FIX broken links from cards on landing page (microsoft#1140)

Author: romanlutz

build_scripts/check_links.py

@@ -27,13 +27,21 @@
 # Updated regex pattern to capture URLs from Markdown and HTML
 URL_PATTERN = re.compile(r'\[.*?\]\((.*?)\)|href="([^"]+)"|src="([^"]+)"')
 
+# Pattern to capture :link: directives from MyST grid-item-cards
+GRID_LINK_PATTERN = re.compile(r"^:link:\s+(.+)$", re.MULTILINE)
+
 
 def extract_urls(file_path):
     with open(file_path, "r", encoding="utf-8") as file:
         content = file.read()
     matches = URL_PATTERN.findall(content)
     # Flatten the list of tuples and filter out empty strings
     urls = [strip_fragment(url) for match in matches for url in match if url]
+
+    # Extract :link: directives from MyST grid-item-cards
+    grid_links = GRID_LINK_PATTERN.findall(content)
+    urls.extend(grid_links)
+
     return urls
 
 
@@ -47,7 +55,17 @@ def strip_fragment(url):
 
 def resolve_relative_url(base_path, url):
     if not url.startswith(("http://", "https://", "mailto:", "attachment:")):
-        return os.path.abspath(os.path.join(os.path.dirname(base_path), url))
+        # Handle MyST doc references (e.g., setup/1b_install_docker)
+        # These can be .md, .rst, or directory paths
+        abs_path = os.path.abspath(os.path.join(os.path.dirname(base_path), url))
+
+        # Check various possible file extensions for doc links
+        if not os.path.exists(abs_path):
+            for ext in [".md", ".ipynb"]:
+                if os.path.exists(abs_path + ext):
+                    return abs_path + ext
+
+        return abs_path
     return url
 
 
@@ -74,6 +92,10 @@ def check_url(url, retries=2, delay=2):
     ):
         return url, True
 
+    # If it's not an HTTP URL at this point, it's likely a broken local file reference
+    if not url.startswith(("http://", "https://")):
+        return url, False
+
     attempts = 0
     while attempts <= retries:
         try:
@@ -91,6 +113,9 @@ def check_url(url, retries=2, delay=2):
                 return url, False
             time.sleep(delay)
 
+    # If we exit the loop without returning, the URL is broken
+    return url, False
+
 
 def check_links_in_file(file_path):
     urls = extract_urls(file_path)

doc/_config.yml

@@ -72,5 +72,5 @@ sphinx:
         - "https://scikit-learn.org/stable"
         - null
       pytorch:
-        - "https://pytorch.org/docs/stable"
+        - "https://docs.pytorch.org/docs/stable"
         - null

doc/contributing/11_release_process.md

@@ -49,9 +49,9 @@ links work properly. Note: There may not be any links to update, but it is
 good practice to check in case our README changes.
 
 Replace all "main" links like
-"https://github.com/Azure/PyRIT/blob/main/doc/README.md" with "raw" links that have
+"https://github.com/Azure/PyRIT/blob/main/doc/index.md" with "raw" links that have
 the correct version number, i.e.,
-"https://raw.githubusercontent.com/Azure/PyRIT/releases/vx.y.z/doc/README.md".
+"https://raw.githubusercontent.com/Azure/PyRIT/releases/vx.y.z/doc/index.md".
 
 For images, update using the "raw" link, e.g.,
 "https://raw.githubusercontent.com/Azure/PyRIT/releases/vx.y.z/assets/pyrit_architecture.png".

doc/index.md

@@ -26,8 +26,10 @@ PyRIT offers flexible installation options to suit different needs. Choose the p
 :gutter: 3
 
 :::{grid-item-card} 🐋 Docker Installation
-:link: ./setup/1b_install_docker.md
+:link: setup/1b_install_docker
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 **For Users - Quick Start** ⭐
 
@@ -47,8 +49,10 @@ Get started immediately with a pre-configured environment:
 :::
 
 :::{grid-item-card} 🐍 Local Pip/Conda Installation
-:link: ./setup/1a_install_conda.md
+:link: setup/1a_install_conda
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 **For Users - Custom Setup**
 
@@ -68,8 +72,10 @@ Install PyRIT directly on your machine:
 :::
 
 :::{grid-item-card} 🐋 DevContainers in VS Code
-:link: ./contributing/1b_install_devcontainers.md
+:link: contributing/1b_install_devcontainers
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 **For Contributors** ⭐
 
@@ -89,8 +95,10 @@ Standardized development environment:
 :::
 
 :::{grid-item-card} 🐍 Local Conda Development
-:link: ./contributing/1a_install_conda.md#local-installation-with-condapython
+:link: contributing/1a_install_conda
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 **For Contributors - Custom Dev Setup**
 
@@ -119,72 +127,90 @@ There is no single way to read the documentation, and it's perfectly fine to jum
 :gutter: 2
 
 :::{grid-item-card} 1️⃣ Cookbooks
-:link: ./cookbooks/README.md
+:link: cookbooks/README
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 This provides an overview of PyRIT in action. A great place to see practical examples and get started quickly.
 
 :::
 
 :::{grid-item-card} 2️⃣ Architecture
-:link: ./code/architecture.md
+:link: code/architecture
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 High-level overview of all PyRIT components. Understanding any single component is easier with knowledge of the others.
 
 :::
 
 :::{grid-item-card} 3️⃣ Attacks
-:link: ./code/executor/0_executor.md
+:link: code/executor/0_executor
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Top-level components implementing end-to-end attack techniques. Shows how all PyRIT components fit together.
 
 :::
 
 :::{grid-item-card} 4️⃣ Datasets
-:link: ./code/datasets/0_dataset.md
+:link: code/datasets/0_dataset
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Building attacks using seed prompts and fetching datasets. The foundation for creating test scenarios.
 
 :::
 
 :::{grid-item-card} 5️⃣ Targets
-:link: ./code/targets/0_prompt_targets.md
+:link: code/targets/0_prompt_targets
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Endpoints that PyRIT sends prompts to. Essential for nearly any PyRIT scenario - learn what targets are available.
 
 :::
 
 :::{grid-item-card} 6️⃣ Converters
-:link: ./code/converters/0_converters.ipynb
+:link: code/converters/0_converters
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Transform prompts from one format to another. One of the most powerful capabilities within PyRIT.
 
 :::
 
 :::{grid-item-card} 7️⃣ Scorers
-:link: ./code/scoring/0_scoring.md
+:link: code/scoring/0_scoring
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 How PyRIT makes decisions and records output. Essential for evaluating AI system responses.
 
 :::
 
 :::{grid-item-card} 8️⃣ Memory
-:link: ./code/memory/0_memory.md
+:link: code/memory/0_memory
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 How PyRIT components communicate state information. Understand the data flow and storage mechanisms.
 
 :::
 
 :::{grid-item-card} 9️⃣ Auxiliary Attacks (Optional)
-:link: ./code/auxiliary_attacks/0_auxiliary_attacks.ipynb
+:link: code/auxiliary_attacks/0_auxiliary_attacks
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Attacks and techniques that don't fit into core PyRIT functionality. Advanced and specialized methods.
 
@@ -198,16 +224,20 @@ Attacks and techniques that don't fit into core PyRIT functionality. Advanced an
 :gutter: 2
 
 :::{grid-item-card} 🚀 Deployment
-:link: ./deployment/README.md
+:link: deployment/README
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Code to download, deploy, and score open-source models (such as Hugging Face) on Azure.
 
 :::
 
 :::{grid-item-card} 📰 Blog
-:link: ./blog/README.md
+:link: blog/README
+:shadow: md
 :class-header: bg-light
+:link-type: doc
 
 Notable new changes and updates. Stay current with the latest PyRIT developments.
 

pyrit/models/storage_io.py

@@ -215,7 +215,7 @@ async def read_file(self, path: Union[Path, str]) -> bytes:
         Asynchronously reads the content of a file (blob) from Azure Blob Storage.
 
         If the provided `path` is a full URL
-        (e.g., "https://<Azure STorage Account>.blob.core.windows.net/<container name>/dir1/dir2/sample.png"),
+        (e.g., "https://account.blob.core.windows.net/container/dir1/dir2/sample.png"),
         it extracts the relative blob path (e.g., "dir1/dir2/sample.png") to correctly access the blob.
         If a relative path is provided, it will use it as-is.
 

tests/unit/score/test_true_false_composite_scorer.py

@@ -164,6 +164,6 @@ async def test_composite_scorer_with_task(mock_request, true_scorer):
 
 
 def test_composite_scorer_empty_scorers_list():
-    """Test that CompositeScorer raises an exception when given an empty list of scorers."""
+    """Test that TrueFalseCompositeScorer raises an exception when given an empty list of scorers."""
     with pytest.raises(ValueError, match="At least one scorer must be provided"):
         TrueFalseCompositeScorer(aggregator=TrueFalseScoreAggregator.AND, scorers=[])