir, current_doc_dir))
-            file_basename = _normalize_tag(tag)
 
+            file_basename = _normalize_tag(tag)
             if self.env.app.config.tags_create_badges:
                 result += self._get_badge_node(tag, file_basename, relative_tag_dir)
                 tag_separator = " "
@@ -68,6 +69,10 @@ def run(self):
                 tag_separator = f"{self.separator} "
             if not count == len(tags):
                 result += nodes.inline(text=tag_separator)
+
+        # register tags to global metadata for document
+        self.env.metadata[self.env.docname]["tags"] = tags
+
         return [result]
 
     def _get_plaintext_node(
@@ -193,30 +198,11 @@ def create_file(
 
 
 class Entry:
-    """Extracted info from source file (*.rst/*.md/*.ipynb)"""
+    """Tags to pages map"""
 
-    def __init__(self, entrypath: Path):
+    def __init__(self, entrypath: Path, tags: list):
         self.filepath = entrypath
-        self.lines = self.filepath.read_text(encoding="utf8").split("\n")
-        if self.filepath.suffix == ".rst":
-            tagstart = ".. tags::"
-            tagend = ""
-        elif self.filepath.suffix == ".md":
-            tagstart = "```{tags}"
-            tagend = "```"
-        elif self.filepath.suffix == ".ipynb":
-            tagstart = '".. tags::'
-            tagend = '"'
-        else:
-            raise ValueError(
-                "Unknown file extension. Currently, only .rst, .md .ipynb are supported."
-            )
-        tagline = [line for line in self.lines if tagstart in line]
-        self.tags = []
-        if tagline:
-            tagline = tagline[0].replace(tagstart, "").rstrip(tagend)
-            self.tags = tagline.split(",")
-            self.tags = [tag.strip() for tag in self.tags]
+        self.tags = tags
 
     def assign_to_tags(self, tag_dict):
         """Append ourself to tags"""
@@ -245,6 +231,7 @@ def tagpage(tags, outdir, title, extension, tags_index_head):
     This page contains a list of all available tags.
 
     """
+
     tags = list(tags.values())
 
     if "md" in extension:
@@ -294,24 +281,18 @@ def assign_entries(app):
     pages = []
     tags = {}
 
-    # Get document paths in the project that match specified file extensions
-    doc_paths = get_matching_files(
-        app.srcdir,
-        include_patterns=[f"**.{extension}" for extension in app.config.tags_extension],
-        exclude_patterns=app.config.exclude_patterns,
-    )
-
-    for path in doc_paths:
-        entry = Entry(Path(app.srcdir) / path)
+    for docname in app.env.found_docs:
+        doctags = app.env.metadata[docname].get("tags", None)
+        if doctags is None:
+            continue  # skip if no tags
+        entry = Entry(app.env.doc2path(docname), doctags)
         entry.assign_to_tags(tags)
         pages.append(entry)
-
     return tags, pages
 
 
 def update_tags(app):
     """Update tags according to pages found"""
-
     if app.config.tags_create_tags:
         tags_output_dir = Path(app.config.tags_output_dir)
 
@@ -324,6 +305,7 @@ def update_tags(app):
 
         # Create pages for each tag
         tags, pages = assign_entries(app)
+
         for tag in tags.values():
             tag.create_file(
                 [item for item in pages if tag.name in item.tags],
@@ -386,4 +368,5 @@ def setup(app):
         "version": __version__,
         "parallel_read_safe": True,
         "parallel_write_safe": True,
+        "env_version": 1,
     }

test/conftest.py

@@ -15,7 +15,6 @@
 
 import pytest
 import sphinx.testing
-import sphinx.testing.path
 
 collect_ignore = ["sources", "outputs"]
 pytest_plugins = "sphinx.testing.fixtures"
@@ -31,11 +30,15 @@ def rootdir():
     resolved and copied as files.
     """
 
-    def copytree(src, dest):
-        shutil.copytree(src, dest, symlinks=True)
+    # patch copytree into Path objects to have sphinx 5 compatibility and
+    class PatchedPath(type(Path())):
+        def __new__(cls, *pathsegments):
+            return super().__new__(cls, *pathsegments)
 
-    with patch.object(sphinx.testing.path.path, "copytree", copytree):
-        yield sphinx.testing.path.path(SOURCE_ROOT_DIR)
+        def copytree(src, dest):
+            shutil.copytree(src, dest, symlinks=True)
+
+    yield PatchedPath(SOURCE_ROOT_DIR)
 
 
 @pytest.fixture(scope="session", autouse=True)

test/outputs/general/_tags/tag-4.txt

@@ -1,5 +1,5 @@
-My tags: [{(tag   4)}]
-**********************
+My tags: [{(tag 4)}]
+********************
 
 
 With this tag

test/outputs/general/_tags/tagsindex.txt

@@ -5,7 +5,7 @@ Tags overview
 Tags
 ^^^^
 
-* [{(tag   4)}] (1)
+* [{(tag 4)}] (1)
 
 * tag 3 (2)
 

test/outputs/general/index.txt

@@ -13,7 +13,7 @@ Test document
 
 * Tags overview
 
-  * [{(tag   4)}] (1)
+  * [{(tag 4)}] (1)
 
     * Page 1
 

test/sources/test-badges/conf.py

@@ -1,4 +1,4 @@
-extensions = ["sphinx_design", "sphinx_tags", "myst_parser"]
+extensions = ["sphinx_tags", "myst_parser", "sphinx_design"]
 tags_create_tags = True
 tags_extension = ["md"]
 tags_create_badges = True

test/test_badges.py

@@ -11,11 +11,17 @@
 EXPECTED_CLASSES = {
     "tag-1": "sd-bg-primary",
     "tag-2": "sd-bg-secondary",
-    "prefix-tag-3": "sd-bg-info",
-    "tag-4": "sd-bg-dark",
+    "prefix:tag-3": "sd-bg-info",
+    "tag 4": "sd-bg-dark",
 }
 
 
+@pytest.mark.sphinx("html", testroot="badges")
+def test_build(app: SphinxTestApp, status: StringIO, warning: StringIO):
+    app.build()
+    assert "build succeeded" in status.getvalue()
+
+
 @pytest.mark.sphinx("html", testroot="badges")
 def test_badges(app: SphinxTestApp, status: StringIO, warning: StringIO):
     """Parse output HTML for a page with badges, find badge links, and check for CSS classes for
@@ -27,11 +33,14 @@ def test_badges(app: SphinxTestApp, status: StringIO, warning: StringIO):
 
     build_dir = Path(app.srcdir) / "_build" / "html"
     page_1 = (build_dir / "page_1.html").read_text()
+    assert page_1 is not None
     soup = BeautifulSoup(page_1, "html.parser")
     # print(soup.prettify())
+    assert soup is not None
+    print(soup)
+    badge_links = soup.find_all("span", class_="sd-badge")
+    print("badge: ", badge_links)
 
-    badge_links = soup.find_all("a", class_="sd-badge")
-    classes_by_tag = {Path(a.get("href")).stem: a.get("class") for a in badge_links}
-
-    for tag, cls in EXPECTED_CLASSES.items():
-        assert cls in classes_by_tag[tag]
+    for (tag, class_), span in zip(EXPECTED_CLASSES.items(), badge_links):
+        assert tag in span.text  # hard to test tag 4 b/c of the icon
+        assert class_ in span["class"]

test/test_general_tags.py

@@ -19,6 +19,7 @@ def run_all_formats():
             pytest.param(marks=pytest.mark.sphinx("text", testroot="symlink")),
             pytest.param(marks=pytest.mark.sphinx("text", testroot="ipynb")),
         ],
+        ids=["myst", "rst", "symlink", "ipynb"],
     )
 
 
@@ -29,7 +30,7 @@ def test_build(app: SphinxTestApp, status: StringIO, warning: StringIO):
 
     # Build with notebooks and text output results in spurious errors "File Not Found: _tags/*.html"
     # For all other formats, ensure no warnings are raised
-    if not app.srcdir.endswith("ipynb"):
+    if not str(app.srcdir).endswith("ipynb"):
         assert not warning.getvalue().strip()