Merge pull request #385 from ocaisa/schema

boegel · web-flow · commit f6517a3b1c2c · 2025-02-20T16:26:38.000+01:00
Add `ld+json` metadata to software detail pages
diff --git a/.gitignore b/.gitignore
@@ -1 +1,3 @@
 site/
+venv*
+
diff --git a/mkdocs-ldjson-plugin/__init__.py b/mkdocs-ldjson-plugin/__init__.py
@@ -0,0 +1 @@
+from .inject_ld_json import InjectLDJsonPlugin
diff --git a/mkdocs-ldjson-plugin/inject_ld_json.py b/mkdocs-ldjson-plugin/inject_ld_json.py
@@ -0,0 +1,28 @@
+import json
+from mkdocs.plugins import BasePlugin
+from bs4 import BeautifulSoup
+
+class InjectLDJsonPlugin(BasePlugin):
+    def on_post_page(self, output, page, config):
+        """Inject JSON-LD metadata into the <head> section of the rendered page."""
+        
+        # Get JSON-LD data from page metadata
+        json_ld = page.meta.get("json_ld")
+        if not json_ld:
+            return output  # No JSON-LD to inject, return page as is
+        
+        # Convert dictionary to JSON string
+        json_ld_str = json.dumps(json_ld, indent=2)
+
+        # Parse the HTML output
+        soup = BeautifulSoup(output, "html.parser")
+        
+        # Create <script> tag for JSON-LD
+        script_tag = soup.new_tag("script", type="application/ld+json")
+        script_tag.string = json_ld_str
+
+        # Find the <head> and insert the script
+        if soup.head:
+            soup.head.append(script_tag)
+
+        return str(soup)
diff --git a/mkdocs-ldjson-plugin/requirements.txt b/mkdocs-ldjson-plugin/requirements.txt
@@ -0,0 +1,2 @@
+mkdocs
+beautifulsoup4
diff --git a/mkdocs-ldjson-plugin/setup.py b/mkdocs-ldjson-plugin/setup.py
@@ -0,0 +1,20 @@
+from setuptools import setup, find_packages
+
+# Read requirements from requirements.txt
+def read_requirements():
+    with open("requirements.txt", encoding="utf-8") as f:
+        return f.read().splitlines()
+
+setup(
+    name="inject_ld_json",
+    version="0.1.0",
+    py_modules=["inject_ld_json"],
+    packages=find_packages(),
+    install_requires=read_requirements(),  # Use requirements.txt
+    include_package_data=True,
+    entry_points={
+        "mkdocs.plugins": [
+            "inject_ld_json = inject_ld_json:InjectLDJsonPlugin",
+        ]
+    },
+)
diff --git a/mkdocs.yml b/mkdocs.yml
@@ -100,6 +100,8 @@ plugins:
         software_layer/adding_software.md: adding_software/overview.md
         pilot.md: repositories/pilot.md
         gpu.md: site_specific_config/gpu.md
+  # Enable our custom plugin for json-ld metadata
+  - inject_ld_json
 markdown_extensions:
   # enable adding HTML attributes and CSS classes
   - attr_list
diff --git a/requirements.txt b/requirements.txt
@@ -4,3 +4,4 @@ mkdocs-redirects
 mkdocs-git-revision-date-localized-plugin
 mkdocs-toc-sidebar-plugin
 mkdocs-macros-plugin
+./mkdocs-ldjson-plugin
diff --git a/scripts/available_software/available_software.py b/scripts/available_software/available_software.py
@@ -11,13 +11,16 @@
 @author: Michiel Lachaert (Ghent University)
 @author: Lara Peeters (Ghent University)
 """
+import copy
 import json
 import os
 import re
 import subprocess
 import sys
 import time
+import yaml
 from typing import Union, Tuple
+from string import Template
 import numpy as np
 from mdutils.mdutils import MdUtils
 from natsort import natsorted
@@ -339,6 +342,41 @@ def generate_software_table_data(software_data: dict, targets: list) -> list:
     return table_data
 
 
+# LD+JSON Template with placeholders
+ldjson_template = Template("""
+{
+    "json_ld": {
+        "@context": "https://schema.org",
+        "@type": "SoftwareApplication",
+        "name": "$name",
+        "url": "$homepage",
+        "softwareVersion": "$version",
+        "description": "$description",
+        "operatingSystem": "LINUX",
+        "applicationCategory": "DeveloperApplication",
+        "softwareRequirements": "See https://www.eessi.io/docs/ for how to make EESSI available on your system",
+        "license": "Not confirmed",
+        "review": {
+            "@type": "Review",
+            "reviewRating": {
+                "@type": "Rating",
+                "ratingValue": 5
+            },
+            "author": {
+                "@type": "Organization",
+                "name": "EESSI"
+            },
+            "reviewBody": "Application has been successfully made available on all architectures supported by EESSI"
+        },
+        "offers": {
+            "@type": "Offer",
+            "price": 0
+        }
+    }
+}
+""")
+
+
 def generate_software_detail_page(
         software_name: str,
         software_data: dict,
@@ -357,15 +395,20 @@ def generate_software_detail_page(
     """
     sorted_versions = dict_sort(software_data["versions"])
     newest_version = list(sorted_versions.keys())[-1]
+    ldjson_software_data = copy.deepcopy(software_data)
 
     filename = f"{path}/{software_name}.md"
     md_file = MdUtils(file_name=filename, title=f"{software_name}")
     if 'description' in software_data.keys():
         description = software_data['description']
         md_file.new_paragraph(f"{description}")
+    else:
+        ldjson_software_data['description'] = ''
     if 'homepage' in software_data.keys():
         homepage = software_data['homepage']
         md_file.new_paragraph(f"{homepage}")
+    else:
+        ldjson_software_data["homepage"] = ''
 
     md_file.new_header(level=1, title="Available modules")
 
@@ -392,11 +435,21 @@ def generate_software_detail_page(
 
     md_file.create_md_file()
 
-    # Remove the TOC
     with open(filename) as f:
         read_data = f.read()
     with open(filename, 'w') as f:
-        f.write("---\nhide:\n  - toc\n---\n" + read_data)
+        # Add the software name
+        ldjson_software_data['name'] = software_name
+        # Just output the supported versions (with toolchains)
+        ldjson_software_data["version"] = list(sorted_versions.keys())
+        # Make the description safe for json (and remove surrounding quotes)
+        ldjson_software_data['description'] = json.dumps(ldjson_software_data['description'])[1:-1]
+        json_str = ldjson_template.substitute(ldjson_software_data)  # Replace placeholders
+        json_topmatter = json.loads(json_str)
+        # Remove the TOC
+        json_topmatter["hide"] = ["toc"]
+        yaml_topmatter = yaml.dump(json_topmatter)
+        f.write("---\n" + yaml_topmatter + "---\n" + read_data)
 
 
 def generate_detail_pages(json_path, dest_path) -> None:
diff --git a/scripts/available_software/requirements.txt b/scripts/available_software/requirements.txt
@@ -1,3 +1,4 @@
 mdutils
 numpy
-natsort
+natsort
+pyyaml
diff --git a/scripts/available_software/tests/data/test_md_template_detailed_science_sol.md b/scripts/available_software/tests/data/test_md_template_detailed_science_sol.md
@@ -1,6 +1,31 @@
 ---
 hide:
-  - toc
+- toc
+json_ld:
+  '@context': https://schema.org
+  '@type': SoftwareApplication
+  applicationCategory: DeveloperApplication
+  description: ''
+  license: Not confirmed
+  name: science
+  offers:
+    '@type': Offer
+    price: 0
+  operatingSystem: LINUX
+  review:
+    '@type': Review
+    author:
+      '@type': Organization
+      name: EESSI
+    reviewBody: Application has been successfully made available on all architectures
+      supported by EESSI
+    reviewRating:
+      '@type': Rating
+      ratingValue: 5
+  softwareRequirements: See https://www.eessi.io/docs/ for how to make EESSI available
+    on your system
+  softwareVersion: '[''science/5.3.0'', ''science/7.2.0'']'
+  url: ''
 ---
 
 science

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+from .inject_ld_json import InjectLDJsonPlugin`
-Original file line number
+Diff line change
@@ @@ -1,3 +1,4 @@ @@
 mdutils
 numpy
 -natsort
 +natsort
 +pyyaml