rdf-construct 0.2.0__py3-none-any.whl

rdf_construct/docs/search.py

@@ -0,0 +1,278 @@
+"""Search index generation for client-side documentation search."""
+
+from __future__ import annotations
+
+import json
+import re
+from dataclasses import asdict, dataclass
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from rdf_construct.docs.config import DocsConfig
+    from rdf_construct.docs.extractors import ClassInfo, ExtractedEntities, InstanceInfo, PropertyInfo
+
+
+@dataclass
+class SearchEntry:
+    """A single entry in the search index."""
+
+    uri: str
+    qname: str
+    entity_type: str  # class, object_property, datatype_property, instance
+    label: str
+    keywords: list[str]
+    url: str
+
+
+def extract_keywords(text: str | None) -> list[str]:
+    """Extract searchable keywords from text.
+
+    Splits on whitespace and punctuation, lowercases, and removes
+    common stop words.
+
+    Args:
+        text: Text to extract keywords from.
+
+    Returns:
+        List of keyword strings.
+    """
+    if not text:
+        return []
+
+    # Split on non-alphanumeric characters
+    words = re.split(r"[^a-zA-Z0-9]+", text.lower())
+
+    # Remove stop words and short words
+    stop_words = {
+        "a", "an", "the", "is", "are", "was", "were", "be", "been",
+        "being", "have", "has", "had", "do", "does", "did", "will",
+        "would", "could", "should", "may", "might", "must", "shall",
+        "can", "need", "of", "to", "in", "for", "on", "with", "at",
+        "by", "from", "as", "or", "and", "but", "if", "this", "that",
+        "which", "who", "whom", "whose", "what", "when", "where", "why",
+        "how", "all", "each", "every", "both", "few", "more", "most",
+        "other", "some", "such", "no", "not", "only", "same", "so",
+        "than", "too", "very", "just", "also", "any",
+    }
+
+    keywords = [w for w in words if w and len(w) > 2 and w not in stop_words]
+    return list(set(keywords))  # Remove duplicates
+
+
+def class_to_search_entry(
+    class_info: "ClassInfo",
+    config: "DocsConfig",
+) -> SearchEntry:
+    """Convert a ClassInfo to a SearchEntry.
+
+    Args:
+        class_info: Class information.
+        config: Documentation configuration.
+
+    Returns:
+        SearchEntry for the class.
+    """
+    # Build keywords from various sources
+    keywords = []
+
+    # QName parts
+    if ":" in class_info.qname:
+        prefix, local = class_info.qname.split(":", 1)
+        keywords.extend(extract_keywords(local))
+        keywords.append(prefix.lower())
+
+    # Label
+    if class_info.label:
+        keywords.extend(extract_keywords(class_info.label))
+
+    # Definition
+    if class_info.definition:
+        keywords.extend(extract_keywords(class_info.definition))
+
+    # Superclass names
+    for uri in class_info.superclasses:
+        if "#" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("#")[-1]))
+        elif "/" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("/")[-1]))
+
+    from .config import entity_to_url
+    return SearchEntry(
+        uri=str(class_info.uri),
+        qname=class_info.qname,
+        entity_type="class",
+        label=class_info.label or class_info.qname,
+        keywords=list(set(keywords)),
+        url=entity_to_url(class_info.qname, "class", config),
+    )
+
+
+def property_to_search_entry(
+    prop_info: "PropertyInfo",
+    config: "DocsConfig",
+) -> SearchEntry:
+    """Convert a PropertyInfo to a SearchEntry.
+
+    Args:
+        prop_info: Property information.
+        config: Documentation configuration.
+
+    Returns:
+        SearchEntry for the property.
+    """
+    keywords = []
+
+    # QName parts
+    if ":" in prop_info.qname:
+        prefix, local = prop_info.qname.split(":", 1)
+        keywords.extend(extract_keywords(local))
+        keywords.append(prefix.lower())
+
+    # Label
+    if prop_info.label:
+        keywords.extend(extract_keywords(prop_info.label))
+
+    # Definition
+    if prop_info.definition:
+        keywords.extend(extract_keywords(prop_info.definition))
+
+    # Property type
+    keywords.append(prop_info.property_type)
+
+    # Domain/range class names
+    for uri in prop_info.domain + prop_info.range:
+        if "#" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("#")[-1]))
+        elif "/" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("/")[-1]))
+
+    entity_type = f"{prop_info.property_type}_property"
+
+    from .config import entity_to_url
+    return SearchEntry(
+        uri=str(prop_info.uri),
+        qname=prop_info.qname,
+        entity_type=entity_type,
+        label=prop_info.label or prop_info.qname,
+        keywords=list(set(keywords)),
+        url=entity_to_url(prop_info.qname, entity_type, config),
+    )
+
+
+def instance_to_search_entry(
+    instance_info: "InstanceInfo",
+    config: "DocsConfig",
+) -> SearchEntry:
+    """Convert an InstanceInfo to a SearchEntry.
+
+    Args:
+        instance_info: Instance information.
+        config: Documentation configuration.
+
+    Returns:
+        SearchEntry for the instance.
+    """
+    keywords = []
+
+    # QName parts
+    if ":" in instance_info.qname:
+        prefix, local = instance_info.qname.split(":", 1)
+        keywords.extend(extract_keywords(local))
+        keywords.append(prefix.lower())
+
+    # Label
+    if instance_info.label:
+        keywords.extend(extract_keywords(instance_info.label))
+
+    # Definition
+    if instance_info.definition:
+        keywords.extend(extract_keywords(instance_info.definition))
+
+    # Type names
+    for uri in instance_info.types:
+        if "#" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("#")[-1]))
+        elif "/" in str(uri):
+            keywords.extend(extract_keywords(str(uri).split("/")[-1]))
+
+    # Add "instance" as a keyword
+    keywords.append("instance")
+
+    from .config import entity_to_url
+    return SearchEntry(
+        uri=str(instance_info.uri),
+        qname=instance_info.qname,
+        entity_type="instance",
+        label=instance_info.label or instance_info.qname,
+        keywords=list(set(keywords)),
+        url=entity_to_url(instance_info.qname, "instance", config),
+    )
+
+
+def generate_search_index(
+    entities: "ExtractedEntities",
+    config: "DocsConfig",
+) -> list[SearchEntry]:
+    """Generate a search index from extracted entities.
+
+    Args:
+        entities: All extracted entities from the ontology.
+        config: Documentation configuration.
+
+    Returns:
+        List of SearchEntry objects for all entities.
+    """
+    entries: list[SearchEntry] = []
+
+    # Classes
+    if config.include_classes:
+        for class_info in entities.classes:
+            entries.append(class_to_search_entry(class_info, config))
+
+    # Properties
+    if config.include_object_properties:
+        for prop_info in entities.object_properties:
+            entries.append(property_to_search_entry(prop_info, config))
+
+    if config.include_datatype_properties:
+        for prop_info in entities.datatype_properties:
+            entries.append(property_to_search_entry(prop_info, config))
+
+    if config.include_annotation_properties:
+        for prop_info in entities.annotation_properties:
+            entries.append(property_to_search_entry(prop_info, config))
+
+    # Instances
+    if config.include_instances:
+        for instance_info in entities.instances:
+            entries.append(instance_to_search_entry(instance_info, config))
+
+    return entries
+
+
+def write_search_index(
+    entries: list[SearchEntry],
+    output_dir: Path,
+) -> Path:
+    """Write the search index to a JSON file.
+
+    Args:
+        entries: List of search entries.
+        output_dir: Output directory.
+
+    Returns:
+        Path to the written search index file.
+    """
+    output_path = output_dir / "search.json"
+
+    # Convert to serialisable format
+    data = {
+        "entities": [asdict(entry) for entry in entries],
+    }
+
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    with open(output_path, "w") as f:
+        json.dump(data, f, indent=2)
+
+    return output_path

rdf_construct/docs/templates/html/base.html.jinja

@@ -0,0 +1,44 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>{% block title %}{{ ontology.title or "Ontology Documentation" }}{% endblock %}</title>
+    <link rel="stylesheet" href="{{ config.base_url }}/assets/style.css">
+    {% block head %}{% endblock %}
+</head>
+<body>
+    <header class="header">
+        <h1>{{ ontology.title or "Ontology Documentation" }}</h1>
+        {% if ontology.description %}
+        <p class="description">{{ ontology.description }}</p>
+        {% endif %}
+    </header>
+
+    <nav class="nav">
+        <a href="{{ config.base_url }}/index.html">Index</a>
+        <a href="{{ config.base_url }}/hierarchy.html">Hierarchy</a>
+        <a href="{{ config.base_url }}/namespaces.html">Namespaces</a>
+    </nav>
+
+    {% if config.include_search %}
+    <div class="search-box">
+        <input type="text" id="search-input" placeholder="Search entities...">
+        <ul id="search-results" class="search-results"></ul>
+    </div>
+    {% endif %}
+
+    <main>
+        {% block content %}{% endblock %}
+    </main>
+
+    <footer class="footer">
+        <p>Generated by <a href="https://github.com/aigora-de/rdf-construct">rdf-construct</a></p>
+    </footer>
+
+    {% if config.include_search %}
+    <script src="{{ config.base_url }}/assets/search.js"></script>
+    {% endif %}
+    {% block scripts %}{% endblock %}
+</body>
+</html>

rdf_construct/docs/templates/html/class.html.jinja

@@ -0,0 +1,152 @@
+{% extends "base.html.jinja" %}
+
+{% block title %}{{ class_info.label or class_info.qname }} - {{ ontology.title or "Documentation" }}{% endblock %}
+
+{% block content %}
+<article class="entity-card">
+    <h1>{{ class_info.label or class_info.qname }} <span class="entity-type">Class</span></h1>
+    <p class="uri">{{ class_info.uri }}</p>
+
+    {% if class_info.definition %}
+    <p class="definition">{{ class_info.definition }}</p>
+    {% endif %}
+
+    {% if class_info.superclasses %}
+    <section class="section">
+        <h3>Superclasses</h3>
+        <ul class="entity-list">
+            {% for uri in class_info.superclasses %}
+            <li>
+                {% set found = namespace(value=false) %}
+                {% for c in classes %}
+                    {% if c.uri|string == uri|string %}
+                    <a href="{{ c.qname | entity_url('class') }}">{{ c.label or c.qname }}</a>
+                    {% set found.value = true %}
+                    {% endif %}
+                {% endfor %}
+                {% if not found.value %}
+                <code>{{ uri }}</code>
+                {% endif %}
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if class_info.subclasses %}
+    <section class="section">
+        <h3>Subclasses</h3>
+        <ul class="entity-list">
+            {% for uri in class_info.subclasses %}
+            <li>
+                {% set found = namespace(value=false) %}
+                {% for c in classes %}
+                    {% if c.uri|string == uri|string %}
+                    <a href="{{ c.qname | entity_url('class') }}">{{ c.label or c.qname }}</a>
+                    {% set found.value = true %}
+                    {% endif %}
+                {% endfor %}
+                {% if not found.value %}
+                <code>{{ uri }}</code>
+                {% endif %}
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if class_info.domain_of %}
+    <section class="section">
+        <h3>Properties (Domain)</h3>
+        <ul class="entity-list">
+            {% for prop in class_info.domain_of %}
+            <li>
+                <a href="{{ prop.qname | entity_url(prop.property_type + '_property') }}">
+                    {{ prop.label or prop.qname }}
+                </a>
+                <span class="entity-type {{ prop.property_type }}">{{ prop.property_type }}</span>
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if inherited_properties %}
+    <section class="section">
+        <h3>Inherited Properties</h3>
+        <ul class="entity-list">
+            {% for prop in inherited_properties %}
+            <li>
+                <a href="{{ prop.qname | entity_url(prop.property_type + '_property') }}">
+                    {{ prop.label or prop.qname }}
+                </a>
+                <span class="entity-type {{ prop.property_type }}">{{ prop.property_type }}</span>
+                <span class="annotation">(inherited)</span>
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if class_info.range_of %}
+    <section class="section">
+        <h3>Used as Range</h3>
+        <ul class="entity-list">
+            {% for prop in class_info.range_of %}
+            <li>
+                <a href="{{ prop.qname | entity_url(prop.property_type + '_property') }}">
+                    {{ prop.label or prop.qname }}
+                </a>
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if class_info.instances and config.include_instances %}
+    <section class="section">
+        <h3>Instances</h3>
+        <ul class="entity-list">
+            {% for uri in class_info.instances %}
+            <li>
+                {% set found = namespace(value=false) %}
+                {% for i in instances %}
+                    {% if i.uri|string == uri|string %}
+                    <a href="{{ i.qname | entity_url('instance') }}">{{ i.label or i.qname }}</a>
+                    {% set found.value = true %}
+                    {% endif %}
+                {% endfor %}
+                {% if not found.value %}
+                <code>{{ uri }}</code>
+                {% endif %}
+            </li>
+            {% endfor %}
+        </ul>
+    </section>
+    {% endif %}
+
+    {% if class_info.annotations %}
+    <section class="section">
+        <h3>Annotations</h3>
+        <table>
+            <thead>
+                <tr>
+                    <th>Property</th>
+                    <th>Value</th>
+                </tr>
+            </thead>
+            <tbody>
+                {% for name, values in class_info.annotations.items() %}
+                {% for value in values %}
+                <tr>
+                    <td><code>{{ name }}</code></td>
+                    <td>{{ value }}</td>
+                </tr>
+                {% endfor %}
+                {% endfor %}
+            </tbody>
+        </table>
+    </section>
+    {% endif %}
+</article>
+{% endblock %}

rdf_construct/docs/templates/html/hierarchy.html.jinja

@@ -0,0 +1,28 @@
+{% extends "base.html.jinja" %}
+
+{% block title %}Class Hierarchy - {{ ontology.title or "Documentation" }}{% endblock %}
+
+{% block content %}
+<h2>Class Hierarchy</h2>
+
+{% macro render_tree(nodes) %}
+<ul class="hierarchy-tree">
+    {% for node in nodes %}
+    <li>
+        <a href="{{ node.class.qname | entity_url('class') }}">
+            {{ node.class.label or node.class.qname }}
+        </a>
+        {% if node.children %}
+        {{ render_tree(node.children) }}
+        {% endif %}
+    </li>
+    {% endfor %}
+</ul>
+{% endmacro %}
+
+{% if hierarchy %}
+{{ render_tree(hierarchy) }}
+{% else %}
+<p>No class hierarchy found.</p>
+{% endif %}
+{% endblock %}

rdf_construct/docs/templates/html/index.html.jinja

@@ -0,0 +1,110 @@
+{% extends "base.html.jinja" %}
+
+{% block title %}{{ ontology.title or "Ontology Documentation" }}{% endblock %}
+
+{% block content %}
+<section class="stats">
+    <div class="stat-card">
+        <div class="number">{{ total_classes }}</div>
+        <div class="label">Classes</div>
+    </div>
+    <div class="stat-card">
+        <div class="number">{{ total_properties }}</div>
+        <div class="label">Properties</div>
+    </div>
+    {% if instances %}
+    <div class="stat-card">
+        <div class="number">{{ total_instances }}</div>
+        <div class="label">Instances</div>
+    </div>
+    {% endif %}
+</section>
+
+{% if classes %}
+<section class="section">
+    <h2>Classes</h2>
+    <ul class="entity-list">
+        {% for class in classes %}
+        <li>
+            <a href="{{ class.qname | entity_url('class') }}">
+                {{ class.label or class.qname }}
+            </a>
+            {% if class.definition %}
+            <span class="definition">— {{ class.definition[:100] }}{% if class.definition|length > 100 %}...{% endif %}</span>
+            {% endif %}
+        </li>
+        {% endfor %}
+    </ul>
+</section>
+{% endif %}
+
+{% if object_properties %}
+<section class="section">
+    <h2>Object Properties</h2>
+    <ul class="entity-list">
+        {% for prop in object_properties %}
+        <li>
+            <a href="{{ prop.qname | entity_url('object_property') }}">
+                {{ prop.label or prop.qname }}
+            </a>
+            <span class="entity-type object">Object</span>
+            {% if prop.definition %}
+            <span class="definition">— {{ prop.definition[:80] }}{% if prop.definition|length > 80 %}...{% endif %}</span>
+            {% endif %}
+        </li>
+        {% endfor %}
+    </ul>
+</section>
+{% endif %}
+
+{% if datatype_properties %}
+<section class="section">
+    <h2>Datatype Properties</h2>
+    <ul class="entity-list">
+        {% for prop in datatype_properties %}
+        <li>
+            <a href="{{ prop.qname | entity_url('datatype_property') }}">
+                {{ prop.label or prop.qname }}
+            </a>
+            <span class="entity-type datatype">Datatype</span>
+            {% if prop.definition %}
+            <span class="definition">— {{ prop.definition[:80] }}{% if prop.definition|length > 80 %}...{% endif %}</span>
+            {% endif %}
+        </li>
+        {% endfor %}
+    </ul>
+</section>
+{% endif %}
+
+{% if annotation_properties %}
+<section class="section">
+    <h2>Annotation Properties</h2>
+    <ul class="entity-list">
+        {% for prop in annotation_properties %}
+        <li>
+            <a href="{{ prop.qname | entity_url('annotation_property') }}">
+                {{ prop.label or prop.qname }}
+            </a>
+            <span class="entity-type annotation">Annotation</span>
+        </li>
+        {% endfor %}
+    </ul>
+</section>
+{% endif %}
+
+{% if instances and config.include_instances %}
+<section class="section">
+    <h2>Instances</h2>
+    <ul class="entity-list">
+        {% for inst in instances %}
+        <li>
+            <a href="{{ inst.qname | entity_url('instance') }}">
+                {{ inst.label or inst.qname }}
+            </a>
+            <span class="entity-type instance">Instance</span>
+        </li>
+        {% endfor %}
+    </ul>
+</section>
+{% endif %}
+{% endblock %}