Don't parse links in code blocks or comments

pyproject.toml

@@ -23,6 +23,10 @@ jsonschema = "^4.23.0"
 natsort = "^8.4.0"
 emoji = "^2.12.1"
 livereload = "^2.7.0"
+markdown = "^3.7"
+beautifulsoup4 = "^4.12.3"
+types-markdown = "^3.7.0.20240822"
+types-beautifulsoup4 = "^4.12.0.20241020"
 
 [tool.poetry.group.dev.dependencies]
 bumpver = "^2023.1129"

src/mkslides/constants.py

@@ -4,46 +4,12 @@
 
 from jinja2 import Environment, FileSystemLoader, PackageLoader, select_autoescape
 
-MD_LINK_REGEX = re.compile(
-    r"""
-    \[.*?\]                     # Alt text
-    \((?P<location>[^()<>]*?)\) # Location without angle brackets
-    """,
-    re.VERBOSE,
-)
-
-MD_ESCAPED_LINK_REGEX = re.compile(
-    r"""
-    \[.*?\]                     # Alt text
-    \(<(?P<location>[^<>]*?)>\) # Location with mandatory angle brackets
-    """,
-    re.VERBOSE,
-)
-
-HTML_IMAGE_REGEX = re.compile(
-    r"""
-    <img                        # Start of the image
-    .+?                         # Any attributes
-    src=                        # src attribute
-        (?P<delimiter>['\"])    # Delimiter
-        (?P<location>.+?)       # Image location
-        (?P=delimiter)          # Delimiter
-    .*?                         # Any attributes
-    >                           # End of the image
-    """,
-    re.VERBOSE,
-)
-
 HTML_BACKGROUND_IMAGE_REGEX = re.compile(
     r"""
-    <!--                        # Start of the comment
-    .*?                         # Any content
     data-background-image=      # data-background-image attribute
         (?P<delimiter>['\"])    # Delimiter
         (?P<location>.+?)       # Image location
         (?P=delimiter)          # Delimiter
-    .*?                         # Any content
-    -->                         # End of the comment
     """,
     re.VERBOSE,
 )

src/mkslides/markupgenerator.py

@@ -9,6 +9,8 @@
 from urllib.parse import urlparse
 
 import frontmatter  # type: ignore[import-untyped]
+import markdown
+from bs4 import BeautifulSoup, Comment
 from emoji import emojize
 from natsort import natsorted
 
@@ -18,10 +20,7 @@
     DEFAULT_SLIDESHOW_TEMPLATE,
     HIGHLIGHTJS_THEMES_RESOURCE,
     HTML_BACKGROUND_IMAGE_REGEX,
-    HTML_IMAGE_REGEX,
     LOCAL_JINJA2_ENVIRONMENT,
-    MD_ESCAPED_LINK_REGEX,
-    MD_LINK_REGEX,
     REVEALJS_RESOURCE,
     REVEALJS_THEMES_RESOURCE,
 )
@@ -107,8 +106,8 @@ def __process_markdown_file(
         # Retrieve the frontmatter metadata and the markdown content
 
         content = md_file.read_text(encoding="utf-8-sig")
-        metadata, markdown = frontmatter.parse(content)
-        markdown = emojize(markdown, language="alias")
+        metadata, markdown_content = frontmatter.parse(content)
+        markdown_content = emojize(markdown_content, language="alias")
 
         # Get the relative path of reveal.js
 
@@ -181,15 +180,15 @@ def __process_markdown_file(
             highlight_theme=relative_highlight_theme_path,
             revealjs_path=relative_revealjs_path,
             markdown_data_options=markdown_data_options,
-            markdown=markdown,
+            markdown=markdown_content,
             revealjs_config=revealjs_config,
             plugins=plugins,
         )
         self.__create_file(output_markup_path, markup)
 
         # Copy local files
 
-        self.__copy_local_files(md_file, md_root_path, markdown)
+        self.__copy_local_files(md_file, md_root_path, markdown_content)
 
         return metadata, output_markup_path
 
@@ -250,20 +249,13 @@ def __copy_local_files(
         self,
         md_file: Path,
         md_root_path: Path,
-        markdown: str,
+        markdown_content: str,
     ) -> None:
-        for regex in [
-            MD_LINK_REGEX,
-            MD_ESCAPED_LINK_REGEX,
-            HTML_IMAGE_REGEX,
-            HTML_BACKGROUND_IMAGE_REGEX,
-        ]:
-            for m in regex.finditer(markdown):
-                location = m.group("location")
-
-                if self.__get_url_type(location) == URLType.RELATIVE:
-                    image = Path(md_file.parent, location).resolve(strict=True)
-                    self.__copy_to_output_relative_to_md_root(image, md_root_path)
+        links = self.__find_all_links(markdown_content)
+        for link in links:
+            if self.__get_url_type(link) == URLType.RELATIVE:
+                image = Path(md_file.parent, link).resolve(strict=True)
+                self.__copy_to_output_relative_to_md_root(image, md_root_path)
 
     def __copy_theme(
         self,
@@ -390,3 +382,23 @@ def __get_url_type(self, url: str) -> URLType:
             return URLType.ABSOLUTE
 
         return URLType.RELATIVE
+
+    def __find_all_links(self, markdown_content: str) -> set[str]:
+        html_content = markdown.markdown(markdown_content)
+        soup = BeautifulSoup(html_content, "html.parser")
+
+        found_links = set()
+
+        for link in soup.find_all("a", href=True):
+            if not link.find_parents(["code", "pre"]):
+                found_links.add(link["href"])
+
+        for link in soup.find_all("img", src=True):
+            if not link.find_parents(["code", "pre"]):
+                found_links.add(link["src"])
+
+        for comment in soup.find_all(string=lambda text: isinstance(text, Comment)):
+            if match := HTML_BACKGROUND_IMAGE_REGEX.search(comment):
+                found_links.add(match.group("location"))
+
+        return found_links

tests/test_files/someslides.md

@@ -79,6 +79,39 @@ Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor i
 
 ## Don't try to parse anchor links
 
-[Go to anchor](#anchor)
+This should not throw an error:
+
+[Go to anchor](#some-random-anchor)
 
 ---
+
+## Don't try to parse links in code blocks
+
+This should not throw an error:
+
+`[test](./some-random-md-link)`
+`![test](./some-random-md-image)`
+`<a href="./some-random-html-link">test</a>`
+`<img src="./some-random-html-image" />`
+
+```markdown
+[test](./some-random-md-link)
+![test](./some-random-md-image)
+```
+
+```html
+<a href="./some-random-html-link">test</a>
+<img src="./some-random-html-image" />
+```
+
+    [test](./some-random-md-link)
+    ![test](./some-random-md-image)
+    <a href="./some-random-html-link">test</a>
+    <img src="./some-random-html-image" />
+
+<!--
+[test](./some-random-md-link)
+![test](./some-random-md-image)
+<a href="./some-random-html-link">test</a>
+<img src="./some-random-html-image" />
+-->