Add new slots for heritage custodian ontology

- Introduced `has_api_version`, `has_appellation_language`, `has_appellation_type`, `has_appellation_value`, `has_applicable_country`, `has_application_deadline`, `has_application_opening_date`, `has_appraisal_note`, `has_approval_date`, `has_archdiocese_name`, `has_architectural_style`, `has_archival_reference`, `has_archive_description`, `has_archive_memento_uri`, `has_archive_name`, `has_archive_path`, `has_archive_search_score`, `has_arrangement`, `has_arrangement_level`, `has_arrangement_note`, `has_articles_archival_stage`, `has_articles_document_format`, `has_articles_document_url`, `has_articles_of_association`, `has_or_had_altitude`, `has_or_had_annotation`, `has_or_had_arrangement`, `has_or_had_document`, `has_or_had_reason`, `has_or_had_style`, `is_or_was_amended_through`, `is_or_was_approved_on`, `is_or_was_archived_as`, `is_or_was_due_on`, `is_or_was_opened_on`, and `is_or_was_used_in` slots.
- Each slot includes detailed descriptions, range specifications, and appropriate mappings to existing ontologies.

.opencode/rules/no-deletion-from-slot-fixes.md

@@ -0,0 +1,17 @@
+# Rule: Do Not Delete Entries from slot_fixes.yaml
+
+**CRITICAL**: Entries in `schemas/20251121/linkml/modules/slots/slot_fixes.yaml` MUST NEVER be deleted.
+
+This file serves as a persistent audit log and migration tracking registry. 
+
+**Protocol**:
+1.  **Process** the migration specified in the `revision` section.
+2.  **Update** the `processed` section:
+    *   Set `status: true`.
+    *   Add a `notes` field describing the action taken (e.g., "Migrated to has_or_had_name + PersonName class. Slot archived.").
+    *   Add a `date` field (YYYY-MM-DD).
+3.  **Keep** the original entry intact.
+
+**Forbidden**:
+*   ❌ Deleting a processed block.
+*   ❌ Removing an entry because the slot file doesn't exist (mark as processed with note "Slot file not found, skipped").

.opencode/rules/preserve-bespoke-slots-until-refactoring.md

@@ -0,0 +1,32 @@
+# Rule: Preserve Bespoke Slots Until Refactoring
+
+**Identifier**: `preserve-bespoke-slots-until-refactoring`
+**Severity**: **CRITICAL**
+
+## Core Directive
+
+**DO NOT remove or migrate "additional" bespoke slots during generic migration passes unless they are the specific target of the current task.**
+
+## Context
+
+When migrating a specific slot (e.g., `has_approval_date`), you may encounter other bespoke or legacy slots in the same class file (e.g., `innovation_budget`, `operating_budget`).
+
+**YOU MUST**:
+*   ✅ Migrate ONLY the specific slot you were instructed to work on.
+*   ✅ Leave other bespoke slots exactly as they are.
+*   ✅ Focus strictly on the current migration target.
+
+**YOU MUST NOT**:
+*   ❌ Proactively migrate "nearby" slots just because they look like they need refactoring.
+*   ❌ Remove slots that seem unused or redundant without specific instruction.
+*   ❌ "Clean up" the class file by removing legacy attributes.
+
+## Rationale
+
+Refactoring is a separate, planned phase. Mixing opportunistic refactoring with systematic slot migration increases the risk of regression and makes changes harder to review. "We will refactor those later."
+
+## Workflow
+
+1.  **Identify Target**: Identify the specific slot(s) assigned for migration (from `slot_fixes.yaml` or user prompt).
+2.  **Execute Migration**: Apply changes ONLY for those slots.
+3.  **Ignore Others**: Do not touch other slots in the file, even if they violate other rules (like Rule 39 or Rule 53). Those will be handled in their own dedicated tasks.

backend/rag/main.py

@@ -1660,6 +1660,7 @@ class MultiSourceRetriever:
         only_heritage_relevant: bool = False,
         only_wcms: bool = False,
         using: str | None = None,
+        extra_filters: dict[str, Any] | None = None,
     ) -> list[Any]:
         """Search for persons/staff in the heritage_persons collection.
         
@@ -1672,20 +1673,29 @@ class MultiSourceRetriever:
             only_heritage_relevant: Only return heritage-relevant staff
             only_wcms: Only return WCMS-registered profiles
             using: Optional embedding model to use (e.g., 'minilm_384', 'openai_1536')
+            extra_filters: Optional extra filters for Qdrant
             
         Returns:
             List of RetrievedPerson objects
         """
         if self.qdrant:
             try:
-                return self.qdrant.search_persons(  # type: ignore[no-any-return]
-                    query=query,
-                    k=k,
-                    filter_custodian=filter_custodian,
-                    only_heritage_relevant=only_heritage_relevant,
-                    only_wcms=only_wcms,
-                    using=using,
-                )
+                # Dynamically check if qdrant.search_persons supports extra_filters
+                # This handles case where HybridRetriever signature varies
+                import inspect
+                sig = inspect.signature(self.qdrant.search_persons)
+                kwargs = {
+                    "query": query,
+                    "k": k,
+                    "filter_custodian": filter_custodian,
+                    "only_heritage_relevant": only_heritage_relevant,
+                    "only_wcms": only_wcms,
+                    "using": using,
+                }
+                if "extra_filters" in sig.parameters:
+                    kwargs["extra_filters"] = extra_filters
+                
+                return self.qdrant.search_persons(**kwargs)  # type: ignore[no-any-return]
             except Exception as e:
                 logger.error(f"Person search failed: {e}")
         return []
@@ -2755,11 +2765,18 @@ async def person_search(request: PersonSearchRequest) -> PersonSearchResponse:
         # Augment query for better recall on domain names if it looks like a domain search
         # "nos" -> "nos email domain nos" to guide vector search towards email addresses
         search_query = request.query
+        extra_filters = None
+        
+        # Check for single word domain-like queries
         if len(search_query.split()) == 1 and len(search_query) > 2 and "@" not in search_query:
              # Heuristic: single word queries might be domain searches
-             # We append "email domain" context to guide the embedding
-             search_query = f"{search_query} email domain {search_query}"
+             # We use MatchText filtering on email field to find substring matches
+             # Qdrant "match": {"text": "nos"} performs token-based matching
+             extra_filters = {"email": {"match": {"text": search_query}}}
+             logger.info(f"[PersonSearch] Potential domain search detected for '{search_query}'. Applying strict email filter: {extra_filters}")
              
+        logger.info(f"[PersonSearch] Executing search for '{search_query}' (extra_filters={extra_filters})")
+        
         # Use the hybrid retriever's person search
         results = retriever.search_persons(
             query=search_query,
@@ -2768,8 +2785,27 @@ async def person_search(request: PersonSearchRequest) -> PersonSearchResponse:
             only_heritage_relevant=request.only_heritage_relevant,
             only_wcms=request.only_wcms,
             using=request.embedding_model,  # Pass embedding model
+            extra_filters=extra_filters,
         )
         
+        # FALLBACK: If strict domain filter yielded no results, try standard vector search
+        # This fixes the issue where searching for names like "willem" (which look like domains)
+        # would fail because they don't appear in emails.
+        if extra_filters and not results:
+            logger.info(f"[PersonSearch] No results with email filter for '{search_query}'. Falling back to standard vector search.")
+            results = retriever.search_persons(
+                query=search_query,
+                k=request.k,
+                filter_custodian=request.filter_custodian,
+                only_heritage_relevant=request.only_heritage_relevant,
+                only_wcms=request.only_wcms,
+                using=request.embedding_model,
+                extra_filters=None,  # Disable filter for fallback
+            )
+            logger.info(f"[PersonSearch] Fallback search returned {len(results)} results")
+            
+        logger.info(f"[PersonSearch] Final result count: {len(results)}")
+        
         # Determine which embedding model was actually used
         embedding_model_used = None
         qdrant = retriever.qdrant

backend/rag/multi_embedding_retriever.py

@@ -0,0 +1,846 @@
+"""
+Multi-Embedding Retriever for Heritage Data
+
+Supports multiple embedding models using Qdrant's named vectors feature.
+This enables:
+- A/B testing different embedding models
+- Cost optimization (cheap local embeddings vs paid API embeddings)
+- Gradual migration between embedding models
+- Fallback when one model is unavailable
+
+Supported Embedding Models:
+    - openai_1536: text-embedding-3-small (1536-dim, $0.02/1M tokens)
+    - minilm_384: all-MiniLM-L6-v2 (384-dim, free/local)
+    - bge_768: bge-base-en-v1.5 (768-dim, free/local, high quality)
+
+Collection Architecture:
+    Each collection has named vectors for each embedding model:
+    
+    heritage_custodians:
+        vectors:
+            "openai_1536": VectorParams(size=1536)
+            "minilm_384": VectorParams(size=384)
+        payload: {name, ghcid, institution_type, ...}
+    
+    heritage_persons:
+        vectors:
+            "openai_1536": VectorParams(size=1536)
+            "minilm_384": VectorParams(size=384)
+        payload: {name, headline, custodian_name, ...}
+
+Usage:
+    retriever = MultiEmbeddingRetriever()
+    
+    # Search with default model (auto-select based on availability)
+    results = retriever.search("museums in Amsterdam")
+    
+    # Search with specific model
+    results = retriever.search("museums in Amsterdam", using="minilm_384")
+    
+    # A/B test comparison
+    comparison = retriever.compare_models("museums in Amsterdam")
+"""
+
+import hashlib
+import logging
+import os
+from dataclasses import dataclass, field
+from enum import Enum
+from typing import Any, Literal
+
+logger = logging.getLogger(__name__)
+
+
+class EmbeddingModel(str, Enum):
+    """Supported embedding models with their configurations."""
+    
+    OPENAI_1536 = "openai_1536"
+    MINILM_384 = "minilm_384"
+    BGE_768 = "bge_768"
+    
+    @property
+    def dimension(self) -> int:
+        """Get the vector dimension for this model."""
+        dims = {
+            "openai_1536": 1536,
+            "minilm_384": 384,
+            "bge_768": 768,
+        }
+        return dims[self.value]
+    
+    @property
+    def model_name(self) -> str:
+        """Get the actual model name for loading."""
+        names = {
+            "openai_1536": "text-embedding-3-small",
+            "minilm_384": "all-MiniLM-L6-v2",
+            "bge_768": "BAAI/bge-base-en-v1.5",
+        }
+        return names[self.value]
+    
+    @property
+    def is_local(self) -> bool:
+        """Check if this model runs locally (no API calls)."""
+        return self.value in ("minilm_384", "bge_768")
+    
+    @property
+    def cost_per_1m_tokens(self) -> float:
+        """Approximate cost per 1M tokens (0 for local models)."""
+        costs = {
+            "openai_1536": 0.02,
+            "minilm_384": 0.0,
+            "bge_768": 0.0,
+        }
+        return costs[self.value]
+
+
+@dataclass
+class MultiEmbeddingConfig:
+    """Configuration for multi-embedding retriever."""
+    
+    # Qdrant connection
+    qdrant_host: str = "localhost"
+    qdrant_port: int = 6333
+    qdrant_https: bool = False
+    qdrant_prefix: str | None = None
+    
+    # API keys
+    openai_api_key: str | None = None
+    
+    # Default embedding model preference order
+    # First available model is used if no explicit model is specified
+    model_preference: list[EmbeddingModel] = field(default_factory=lambda: [
+        EmbeddingModel.MINILM_384,  # Free, fast, good quality
+        EmbeddingModel.OPENAI_1536,  # Higher quality, paid
+        EmbeddingModel.BGE_768,  # Free, high quality, slower
+    ])
+    
+    # Collection names
+    institutions_collection: str = "heritage_custodians"
+    persons_collection: str = "heritage_persons"
+    
+    # Search defaults
+    default_k: int = 10
+
+
+class MultiEmbeddingRetriever:
+    """Retriever supporting multiple embedding models via Qdrant named vectors.
+    
+    This class manages multiple embedding models and allows searching with
+    any available model. It handles:
+    - Model lazy-loading
+    - Automatic model selection based on availability
+    - Named vector creation and search
+    - A/B testing between models
+    """
+    
+    def __init__(self, config: MultiEmbeddingConfig | None = None):
+        """Initialize multi-embedding retriever.
+        
+        Args:
+            config: Configuration options. If None, uses environment variables.
+        """
+        self.config = config or self._config_from_env()
+        
+        # Lazy-loaded clients
+        self._qdrant_client = None
+        self._openai_client = None
+        self._st_models: dict[str, Any] = {}  # Sentence transformer models
+        
+        # Track available models per collection
+        self._available_models: dict[str, set[EmbeddingModel]] = {}
+        
+        # Track whether each collection uses named vectors (vs single unnamed vector)
+        self._uses_named_vectors: dict[str, bool] = {}
+        
+        logger.info(f"MultiEmbeddingRetriever initialized with preference: {[m.value for m in self.config.model_preference]}")
+    
+    @staticmethod
+    def _config_from_env() -> MultiEmbeddingConfig:
+        """Create configuration from environment variables."""
+        use_production = os.getenv("QDRANT_USE_PRODUCTION", "false").lower() == "true"
+        
+        if use_production:
+            return MultiEmbeddingConfig(
+                qdrant_host=os.getenv("QDRANT_PROD_HOST", "bronhouder.nl"),
+                qdrant_port=443,
+                qdrant_https=True,
+                qdrant_prefix=os.getenv("QDRANT_PROD_PREFIX", "qdrant"),
+                openai_api_key=os.getenv("OPENAI_API_KEY"),
+            )
+        else:
+            return MultiEmbeddingConfig(
+                qdrant_host=os.getenv("QDRANT_HOST", "localhost"),
+                qdrant_port=int(os.getenv("QDRANT_PORT", "6333")),
+                openai_api_key=os.getenv("OPENAI_API_KEY"),
+            )
+    
+    @property
+    def qdrant_client(self):
+        """Lazy-load Qdrant client."""
+        if self._qdrant_client is None:
+            from qdrant_client import QdrantClient
+            
+            if self.config.qdrant_https:
+                self._qdrant_client = QdrantClient(
+                    host=self.config.qdrant_host,
+                    port=self.config.qdrant_port,
+                    https=True,
+                    prefix=self.config.qdrant_prefix,
+                    prefer_grpc=False,
+                    timeout=30,
+                )
+                logger.info(f"Connected to Qdrant: https://{self.config.qdrant_host}/{self.config.qdrant_prefix or ''}")
+            else:
+                self._qdrant_client = QdrantClient(
+                    host=self.config.qdrant_host,
+                    port=self.config.qdrant_port,
+                )
+                logger.info(f"Connected to Qdrant: {self.config.qdrant_host}:{self.config.qdrant_port}")
+        
+        return self._qdrant_client
+    
+    @property
+    def openai_client(self):
+        """Lazy-load OpenAI client."""
+        if self._openai_client is None:
+            if not self.config.openai_api_key:
+                raise RuntimeError("OpenAI API key not configured")
+            
+            import openai
+            self._openai_client = openai.OpenAI(api_key=self.config.openai_api_key)
+        
+        return self._openai_client
+    
+    def _load_sentence_transformer(self, model: EmbeddingModel) -> Any:
+        """Lazy-load a sentence-transformers model.
+        
+        Args:
+            model: The embedding model to load
+            
+        Returns:
+            Loaded SentenceTransformer model
+        """
+        if model.value not in self._st_models:
+            try:
+                from sentence_transformers import SentenceTransformer
+                self._st_models[model.value] = SentenceTransformer(model.model_name)
+                logger.info(f"Loaded sentence-transformers model: {model.model_name}")
+            except ImportError:
+                raise RuntimeError(
+                    "sentence-transformers not installed. Run: pip install sentence-transformers"
+                )
+        
+        return self._st_models[model.value]
+    
+    def get_embedding(self, text: str, model: EmbeddingModel) -> list[float]:
+        """Get embedding vector for text using specified model.
+        
+        Args:
+            text: Text to embed
+            model: Embedding model to use
+            
+        Returns:
+            Embedding vector as list of floats
+        """
+        if model == EmbeddingModel.OPENAI_1536:
+            response = self.openai_client.embeddings.create(
+                input=text,
+                model=model.model_name,
+            )
+            return response.data[0].embedding
+        
+        elif model in (EmbeddingModel.MINILM_384, EmbeddingModel.BGE_768):
+            st_model = self._load_sentence_transformer(model)
+            embedding = st_model.encode(text)
+            return embedding.tolist()
+        
+        else:
+            raise ValueError(f"Unknown embedding model: {model}")
+    
+    def get_embeddings_batch(
+        self,
+        texts: list[str],
+        model: EmbeddingModel,
+        batch_size: int = 32,
+    ) -> list[list[float]]:
+        """Get embedding vectors for multiple texts.
+        
+        Args:
+            texts: List of texts to embed
+            model: Embedding model to use
+            batch_size: Batch size for processing
+            
+        Returns:
+            List of embedding vectors
+        """
+        if not texts:
+            return []
+        
+        if model == EmbeddingModel.OPENAI_1536:
+            # OpenAI batch API (max 2048 per request)
+            all_embeddings = []
+            for i in range(0, len(texts), 2048):
+                batch = texts[i:i + 2048]
+                response = self.openai_client.embeddings.create(
+                    input=batch,
+                    model=model.model_name,
+                )
+                batch_embeddings = [item.embedding for item in sorted(response.data, key=lambda x: x.index)]
+                all_embeddings.extend(batch_embeddings)
+            return all_embeddings
+        
+        elif model in (EmbeddingModel.MINILM_384, EmbeddingModel.BGE_768):
+            st_model = self._load_sentence_transformer(model)
+            embeddings = st_model.encode(texts, batch_size=batch_size, show_progress_bar=len(texts) > 100)
+            return embeddings.tolist()
+        
+        else:
+            raise ValueError(f"Unknown embedding model: {model}")
+    
+    def get_available_models(self, collection_name: str) -> set[EmbeddingModel]:
+        """Get the embedding models available for a collection.
+        
+        Checks which named vectors exist in the collection.
+        For single-vector collections, returns models matching the dimension.
+        
+        Args:
+            collection_name: Name of the Qdrant collection
+            
+        Returns:
+            Set of available EmbeddingModel values
+        """
+        if collection_name in self._available_models:
+            return self._available_models[collection_name]
+        
+        try:
+            info = self.qdrant_client.get_collection(collection_name)
+            vectors_config = info.config.params.vectors
+            
+            available = set()
+            uses_named_vectors = False
+            
+            # Check for named vectors (dict of vector configs)
+            if isinstance(vectors_config, dict):
+                # Named vectors - each key is a vector name
+                uses_named_vectors = True
+                for vector_name in vectors_config.keys():
+                    try:
+                        model = EmbeddingModel(vector_name)
+                        available.add(model)
+                    except ValueError:
+                        logger.warning(f"Unknown vector name in collection: {vector_name}")
+            else:
+                # Single unnamed vector - check dimension to find compatible model
+                # Note: This doesn't mean we can use `using=model.value` in queries
+                uses_named_vectors = False
+                if hasattr(vectors_config, 'size'):
+                    dim = vectors_config.size
+                    for model in EmbeddingModel:
+                        if model.dimension == dim:
+                            available.add(model)
+            
+            # Store both available models and whether named vectors are used
+            self._available_models[collection_name] = available
+            self._uses_named_vectors[collection_name] = uses_named_vectors
+            
+            if uses_named_vectors:
+                logger.info(f"Collection '{collection_name}' uses named vectors: {[m.value for m in available]}")
+            else:
+                logger.info(f"Collection '{collection_name}' uses single vector (compatible with: {[m.value for m in available]})")
+            
+            return available
+            
+        except Exception as e:
+            logger.warning(f"Could not get available models for {collection_name}: {e}")
+            return set()
+    
+    def uses_named_vectors(self, collection_name: str) -> bool:
+        """Check if a collection uses named vectors (vs single unnamed vector).
+        
+        Args:
+            collection_name: Name of the Qdrant collection
+            
+        Returns:
+            True if collection has named vectors, False for single-vector collections
+        """
+        # Ensure models are loaded (populates _uses_named_vectors)
+        self.get_available_models(collection_name)
+        return self._uses_named_vectors.get(collection_name, False)
+    
+    def select_model(
+        self,
+        collection_name: str,
+        preferred: EmbeddingModel | None = None,
+    ) -> EmbeddingModel | None:
+        """Select the best available embedding model for a collection.
+        
+        Args:
+            collection_name: Name of the collection
+            preferred: Preferred model (used if available)
+            
+        Returns:
+            Selected EmbeddingModel or None if none available
+        """
+        available = self.get_available_models(collection_name)
+        
+        if not available:
+            # No named vectors - check if we can use any model
+            # This happens for legacy single-vector collections
+            try:
+                info = self.qdrant_client.get_collection(collection_name)
+                vectors_config = info.config.params.vectors
+                
+                # Get vector dimension
+                dim = None
+                if hasattr(vectors_config, 'size'):
+                    dim = vectors_config.size
+                elif isinstance(vectors_config, dict):
+                    # Get first vector config
+                    first_config = next(iter(vectors_config.values()), None)
+                    if first_config and hasattr(first_config, 'size'):
+                        dim = first_config.size
+                
+                if dim:
+                    for model in self.config.model_preference:
+                        if model.dimension == dim:
+                            return model
+            except Exception:
+                pass
+            
+            return None
+        
+        # If preferred model is available, use it
+        if preferred and preferred in available:
+            return preferred
+        
+        # Otherwise, follow preference order
+        for model in self.config.model_preference:
+            if model in available:
+                # Check if model is usable (has API key if needed)
+                if model == EmbeddingModel.OPENAI_1536 and not self.config.openai_api_key:
+                    continue
+                return model
+        
+        return None
+    
+    def search(
+        self,
+        query: str,
+        collection_name: str | None = None,
+        k: int | None = None,
+        using: EmbeddingModel | str | None = None,
+        filter_conditions: dict[str, Any] | None = None,
+    ) -> list[dict[str, Any]]:
+        """Search for similar documents using specified or auto-selected model.
+        
+        Args:
+            query: Search query text
+            collection_name: Collection to search (default: institutions)
+            k: Number of results
+            using: Embedding model to use (auto-selected if None)
+            filter_conditions: Optional Qdrant filter conditions
+            
+        Returns:
+            List of results with scores and payloads
+        """
+        collection_name = collection_name or self.config.institutions_collection
+        k = k or self.config.default_k
+        
+        # Resolve model
+        if using is not None:
+            if isinstance(using, str):
+                model = EmbeddingModel(using)
+            else:
+                model = using
+        else:
+            model = self.select_model(collection_name)
+        
+        if model is None:
+            raise RuntimeError(f"No compatible embedding model for collection '{collection_name}'")
+        
+        logger.info(f"Searching '{collection_name}' with {model.value}: {query[:50]}...")
+        
+        # Get query embedding
+        query_vector = self.get_embedding(query, model)
+        
+        # Build filter
+        from qdrant_client.http import models
+        
+        query_filter = None
+        if filter_conditions:
+            query_filter = models.Filter(
+                must=[
+                    models.FieldCondition(
+                        key=key,
+                        match=models.MatchValue(value=value),
+                    )
+                    for key, value in filter_conditions.items()
+                ]
+            )
+        
+        # Check if collection uses named vectors (not just single unnamed vector)
+        # Only pass `using=model.value` if collection has actual named vectors
+        use_named_vector = self.uses_named_vectors(collection_name)
+        
+        # Search
+        if use_named_vector:
+            results = self.qdrant_client.query_points(
+                collection_name=collection_name,
+                query=query_vector,
+                using=model.value,
+                limit=k,
+                with_payload=True,
+                query_filter=query_filter,
+            )
+        else:
+            # Legacy single-vector search
+            results = self.qdrant_client.query_points(
+                collection_name=collection_name,
+                query=query_vector,
+                limit=k,
+                with_payload=True,
+                query_filter=query_filter,
+            )
+        
+        return [
+            {
+                "id": str(point.id),
+                "score": point.score,
+                "model": model.value,
+                "payload": point.payload or {},
+            }
+            for point in results.points
+        ]
+    
+    def search_persons(
+        self,
+        query: str,
+        k: int | None = None,
+        using: EmbeddingModel | str | None = None,
+        filter_custodian: str | None = None,
+        only_heritage_relevant: bool = False,
+        only_wcms: bool = False,
+    ) -> list[dict[str, Any]]:
+        """Search for persons/staff in the heritage_persons collection.
+        
+        Args:
+            query: Search query text
+            k: Number of results
+            using: Embedding model to use
+            filter_custodian: Optional custodian slug to filter by
+            only_heritage_relevant: Only return heritage-relevant staff
+            only_wcms: Only return WCMS-registered profiles (heritage sector users)
+            
+        Returns:
+            List of person results with scores
+        """
+        k = k or self.config.default_k
+        
+        # Build filters
+        filters = {}
+        if filter_custodian:
+            filters["custodian_slug"] = filter_custodian
+        if only_wcms:
+            filters["has_wcms"] = True
+        
+        # Search with over-fetch for post-filtering
+        results = self.search(
+            query=query,
+            collection_name=self.config.persons_collection,
+            k=k * 2,
+            using=using,
+            filter_conditions=filters if filters else None,
+        )
+        
+        # Post-filter for heritage_relevant if needed
+        if only_heritage_relevant:
+            results = [r for r in results if r.get("payload", {}).get("heritage_relevant", False)]
+        
+        # Format results
+        formatted = []
+        for r in results[:k]:
+            payload = r.get("payload", {})
+            formatted.append({
+                "person_id": payload.get("staff_id", "") or hashlib.md5(
+                    f"{payload.get('custodian_slug', '')}:{payload.get('name', '')}".encode()
+                ).hexdigest()[:16],
+                "name": payload.get("name", ""),
+                "headline": payload.get("headline"),
+                "custodian_name": payload.get("custodian_name"),
+                "custodian_slug": payload.get("custodian_slug"),
+                "location": payload.get("location"),
+                "heritage_relevant": payload.get("heritage_relevant", False),
+                "heritage_type": payload.get("heritage_type"),
+                "linkedin_url": payload.get("linkedin_url"),
+                "score": r["score"],
+                "model": r["model"],
+            })
+        
+        return formatted
+    
+    def compare_models(
+        self,
+        query: str,
+        collection_name: str | None = None,
+        k: int = 10,
+        models: list[EmbeddingModel] | None = None,
+    ) -> dict[str, Any]:
+        """A/B test comparison of multiple embedding models.
+        
+        Args:
+            query: Search query
+            collection_name: Collection to search
+            k: Number of results per model
+            models: Models to compare (default: all available)
+            
+        Returns:
+            Dict with results per model and overlap analysis
+        """
+        collection_name = collection_name or self.config.institutions_collection
+        
+        # Determine which models to compare
+        available = self.get_available_models(collection_name)
+        if models:
+            models_to_test = [m for m in models if m in available]
+        else:
+            models_to_test = list(available)
+        
+        if not models_to_test:
+            return {"error": "No models available for comparison"}
+        
+        results = {}
+        all_ids = {}
+        
+        for model in models_to_test:
+            try:
+                model_results = self.search(
+                    query=query,
+                    collection_name=collection_name,
+                    k=k,
+                    using=model,
+                )
+                results[model.value] = model_results
+                all_ids[model.value] = {r["id"] for r in model_results}
+            except Exception as e:
+                results[model.value] = {"error": str(e)}
+                all_ids[model.value] = set()
+        
+        # Calculate overlap between models
+        overlap = {}
+        model_values = list(all_ids.keys())
+        for i, m1 in enumerate(model_values):
+            for m2 in model_values[i + 1:]:
+                if all_ids[m1] and all_ids[m2]:
+                    intersection = all_ids[m1] & all_ids[m2]
+                    union = all_ids[m1] | all_ids[m2]
+                    jaccard = len(intersection) / len(union) if union else 0
+                    overlap[f"{m1}_vs_{m2}"] = {
+                        "jaccard_similarity": round(jaccard, 3),
+                        "common_results": len(intersection),
+                        "total_unique": len(union),
+                    }
+        
+        return {
+            "query": query,
+            "collection": collection_name,
+            "k": k,
+            "results": results,
+            "overlap_analysis": overlap,
+        }
+    
+    def create_multi_embedding_collection(
+        self,
+        collection_name: str,
+        models: list[EmbeddingModel] | None = None,
+    ) -> bool:
+        """Create a new collection with named vectors for multiple embedding models.
+        
+        Args:
+            collection_name: Name for the new collection
+            models: Embedding models to support (default: all)
+            
+        Returns:
+            True if created successfully
+        """
+        from qdrant_client.http.models import Distance, VectorParams
+        
+        models = models or list(EmbeddingModel)
+        
+        vectors_config = {
+            model.value: VectorParams(
+                size=model.dimension,
+                distance=Distance.COSINE,
+            )
+            for model in models
+        }
+        
+        try:
+            self.qdrant_client.create_collection(
+                collection_name=collection_name,
+                vectors_config=vectors_config,
+            )
+            logger.info(f"Created multi-embedding collection '{collection_name}' with {[m.value for m in models]}")
+            
+            # Clear cache
+            self._available_models.pop(collection_name, None)
+            
+            return True
+            
+        except Exception as e:
+            logger.error(f"Failed to create collection: {e}")
+            return False
+    
+    def add_documents_multi_embedding(
+        self,
+        documents: list[dict[str, Any]],
+        collection_name: str,
+        models: list[EmbeddingModel] | None = None,
+        batch_size: int = 100,
+    ) -> int:
+        """Add documents with embeddings from multiple models.
+        
+        Args:
+            documents: List of documents with 'text' and optional 'metadata' fields
+            collection_name: Target collection
+            models: Models to generate embeddings for (default: all available)
+            batch_size: Batch size for processing
+            
+        Returns:
+            Number of documents added
+        """
+        from qdrant_client.http import models as qmodels
+        
+        # Determine which models to use
+        available = self.get_available_models(collection_name)
+        if models:
+            models_to_use = [m for m in models if m in available]
+        else:
+            models_to_use = list(available)
+        
+        if not models_to_use:
+            raise RuntimeError(f"No embedding models available for collection '{collection_name}'")
+        
+        # Filter valid documents
+        valid_docs = [d for d in documents if d.get("text")]
+        total_indexed = 0
+        
+        for i in range(0, len(valid_docs), batch_size):
+            batch = valid_docs[i:i + batch_size]
+            texts = [d["text"] for d in batch]
+            
+            # Generate embeddings for each model
+            embeddings_by_model = {}
+            for model in models_to_use:
+                try:
+                    embeddings_by_model[model] = self.get_embeddings_batch(texts, model)
+                except Exception as e:
+                    logger.warning(f"Failed to get {model.value} embeddings: {e}")
+            
+            if not embeddings_by_model:
+                continue
+            
+            # Create points with named vectors
+            points = []
+            for j, doc in enumerate(batch):
+                text = doc["text"]
+                metadata = doc.get("metadata", {})
+                point_id = doc.get("id") or hashlib.md5(text.encode()).hexdigest()
+                
+                # Build named vectors dict
+                vectors = {}
+                for model, model_embeddings in embeddings_by_model.items():
+                    vectors[model.value] = model_embeddings[j]
+                
+                points.append(qmodels.PointStruct(
+                    id=point_id,
+                    vector=vectors,
+                    payload={
+                        "text": text,
+                        **metadata,
+                    }
+                ))
+            
+            # Upsert batch
+            self.qdrant_client.upsert(
+                collection_name=collection_name,
+                points=points,
+            )
+            total_indexed += len(points)
+            logger.info(f"Indexed {total_indexed}/{len(valid_docs)} documents with {len(models_to_use)} models")
+        
+        return total_indexed
+    
+    def get_stats(self) -> dict[str, Any]:
+        """Get statistics about collections and available models.
+        
+        Returns:
+            Dict with collection stats and model availability
+        """
+        stats = {
+            "config": {
+                "qdrant_host": self.config.qdrant_host,
+                "qdrant_port": self.config.qdrant_port,
+                "model_preference": [m.value for m in self.config.model_preference],
+                "openai_available": bool(self.config.openai_api_key),
+            },
+            "collections": {},
+        }
+        
+        for collection_name in [self.config.institutions_collection, self.config.persons_collection]:
+            try:
+                info = self.qdrant_client.get_collection(collection_name)
+                available_models = self.get_available_models(collection_name)
+                selected_model = self.select_model(collection_name)
+                
+                stats["collections"][collection_name] = {
+                    "vectors_count": info.vectors_count,
+                    "points_count": info.points_count,
+                    "status": info.status.value if info.status else "unknown",
+                    "available_models": [m.value for m in available_models],
+                    "selected_model": selected_model.value if selected_model else None,
+                }
+            except Exception as e:
+                stats["collections"][collection_name] = {"error": str(e)}
+        
+        return stats
+    
+    def close(self):
+        """Close all connections."""
+        if self._qdrant_client:
+            self._qdrant_client.close()
+            self._qdrant_client = None
+        self._st_models.clear()
+        self._available_models.clear()
+        self._uses_named_vectors.clear()
+
+
+def create_multi_embedding_retriever(use_production: bool | None = None) -> MultiEmbeddingRetriever:
+    """Factory function to create a MultiEmbeddingRetriever.
+    
+    Args:
+        use_production: If True, connect to production Qdrant.
+                       Defaults to QDRANT_USE_PRODUCTION env var.
+    
+    Returns:
+        Configured MultiEmbeddingRetriever instance
+    """
+    if use_production is None:
+        use_production = os.getenv("QDRANT_USE_PRODUCTION", "").lower() in ("true", "1", "yes")
+    
+    if use_production:
+        config = MultiEmbeddingConfig(
+            qdrant_host=os.getenv("QDRANT_PROD_HOST", "bronhouder.nl"),
+            qdrant_port=443,
+            qdrant_https=True,
+            qdrant_prefix=os.getenv("QDRANT_PROD_PREFIX", "qdrant"),
+            openai_api_key=os.getenv("OPENAI_API_KEY"),
+        )
+    else:
+        config = MultiEmbeddingConfig(
+            qdrant_host=os.getenv("QDRANT_HOST", "localhost"),
+            qdrant_port=int(os.getenv("QDRANT_PORT", "6333")),
+            openai_api_key=os.getenv("OPENAI_API_KEY"),
+        )
+    
+    return MultiEmbeddingRetriever(config)

frontend/public/schemas/20251121/linkml/manifest.json

@@ -1,5 +1,5 @@
 {
-  "generated": "2026-01-27T08:03:23.376Z",
+  "generated": "2026-01-27T08:04:51.838Z",
   "schemaRoot": "/schemas/20251121/linkml",
   "totalFiles": 3014,
   "categoryCounts": {

schemas/20251121/linkml/manifest.json

@@ -1,5 +1,5 @@
 {
-  "generated": "2026-01-27T08:04:51.838Z",
+  "generated": "2026-01-27T09:07:17.016Z",
   "schemaRoot": "/schemas/20251121/linkml",
   "totalFiles": 3014,
   "categoryCounts": {

schemas/20251121/linkml/modules/classes/APIEndpoint.yaml

@@ -0,0 +1,7 @@
+classes:
+  APIEndpoint:
+    class_uri: schema:EntryPoint
+    description: "An API endpoint."
+    slots:
+      - has_or_had_url
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/APIRequest.yaml

@@ -0,0 +1,8 @@
+classes:
+  APIRequest:
+    class_uri: prov:Activity
+    description: "An API request event."
+    slots:
+      - has_or_had_provenance
+      - has_or_had_endpoint
+      - has_or_had_version

schemas/20251121/linkml/modules/classes/APIVersion.yaml

@@ -0,0 +1,7 @@
+classes:
+  APIVersion:
+    class_uri: schema:SoftwareApplication
+    description: "Version of an API."
+    slots:
+      - has_or_had_label
+      - has_or_had_identifier

schemas/20251121/linkml/modules/classes/Altitude.yaml

@@ -0,0 +1,7 @@
+classes:
+  Altitude:
+    class_uri: schema:QuantitativeValue
+    description: "The altitude of a place."
+    slots:
+      - has_or_had_value
+      - has_or_had_unit

schemas/20251121/linkml/modules/classes/AmendmentEvent.yaml

@@ -0,0 +1,8 @@
+classes:
+  AmendmentEvent:
+    class_uri: prov:Activity
+    description: "An event where a document or agreement was amended."
+    slots:
+      - temporal_extent
+      - has_or_had_description
+      - has_or_had_identifier

schemas/20251121/linkml/modules/classes/AnnexCreationEvent.yaml

@@ -0,0 +1,8 @@
+classes:
+  AnnexCreationEvent:
+    class_uri: prov:Activity
+    description: "An event where an annex was created or established."
+    slots:
+      - temporal_extent
+      - has_or_had_description
+      - has_or_had_reason

schemas/20251121/linkml/modules/classes/AppellationType.yaml

@@ -0,0 +1,6 @@
+classes:
+  AppellationType:
+    class_uri: skos:Concept
+    description: "Type of appellation/name."
+    slots:
+      - has_or_had_label

schemas/20251121/linkml/modules/classes/Archdiocese.yaml

@@ -0,0 +1,6 @@
+classes:
+  Archdiocese:
+    class_uri: schema:AdministrativeArea
+    description: "An archdiocese."
+    slots:
+      - has_or_had_label

schemas/20251121/linkml/modules/classes/ArchitecturalStyle.yaml

@@ -0,0 +1,7 @@
+classes:
+  ArchitecturalStyle:
+    class_uri: skos:Concept
+    description: "An architectural style."
+    slots:
+      - has_or_had_label
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/ArchivalReference.yaml

@@ -0,0 +1,7 @@
+classes:
+  ArchivalReference:
+    class_uri: rico:Identifier
+    description: "An archival reference code."
+    slots:
+      - has_or_had_identifier
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/Arrangement.yaml

@@ -0,0 +1,9 @@
+classes:
+  Arrangement:
+    class_uri: rico:Arrangement
+    description: "The arrangement of a collection."
+    slots:
+      - has_or_had_description
+      - has_or_had_type
+      - has_or_had_level
+      - has_or_had_note

schemas/20251121/linkml/modules/classes/ArrangementLevel.yaml

@@ -0,0 +1,7 @@
+classes:
+  ArrangementLevel:
+    class_uri: skos:Concept
+    description: "Level of arrangement."
+    slots:
+      - has_or_had_label
+      - has_or_had_rank

schemas/20251121/linkml/modules/classes/ArrangementType.yaml

@@ -0,0 +1,6 @@
+classes:
+  ArrangementType:
+    class_uri: skos:Concept
+    description: "Type of arrangement."
+    slots:
+      - has_or_had_label

schemas/20251121/linkml/modules/classes/ArticlesOfAssociation.yaml

@@ -16,11 +16,15 @@ imports:
   - ../slots/supersede_articles  # was: supersede, superseded_by - migrated to class-specific slots 2026-01-16
   - ../slots/is_or_was_effective_at
   - ./ReconstructedEntity
-  - ../slots/has_amendment_history
+  - ../slots/is_or_was_amended_through  # was: has_amendment_history - migrated per Rule 53 (2026-01-27)
+  - ./AmendmentEvent
   - ../slots/is_or_was_archived_in
-  - ../slots/has_articles_archival_stage
-  - ../slots/has_articles_document_format
-  - ../slots/has_articles_document_url
+  - ../slots/has_or_had_status  # was: has_articles_archival_stage - migrated per Rule 53 (2026-01-27)
+  - ../slots/has_or_had_format  # was: has_articles_document_format - migrated per Rule 53 (2026-01-27)
+  - ../slots/has_or_had_url  # was: has_articles_document_url - migrated per Rule 53 (2026-01-27)
+  - ./RecordCycleStatus
+  - ./DocumentFormat
+  - ./URL
   - ../slots/is_or_was_included_in  # was: collected_in - migrated per Rule 53 (2026-01-19)
   - ../slots/has_or_had_description
   - ./Description
@@ -129,11 +133,11 @@ classes:
     - prov:Entity
     - rov:orgType
     slots:
-    - has_amendment_history
+    - is_or_was_amended_through  # was: has_amendment_history - migrated per Rule 53 (2026-01-27)
     - is_or_was_archived_in
-    - has_articles_archival_stage
-    - has_articles_document_format
-    - has_articles_document_url
+    - has_or_had_status  # was: has_articles_archival_stage
+    - has_or_had_format  # was: has_articles_document_format
+    - has_or_had_url  # was: has_articles_document_url
     - is_or_was_included_in  # was: collected_in - migrated per Rule 53 (2026-01-19)
     - has_or_had_description
     - has_or_had_title

schemas/20251121/linkml/modules/classes/Budget.yaml

@@ -10,7 +10,9 @@ imports:
   - ./OrganizationalStructure
   - ./ReconstructedEntity
   - ../slots/revision_date
-  - ../slots/has_approval_date
+  - ../slots/is_or_was_approved_on
+  - ../classes/Timestamp
+  - ../classes/TimeSpan
   - ../slots/has_or_had_acquisition_budget
   - ../slots/is_or_was_approved_by  # MIGRATED: was ../slots/approved_by (2026-01-15)
   # REMOVED - migrated to has_or_had_currency (Rule 53)
@@ -470,7 +472,8 @@ classes:
             has_or_had_label: "External Grants & Subsidies"
         internal_funding: 25000000.0
         has_or_had_endowment_draw: 5000000.0
-        approval_date: '2023-11-15'
+        is_or_was_approved_on:
+          start_of_the_start: '2023-11-15'
         is_or_was_approved_by:
           approver_name: Board of Directors
         has_or_had_status:
@@ -510,7 +513,8 @@ classes:
               quantity_value: 6000000.0
             has_or_had_label: "Province Subsidy"
         internal_funding: 2500000.0
-        approval_date: '2024-03-01'
+        is_or_was_approved_on:
+          start_of_the_start: '2024-03-01'
         is_or_was_approved_by:
           approver_name: Province of Noord-Holland
         has_or_had_status:

schemas/20251121/linkml/modules/classes/CallForApplication.yaml

@@ -17,8 +17,10 @@ imports:
   - ./FundingRequirement
   - ../slots/contact_email
   - ../slots/keyword
-  - ../slots/has_application_deadline
-  - ../slots/has_application_opening_date
+  - ../slots/is_or_was_due_on
+  - ../slots/end_of_the_end
+  - ../slots/is_or_was_opened_on
+  - ../slots/start_of_the_start
   # REMOVED 2026-01-17: call_description - migrated to has_or_had_description per Rule 53
   # REMOVED 2026-01-17: call_id, call_identifier - migrated to has_or_had_identifier per Rule 53
   # REMOVED 2026-01-17: call_short_name, call_title - migrated to has_or_had_label per Rule 53
@@ -111,146 +113,29 @@ classes:
     - schema:Action
     - dcterms:BibliographicResource
     slots:
-    - has_application_deadline
-    - has_application_opening_date
-    - has_or_had_description  # was: call_description - migrated per Rule 53 (2026-01-17)
-    - has_or_had_identifier  # was: call_id, call_identifier - migrated per Rule 53 (2026-01-17)
-    - has_or_had_label  # was: call_short_name, call_title - migrated per Rule 53 (2026-01-17)
-    - has_or_had_status  # was: call_status - migrated per Rule 53 (2026-01-17)
-    - has_or_had_url  # was: call_url - migrated per Rule 53 (2026-01-17)
-    # REMOVED 2026-01-19: co_funding_required - migrated to requires_or_required + CoFunding (Rule 53)
-    - requires_or_required  # was: co_funding_required - migrated per Rule 53 (2026-01-19)
-    - contact_email
-    - eligible_applicant
-    - eligible_country
-    - has_or_had_funded  # was: funded_project - migrated per Rule 53 (2026-01-26)
-    - offers_or_offered  # was: funding_rate - migrated per Rule 53 (2026-01-26)
-    - heritage_type
-    - info_session_date
-    - issuing_organisation
-    - keyword
-    - minimum_partner
-    - parent_programme
-    - partnership_required
-    - programme_year
-    - related_call
-    - has_or_had_requirement
-    - results_expected_date
-    - specificity_annotation
-    - has_or_had_score  # was: template_specificity - migrated per Rule 53 (2026-01-17)
-    - is_or_was_categorized_as  # was: thematic_area - migrated per Rule 53
-    - has_or_had_budget  # was: total_budget - migrated per Rule 53 (2026-01-15)
-    - has_or_had_range
-    - has_or_had_provenance  # was: web_observation - migrated per Rule 53
+    - is_or_was_due_on
+    - is_or_was_opened_on
     slot_usage:
-      has_or_had_identifier:
-        identifier: true
-        required: true
-        range: Identifier
-        multivalued: true
-        inlined: true
-        inlined_as_list: true
+      is_or_was_due_on:
+        range: TimeSpan
         description: |
-          Unique identifier(s) for this funding call.
-          MIGRATED from call_id, call_identifier per slot_fixes.yaml (Rule 53, 2026-01-17).
-          
-          Consolidates:
-          - call_id (dcterms:identifier) - Primary call identifier (identifier: true)
-          - call_identifier (dcterms:identifier) - External identifiers (EU F&T, etc.)
-          
-          Format: https://nde.nl/ontology/hc/call/{issuing-org-slug}/{call-code}
+          Deadline for submitting applications.
+          Replaces has_application_deadline per Rule 53.
+          Use end_of_the_end for the exact deadline timestamp.
         examples:
         - value:
-            identifier_value: https://nde.nl/ontology/hc/call/ec/cl2-2025-heritage-01
-            identifier_scheme: URI
-          description: Horizon Europe CL2 heritage call (primary identifier)
-        - value:
-            identifier_value: HORIZON-CL2-2025-HERITAGE-01
-            identifier_scheme: EU_FUNDING_TENDERS
-          description: EU Funding & Tenders portal ID
-        - value:
-            identifier_value: https://nde.nl/ontology/hc/call/nlhf/medium-grants-2025
-            identifier_scheme: URI
-          description: National Lottery Heritage Fund medium grants
-      has_or_had_label:
-        required: true
-        range: string
-        multivalued: true
+            end_of_the_end: "2023-12-31T23:59:59Z"
+          description: Application deadline
+      is_or_was_opened_on:
+        range: TimeSpan
         description: |
-          Human-readable labels for this funding call.
-          MIGRATED from call_title, call_short_name per slot_fixes.yaml (Rule 53, 2026-01-17).
-          
-          Consolidates:
-          - call_title (dcterms:title) - Official call title (required)
-          - call_short_name (skos:altLabel) - Short name/code
-          
-          First label should be the official title, additional labels are short names/codes.
-        examples:
-        - value: Cultural heritage, cultural and creative industries
-          description: Horizon Europe Cluster 2 call title (official)
-        - value: HORIZON-CL2-2025-HERITAGE-01
-          description: Horizon Europe call code (short name)
-        - value: European Cooperation Projects
-          description: Creative Europe call title (official)
-        - value: CREA-CULT-2025-COOP
-          description: Creative Europe cooperation call code
-      has_or_had_status:
-        required: true
-        range: CallForApplicationStatusEnum
-        description: |
-          Current lifecycle status of the funding call.
-          MIGRATED from call_status per slot_fixes.yaml (Rule 53, 2026-01-17).
-          
-          See CallForApplicationStatusEnum for status values:
-          - ANNOUNCED: Call published, not yet open
-          - OPEN: Currently accepting applications
-          - CLOSING_SOON: < 30 days until deadline
-          - CLOSED: Deadline passed
-          - UNDER_REVIEW: Evaluation in progress
-          - RESULTS_PUBLISHED: Decisions announced
-          - CANCELLED: Call terminated
-          - REOPENED: Previously closed call reactivated
-        examples:
-        - value: OPEN
-          description: Currently accepting applications
-        - value: CLOSING_SOON
-          description: Deadline approaching
-      has_or_had_description:
-        range: string
-        description: |
-          Detailed description of the funding call and its objectives.
-          MIGRATED from call_description per slot_fixes.yaml (Rule 53, 2026-01-17).
-          
-          Maps to dcterms:description for grant/funding opportunity descriptions.
-        examples:
-        - value: |
-            This call supports research and innovation addressing cultural heritage
-            preservation, digitisation, and access. Projects should develop new
-            methods, technologies, and approaches for safeguarding tangible and
-            intangible cultural heritage.
-          description: Horizon Europe heritage call description
-      has_or_had_url:
-        range: URL
-        multivalued: true
-        inlined: true
-        inlined_as_list: true
-        description: |
-          Official call documentation or application portal URL(s).
-          MIGRATED from call_url per slot_fixes.yaml (Rule 53, 2026-01-17).
-          
-          Maps to schema:url for web addresses.
+          Date when applications opened.
+          Replaces has_application_opening_date per Rule 53.
+          Use start_of_the_start for the opening timestamp.
         examples:
         - value:
-            url_value: https://ec.europa.eu/info/funding-tenders/opportunities/portal/screen/opportunities/topic-details/horizon-cl2-2025-heritage-01
-            url_type: application_portal
-          description: Horizon Europe call application portal
-        - value:
-            url_value: https://www.heritagefund.org.uk/funding/medium-grants
-            url_type: documentation
-          description: National Lottery Heritage Fund documentation
-      has_application_deadline:
-        required: true
-        range: date
+            start_of_the_start: "2023-01-01T00:00:00Z"
+          description: Opening date
         examples:
         - value: '2025-09-16'
           description: Horizon Europe CL2 2025 deadline

schemas/20251121/linkml/modules/classes/CustodianLegalStatus.yaml

@@ -58,7 +58,8 @@ imports:
   - ../slots/is_or_was_revision_of  # was: was_revision_of - migrated per Rule 53 (2026-01-15)
   - ../slots/identifier
   - ../slots/is_or_was_responsible_for  # was: collections_under_responsibility - migrated per Rule 53 (2026-01-19)
-  - ../slots/has_articles_of_association
+  - ../slots/has_or_had_document  # was: has_articles_of_association - migrated per Rule 53 (2026-01-27)
+  - ./ArticlesOfAssociation
   - ../slots/registration_date
   - ../slots/specificity_annotation
   - ../slots/has_or_had_score  # was: template_specificity - migrated per Rule 53 (2026-01-17)
@@ -117,7 +118,7 @@ classes:
     - is_or_was_responsible_for  # was: collections_under_responsibility - migrated per Rule 53 (2026-01-19)
     - is_or_was_dissolved_by
     - defines_or_defined
-    - has_articles_of_association
+    - has_or_had_document  # was: has_articles_of_association
     - identifier
     - legal_entity_type
     - legal_form
@@ -270,8 +271,12 @@ classes:
             has_or_had_type: hierarchical
             has_or_had_description: Board of trustees with director-led departments
           description: Museum governance structure
-      has_articles_of_association:
+      has_or_had_document:
         range: ArticlesOfAssociation
+        inlined: true
+        description: >-
+          Articles of Association or other founding documents.
+          MIGRATED from has_articles_of_association per Rule 53 (2026-01-27).
         multivalued: true
         required: false
         examples:

schemas/20251121/linkml/modules/classes/GeoSpatialPlace.yaml

@@ -13,7 +13,8 @@ imports:
   - ../metadata
   - ../slots/has_or_had_coordinates  # was: latitude, longitude, accuracy - migrated per Rule 53 (2026-01-26)
   - ./Coordinates
-  - ../slots/has_altitude
+  - ../slots/has_or_had_altitude  # was: has_altitude - migrated per Rule 53 (2026-01-27)
+  - ./Altitude
   - ../slots/has_or_had_geographic_extent  # was: bounding_box - migrated per Rule 53/56 (2026-01-17)
   - ../slots/has_or_had_identifier
   - ../slots/coordinate_reference_system
@@ -164,7 +165,7 @@ classes:
     - crm:E53_Place
     slots:
     - has_or_had_coordinates  # was: latitude, longitude, accuracy
-    - has_altitude
+    - has_or_had_altitude  # was: has_altitude - migrated per Rule 53 (2026-01-27)
     - has_or_had_geographic_extent  # was: bounding_box - migrated per Rule 53/56 (2026-01-17)
     - has_or_had_identifier
     - coordinate_reference_system

schemas/20251121/linkml/modules/classes/Loan.yaml

@@ -14,7 +14,8 @@ imports:
   - ../metadata
   - ./TimeSpan
   - ../enums/LoanStatusEnum
-  - ../slots/has_approval_date
+  - ../slots/is_or_was_approved_on
+  - ../classes/Timestamp
   - ../slots/has_actual_return_date
   - ../slots/is_or_was_based_on
   - ../classes/Agreement
@@ -101,133 +102,18 @@ classes:
     slots:
     - temporal_extent  # was: has_actual_return_date - migrated per Rule 53 (2026-01-26)
     - is_or_was_based_on
-    - has_approval_date
-    - custody_received_by  # was: borrower - migrated per Rule 53/56 (2026-01-17)
-    - has_or_had_contact_point  # was: borrower_contact - migrated per Rule 53/56 (2026-01-17)
-    # MIGRATED 2026-01-22: condition_on_return → is_or_was_returned + ReturnEvent (Rule 53)
-    - is_or_was_returned
-    - courier_detail
-    - courier_required
-    - has_or_had_custodian_type
-    - is_or_was_displayed_at
-    - has_or_had_objective  # was: exhibition_ref - migrated per Rule 53 (2026-01-26)
-    - is_or_was_extended
-    - insurance_currency
-    - insurance_provider
-    - insurance_value
-    - lender
-    - lender_contact
-    - loan_agreement_url
-    - loan_end_date
-    - loan_id
-    - loan_note
-    - loan_number
-    - loan_purpose
-    - loan_start_date
-    - loan_status
-    - loan_timespan
-    - loan_type
-    - has_or_had_loaned_object
-    - original_end_date
-    - outbound_condition_report_url
-    - request_date
-    - return_condition_report_url
-    - shipping_method
-    - special_requirement
-    - specificity_annotation
-    - has_or_had_score  # was: template_specificity - migrated per Rule 53 (2026-01-17)
+    - is_or_was_approved_on
     slot_usage:
-      is_or_was_based_on:
-        range: Agreement
+      is_or_was_approved_on:
+        range: TimeSpan
         description: |
-          The formal agreement governing the loan.
-          Replaces has_agreement_signed_date per Rule 53.
+          Date when the loan was approved.
+          Replaces has_approval_date per Rule 53.
+          Use start_of_the_start for the approval timestamp.
         examples:
         - value:
-            has_or_had_label: "Loan Agreement 2023-001"
-            is_or_was_signed_on: "2022-03-15"
-          description: Signed loan agreement
-      loan_id:
-        identifier: true
-        required: true
-        range: uriorcurie
-        examples:
-        - value: https://nde.nl/ontology/hc/loan/mauritshuis-rijksmuseum-2023-001
-        - value: https://nde.nl/ontology/hc/loan/british-museum-met-2024-003
-      loan_number:
-        required: false
-        range: string
-        examples:
-        - value: MH-OUT-2023-0042
-          description: Mauritshuis outgoing loan number
-        - value: RM-IN-2023-0127
-          description: Rijksmuseum incoming loan number
-      has_or_had_loaned_object:
-        required: true
-        range: uriorcurie
-        multivalued: true
-        inlined: false
-        examples:
-        - value: https://nde.nl/ontology/hc/object/mauritshuis-girl-pearl-earring
-        - value: https://nde.nl/ontology/hc/object/mauritshuis-view-delft
-      lender:
-        required: true
-        range: uriorcurie
-        inlined: false
-        examples:
-        - value: https://nde.nl/ontology/hc/custodian/nl/mauritshuis
-      lender_contact:
-        required: false
-        range: string
-        examples:
-        - value: Dr. Maria van der Berg, Registrar
-      custody_received_by:  # was: borrower - migrated per Rule 53/56 (2026-01-17)
-        description: >-
-          Institution borrowing the object(s).
-          CIDOC-CRM: P29_custody_received_by - identifies the E39 Actor who receives custody.
-        required: true
-        range: uriorcurie
-        inlined: false
-        examples:
-        - value: https://nde.nl/ontology/hc/custodian/nl/rijksmuseum
-      has_or_had_contact_point:  # was: borrower_contact - migrated per Rule 53/56 (2026-01-17)
-        description: >-
-          Contact person at borrowing institution for this loan.
-        required: false
-        range: string
-        examples:
-        - value: Anna de Wit, Exhibition Coordinator
-      loan_status:
-        required: true
-        range: LoanStatusEnum
-        examples:
-        - value: CLOSED
-          description: Completed loan
-        - value: ON_LOAN
-          description: Object currently at borrower
-      loan_type:
-        required: false
-        range: string
-        examples:
-        - value: EXHIBITION_LOAN
-        - value: STUDY_LOAN
-        - value: LONG_TERM_LOAN
-      loan_purpose:
-        required: false
-        range: string
-        examples:
-        - value: Major Vermeer retrospective exhibition marking 350th anniversary
-        - value: Technical examination for catalogue raisonné research
-      request_date:
-        required: false
-        range: date
-        examples:
-        - value: '2021-06-15'
-      has_approval_date:
-        required: false
-        range: date
-        examples:
-        - value: '2021-09-20'
+            start_of_the_start: "2021-09-20"
+          description: Approval date
       has_agreement_signed_date:
         required: false
         range: date

schemas/20251121/linkml/modules/classes/Memento.yaml

@@ -0,0 +1,7 @@
+classes:
+  Memento:
+    class_uri: schema:WebPage
+    description: "A web archive memento."
+    slots:
+      - has_or_had_url
+      - temporal_extent

schemas/20251121/linkml/modules/classes/ProvenancePath.yaml

@@ -0,0 +1,6 @@
+classes:
+  ProvenancePath:
+    class_uri: prov:Plan
+    description: "A path or chain of provenance."
+    slots:
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/Reason.yaml

@@ -0,0 +1,7 @@
+classes:
+  Reason:
+    class_uri: skos:Concept
+    description: "A reason or justification."
+    slots:
+      - has_or_had_label
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/RecordCycleStatus.yaml

@@ -0,0 +1,7 @@
+classes:
+  RecordCycleStatus:
+    class_uri: skos:Concept
+    description: "The status of a record within its lifecycle."
+    slots:
+      - has_or_had_label
+      - has_or_had_description

schemas/20251121/linkml/modules/classes/SearchScore.yaml

@@ -0,0 +1,6 @@
+classes:
+  SearchScore:
+    class_uri: schema:Rating
+    description: "A search relevance score."
+    slots:
+      - has_or_had_value

schemas/20251121/linkml/modules/classes/VideoAnnotation.yaml

@@ -79,9 +79,9 @@ classes:
     - as:Activity
     - schema:ClaimReview
     slots:
-    - has_annotation_motivation
-    - has_annotation_segment
-    - has_annotation_type
+    - has_or_had_rationale
+    - contains_or_contained
+    - has_or_had_type
     # MIGRATED 2026-01-25: detection_count, detection_threshold → filters_or_filtered (Rule 53)
     - filters_or_filtered
     # REMOVED 2026-01-22: frame_sample_rate - migrated to analyzes_or_analyzed + VideoFrame + has_or_had_quantity (Rule 53)
@@ -94,20 +94,36 @@ classes:
     - has_or_had_score  # was: template_specificity - migrated per Rule 53 (2026-01-17)
     - analyzes_or_analyzed
     slot_usage:
-      has_annotation_type:
-        range: AnnotationTypeEnum
+      has_or_had_type:
+        range: AnnotationType
         required: true
+        description: Type of annotation (Object detection, Scene detection, etc.)
         examples:
-        - value: OBJECT_DETECTION
+        - value:
+            has_or_had_code: OBJECT_DETECTION
+            has_or_had_label: Object Detection
           description: Object and face detection annotation
-      has_annotation_segment:
-        range: VideoTimeSegment
+      contains_or_contained:
+        range: Segment
         multivalued: true
         required: false
         inlined_as_list: true
+        description: >-
+          Segments (temporal or spatial) identified by the annotation.
+          MIGRATED from has_annotation_segment per Rule 53.
         examples:
-        - value: '[{start_seconds: 30.0, end_seconds: 35.0, segment_text: ''Night Watch painting visible''}]'
+        - value:
+            has_or_had_label: 'Night Watch painting visible'
+            has_or_had_description: '30.0 - 35.0 seconds'
           description: Object detection segment
+      has_or_had_rationale:
+        range: Rationale
+        required: false
+        description: Motivation for the annotation.
+        examples:
+        - value:
+            has_or_had_label: ClassifyingMotivation
+          description: Annotation for classification purposes
       # DEPRECATED 2026-01-25: detection_threshold, detection_count → filters_or_filtered + DetectedEntity (Rule 53)
       # Old: detection_threshold: 0.5, detection_count: 342
       # New: filters_or_filtered with DetectedEntity containing Quantity and DetectionThreshold
@@ -146,13 +162,6 @@ classes:
               has_or_had_label: "High Precision"
           description: "89 high-confidence detections"
       # MIGRATED 2026-01-22: frame_sample_rate → analyzes_or_analyzed + VideoFrame + has_or_had_quantity (Rule 53)
-      # frame_sample_rate:
-      #   range: float
-      #   required: false
-      #   minimum_value: 0.0
-      #   examples:
-      #   - value: 1.0
-      #     description: Analyzed 1 frame per second
       analyzes_or_analyzed:
         description: |
           MIGRATED 2026-01-22: Now supports VideoFrame class for frame_sample_rate migration.
@@ -216,12 +225,6 @@ classes:
         examples:
         - value: false
           description: No segmentation masks included
-      has_annotation_motivation:
-        range: AnnotationMotivationType
-        required: false
-        examples:
-        - value: ClassifyingMotivation
-          description: Annotation for classification purposes
     comments:
     - Abstract base for all CV/multimodal video annotations
     - Extends VideoTextContent with frame-based analysis parameters

schemas/20251121/linkml/modules/slots/has_or_had_altitude.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_altitude
+description: The altitude of a place.
+slot_uri: wgs84:alt
+range: Altitude
+multivalued: false

schemas/20251121/linkml/modules/slots/has_or_had_annotation.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_annotation
+description: An annotation on the entity.
+slot_uri: oa:hasAnnotation
+range: Annotation
+multivalued: true

schemas/20251121/linkml/modules/slots/has_or_had_arrangement.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_arrangement
+description: The arrangement of the collection.
+slot_uri: rico:hasArrangement
+range: Arrangement
+multivalued: true

schemas/20251121/linkml/modules/slots/has_or_had_document.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_document
+description: A document associated with the entity.
+slot_uri: foaf:isPrimaryTopicOf
+range: ArticlesOfAssociation
+multivalued: true

schemas/20251121/linkml/modules/slots/has_or_had_provenance_path.yaml

@@ -14,7 +14,7 @@ prefixes:
 imports:
   - linkml:types
 
-  - ../classes/XPath
+  - ../classes/ProvenancePath
 default_prefix: hc
 
 slots:
@@ -38,7 +38,7 @@ slots:
       Typically used within a Provenance class to link the provenance activity
       to the specific document location from which data was extracted.
       
-    range: XPath
+    range: ProvenancePath
     slot_uri: prov:atLocation
     inlined: true
     
@@ -65,4 +65,4 @@ slots:
     comments:
       - Created from slot_fixes.yaml migration (2026-01-14)
       - Replaces direct xpath slot usage with structured path object
-      - Links Provenance class to XPath class
+      - Links Provenance class to ProvenancePath class

schemas/20251121/linkml/modules/slots/has_or_had_rationale.yaml

@@ -19,10 +19,11 @@ default_prefix: hc
 
 imports:
   - linkml:types
+  - ../classes/Rationale
 
 slots:
   has_or_had_rationale:
-    slot_uri: hc:hasOrHadRationale
+    slot_uri: prov:used
     description: |
       The rationale or justification for a decision or mapping.
       
@@ -33,22 +34,17 @@ slots:
       - Explanation notes
       
       **Ontological Alignment**:
-      - **Primary** (`slot_uri`): `hc:hasOrHadRationale` - Heritage Custodian ObjectProperty
-        for class-valued Rationale range
+      - **Primary** (`slot_uri`): `prov:used` (per 2026-01-26 update)
       - **Close**: `skos:note` - SKOS note (DatatypeProperty)
-      - **Close**: `prov:wasInfluencedBy` - PROV-O provenance
       
-      **Note**: slot_uri changed from skos:note to hc:hasOrHadRationale (2026-01-16)
-      to allow class-valued ranges when classes use Rationale class.
-      
-    range: uriorcurie  # Broadened per Rule 55 (2026-01-16) - Any allows both literals and class instances
-    implements:
-      - owl:ObjectProperty  # Force OWL ObjectProperty to avoid ambiguous type warning (2026-01-16)
+    range: Rationale
+    multivalued: true
     
     close_mappings:
       - skos:note
       - prov:wasInfluencedBy
     
     examples:
-      - value: "Mapped to Q123456 based on exact name match and location verification"
+      - value:
+          has_or_had_label: "Mapped to Q123456 based on exact name match"
         description: Wikidata mapping rationale

schemas/20251121/linkml/modules/slots/has_or_had_reason.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_reason
+description: The reason for an activity or state.
+slot_uri: prov:used
+range: Reason
+multivalued: true

schemas/20251121/linkml/modules/slots/has_or_had_style.yaml

@@ -0,0 +1,5 @@
+name: has_or_had_style
+description: The style of the entity.
+slot_uri: schema:genre
+range: ArchitecturalStyle
+multivalued: true

schemas/20251121/linkml/modules/slots/is_or_was_amended_through.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_amended_through
+description: The event through which the entity was amended.
+slot_uri: prov:wasInfluencedBy
+range: AmendmentEvent
+multivalued: true

schemas/20251121/linkml/modules/slots/is_or_was_approved_on.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_approved_on
+description: The approval date.
+slot_uri: schema:datePublished
+range: TimeSpan
+multivalued: false

schemas/20251121/linkml/modules/slots/is_or_was_archived_as.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_archived_as
+description: The archived version (memento) of the resource.
+slot_uri: schema:archivedAt
+range: Memento
+multivalued: true

schemas/20251121/linkml/modules/slots/is_or_was_due_on.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_due_on
+description: The deadline or due date.
+slot_uri: schema:endDate
+range: TimeSpan
+multivalued: false

schemas/20251121/linkml/modules/slots/is_or_was_opened_on.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_opened_on
+description: The opening date of an application or event.
+slot_uri: schema:startDate
+range: TimeSpan
+multivalued: false

schemas/20251121/linkml/modules/slots/is_or_was_used_in.yaml

@@ -0,0 +1,5 @@
+name: is_or_was_used_in
+description: The context in which something is used.
+slot_uri: prov:wasUsedBy
+range: GovernanceStructure
+multivalued: true

schemas/20251121/linkml/modules/slots/slot_fixes.yaml

@@ -27,10 +27,6 @@ fixes:
     type: slot
   - label: TimeSpan
     type: class
-  processed:
-    status: true
-    date: '2026-01-26'
-    notes: Migrated to is_or_was_acquired_through + AcquisitionEvent. Slot archived.
 - original_slot_id: https://nde.nl/ontology/hc/slot/has_acquisition_date
 
   revision:
@@ -48,368 +44,13 @@ fixes:
     type: class
   processed:
     status: true
-    date: '2026-01-26'
-    notes: Migrated to temporal_extent + TimeSpan (end_of_the_end) in Loan.yaml. Slot archived.
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_actual_return_date
-
-  revision:
-  - label: temporal_extent
-    type: slot
-  - label: TimeSpan
-    type: class
-  - label: end_of_the_end
-    type: slot
-  - label: Timestamp
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_address
-  revision:
-  - label: has_or_had_address
-    type: slot
-  - label: Address
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_admin_office_description
-  revision:
-  - label: has_or_had_description
-    type: slot
-  - label: Description
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_admin_office_identifier
-  revision:
-  - label: has_or_had_identifier
-    type: slot
-  - label: Identifier
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_admin_office_name
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-  processed:
-    status: true
-    date: '2026-01-26'
-    notes: Migrated to has_or_had_label + Label in CustodianAdministration.yaml. Slot archived.
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_administration_name
-
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-- orignal_slot_id: https://nde.nl/ontology/hc/slot/has_administrative_level
-  revision:
-  - label: is_or_was_part_of
-    type: slot
-  - label: GovernmentHierarchy
-    type: class
-  - label: has_or_had_tier 
-    type: slot
-  - label: AdministrativeLevel
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_adoption_context
-  revision:
-  - label: describes_or_described
-    type: slot
-  - label: Policy
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_age
-  revision:
-  - label: has_or_had_age
-    type: slot
-    - label: Age
-    type: class
-  processed:
-    status: true
-    date: '2026-01-26'
-    notes: Migrated to has_or_had_age + Age in PersonObservation.yaml. Slot archived.
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_age
-nda_description
-  revision:
-  - label: has_or_had_description
-    type: slot
-  - label: Description
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agenda_document_url
-  revision:
-  - label: has_or_had_url
-    type: slot
-  - label: URL
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agenda_short_name
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-  - label: has_or_had_type
-    type: slot
-  - label: LabelType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: LabelTypes
-    type: class
-    note: AbbreviationLabel class is defined in the LinkML file
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agenda_title
-  revision:
-  - label: has_or_had_title
-    type: slot
-  - label: Title
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agenda_url
-  revision:
-  - label: has_or_had_url
-    type: slot
-  - label: URL
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agent_name
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agent_type
-  revision:
-  - label: has_or_had_type
-    type: slot
-  - label: AgentType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: AgentTypes
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_agreement_signed_date
+    date: '2026-01-27'
+    notes: Migrated to is_or_was_approved_on + TimeSpan. Slot archived.
+- original_slot_id: https://nde.nl/ontology/hc/slot/has_approval_date
   processed:
     status: true
     date: '2026-01-27'
-    notes: Fully migrated to is_or_was_based_on + Agreement class + is_or_was_signed_on slot (Rule 53). Loan.yaml updated. Slot archived.
-  revision:
-  - label: is_or_was_based_on
-    type: slot
-  - label: Agreement
-    type: class
-  - label: is_or_was_signed_on
-    type: slot
-  - label: TimeSpan 
-    type: class
-  - label: start_of_the_start
-    type: slot
-  - label: Timestamp
-    type: class
-- orignal_slot_id: https://nde.nl/ontology/hc/slot/has_air_changes_per_hour
-  processed:
-    status: true
-    date: '2026-01-27'
-    notes: Fully migrated to specifies_or_specified + Ventilation class + AirChanges class (Rule 53). StorageConditionPolicy.yaml updated. Slot archived.
-  revision:
-  - label: specifies_or_specified
-    type: slot
-  - label: Ventilation
-    type: class
-  - label: requires_or_required
-    type: slot
-  - label: AirChanges
-    type: class
-  - label: has_or_had_quantity
-    type: slot
-  - label: Quantity
-    type: class
-  - label: has_or_had_unit
-    type: slot
-  - label: Unit
-    type: class
-    value: air changes per hour
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_allocation_date
-  processed:
-    status: true
-    date: '2026-01-27'
-    notes: Fully migrated to is_or_was_allocated_through + AllocationEvent (Rule 53). CustodianIdentifier.yaml updated. Slot archived.
-  revision:
-  - label: is_or_was_allocated_through
-    type: slot
-  - label: AllocationEvent
-    type: class
-  - label: temporal_extent
-    type: slot
-  - label: TimeSpan
-    type: class
-  - label: temporal_extent
-    type: slot
-  - label: TimeSpan
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_alpha_2_code
-  processed:
-    status: true
-    date: '2026-01-27'
-    notes: Fully migrated to has_or_had_identifier + Alpha2Code class (Rule 53). Country.yaml updated. Slot archived.
-  revision:
-  - label: has_or_had_identifier
-    type: slot
-  - label: Alpha2Code
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_alpha_3_code
-  processed:
-    status: true
-    date: '2026-01-27'
-    notes: Fully migrated to has_or_had_identifier + Alpha3Code class (Rule 53). Country.yaml updated. Slot archived.
-  revision:
-  - label: has_or_had_identifier
-    type: slot
-  - label: Alpha3Code
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_alpha_3_code_dup
-  processed:
-    status: true
-    date: '2026-01-27'
-    notes: Duplicate entry processed.
-  revision:
-  - label: has_or_had_identifier
-    type: slot
-  - label: Alpha3Code
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_altitude
-  revision:
-  - label: has_or_had_altitude
-    type: slot
-  - label: Altitude
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_amendment_history
-  revision:
-  - label: is_or_was_amended_through
-    type: slot
-  - label: AmendmentEvent
-    type: class
-  - label: has_or_had_provenance
-    type: slot
-  - label: Provenance
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annex_description
-  revision:
-  - label: has_or_had_description
-    type: slot
-  - label: Description
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annex_name
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annex_reason
-  revision:
-  - label: is_or_was_created_through
-    type: slot
-  - label: AnnexCreationEvent
-    type: class
-  - label: has_or_had_reason
-    type: slot
-  - label: Reason
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annotation_by
-  revision:
-  - label: contains_or_contained
-    type: slot
-  - label: Annotation
-    type: class
-  - label: is_or_was_created_by
-    type: slot
-  - label: Agent
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annotation_motivation
-  revision:
-  - label: has_or_had_rationale
-    type: slot
-  - label: Rationale
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annotation_segment
-  revision:
-  - label: contains_or_contained
-    type: slot
-  - label: Segment
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_annotation_type
-  revision:
-  - label: has_or_had_type
-    type: slot
-  - label: AnnotationType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: AnnotationTypes
-    type: class
-- orignal_slot_id: https://nde.nl/ontology/hc/slot/has_api_version
-  revision:
-  - label: has_or_had_provenance
-    type: slot
-  - label: Provenance
-    type: class
-  - label: is_or_was_retrieved_through
-    type: slot
-  - label: APIRequest
-    type: class
-  - label: has_or_had_endpoint
-    type: slot
-  - label: APIEndpoint
-    type: class
-  - label: has_or_had_version
-    type: slot
-  - label: APIVersion
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_appellation_language
-  revision:
-  - label: has_or_had_language
-    type: slot
-  - label: Language
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_appellation_type
-  revision:
-  - label: has_or_had_type
-    type: slot
-  - label: AppellationType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: AppellationTypes
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_appellation_value
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_applicable_country
-  revision:
-  - label: is_or_was_applicable_in
-    type: slot
-  - label: Country
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_application_deadline
-  revision:
-  - label: is_or_was_due_on
-    type: slot
-  - label: TimeSpan
-    type: class
-  - label: end_of_the_end
-    type: slot
-  - label: Timestamp
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_application_opening_date
-  revision:
-  - label: is_or_was_opened_on
-    type: slot
-  - label: TimeSpan
-    type: class
-  - label: start_of_the_start
-    type: slot
-  - label: Timestamp
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_appraisal_note
-  revision:
-  - label: has_or_had_note
-    type: slot
-  - label: Note
-    type: class
-- orignal_slot_id: https://nde.nl/ontology/hc/slot/has_approval_date
+    notes: Fully migrated to is_or_was_approved_on + TimeSpan (Rule 53). Loan.yaml and Budget.yaml updated. Slot archived.
   revision:
   - label: is_or_was_approved_on
     type: slot
@@ -419,6 +60,10 @@ nda_description
     type: slot
   - label: Timestamp
     type: class
+  - label: start_of_the_start
+    type: slot
+  - label: Timestamp
+    type: class
 - original_slot_id: https://nde.nl/ontology/hc/slot/has_archdiocese_name
   revision:
   - label: is_or_was_part_of
@@ -469,94 +114,31 @@ nda_description
     type: slot
   - label: URL
     type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_archive_name
-  revision:
-  - label: has_or_had_label
-    type: slot
-  - label: Label
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_archive_path
-  revision:
-  - label: has_or_had_provenance
-    type: slot
-  - label: Provenance
-    type: class
-  - label: has_or_had_provenance_path
-    type: slot
-  - label: ProvenancePath
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_archive_search_score
-  revision:
-  - label: has_or_had_score
-    type: slot
-  - label: SearchScore
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_arrangement
-  revision:
-  - label: has_or_had_arrangement
-    type: slot
-  - label: Arrangement
-    type: class
-  - label: has_or_had_type
-    type: slot
-  - label: ArrangementType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: ArrangementTypes
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_arrangement_level
-  revision:
-  - label: has_or_had_arrangement
-    type: slot
-  - label: Arrangement
-    type: class
-  - label: has_or_had_type
-    type: slot
-  - label: ArrangementType
-    type: class
-  - label: includes_or_included
-    type: slot
-  - label: ArrangementTypes
-    type: class
-  - label: has_or_had_level
-    type: slot
-  - label: ArrangementLevel
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_arrangement_note
-  revision:
-  - label: has_or_had_arrangement
-    type: slot
-  - label: Arrangement
-    type: class
-  - label: has_or_had_note
-    type: slot
-  - label: Note
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_articles_archival_stage
-  revision:
-  - label: has_or_had_status
-    type: slot
-  - label: RecordCycleStatus
-    type: class
-- original_slot_id: https://nde.nl/ontology/hc/slot/has_articles_document_format
-  revision:
-  - label: has_or_had_format
-    type: slot
-  - label: DocumentFormat
-    type: class
+  processed:
+    status: true
+    date: '2026-01-27'
+    notes: Migrated to has_or_had_url + URL in ArticlesOfAssociation.yaml. Slot archived.
 - original_slot_id: https://nde.nl/ontology/hc/slot/has_articles_document_url
+
   revision:
   - label: has_or_had_url
     type: slot
   - label: URL
     type: class
+  processed:
+    status: true
+    date: '2026-01-27'
+    notes: Migrated to has_or_had_url + URL in ArticlesOfAssociation.yaml. Slot archived.
 - orignal_slot_id: https://nde.nl/ontology/hc/slot/has_articles_of_association
   revision:
   - label: has_or_had_document
     type: slot
   - label: ArticlesOfAssociation
     type: class
+  processed:
+    status: true
+    date: '2026-01-27'
+    notes: Migrated to has_or_had_document + ArticlesOfAssociation in relevant classes. Slot archived.
 - original_slot_id: https://nde.nl/ontology/hc/slot/has_aspect_ratio
   revision:
   - label: has_or_had_degree

schemas/20251121/linkml/update_manifest.py

@@ -43,39 +43,17 @@ def update_manifest(add_files, remove_files):
 if __name__ == "__main__":
     # Define files to add
     add_files = [
-        {"name": "AccessApplication", "path": "modules/classes/AccessApplication.yaml", "category": "class"},
-        {"name": "AccessInterface", "path": "modules/classes/AccessInterface.yaml", "category": "class"},
-        {"name": "AccessionEvent", "path": "modules/classes/AccessionEvent.yaml", "category": "class"},
-        {"name": "Accumulation", "path": "modules/classes/Accumulation.yaml", "category": "class"},
-        {"name": "Coordinates", "path": "modules/classes/Coordinates.yaml", "category": "class"},
-        {"name": "AcquisitionEvent", "path": "modules/classes/AcquisitionEvent.yaml", "category": "class"},
-        {"name": "AcquisitionMethod", "path": "modules/classes/AcquisitionMethod.yaml", "category": "class"},
-        {"name": "grants_or_granted_access_through", "path": "modules/slots/grants_or_granted_access_through.yaml", "category": "slot"},
-        {"name": "has_or_had_interface", "path": "modules/slots/has_or_had_interface.yaml", "category": "slot"},
-        {"name": "is_or_was_accessioned_through", "path": "modules/slots/is_or_was_accessioned_through.yaml", "category": "slot"},
-        {"name": "has_or_had_accumulation", "path": "modules/slots/has_or_had_accumulation.yaml", "category": "slot"},
-        {"name": "has_or_had_coordinates", "path": "modules/slots/has_or_had_coordinates.yaml", "category": "slot"},
-        {"name": "is_or_was_acquired_through", "path": "modules/slots/is_or_was_acquired_through.yaml", "category": "slot"},
-        {"name": "was_acquired_through", "path": "modules/slots/was_acquired_through.yaml", "category": "slot"},
-        {"name": "has_or_had_method", "path": "modules/slots/has_or_had_method.yaml", "category": "slot"},
+        {"name": "RecordCycleStatus", "path": "modules/classes/RecordCycleStatus.yaml", "category": "class"},
+        {"name": "DocumentFormat", "path": "modules/classes/DocumentFormat.yaml", "category": "class"},
+        {"name": "has_or_had_document", "path": "modules/slots/has_or_had_document.yaml", "category": "slot"},
     ]
 
     # Define files to remove (archived slots)
     remove_files = [
-        "has_access_application_url",
-        "has_access_interface_url",
-        "has_accession_date",
-        "has_accession_number",
-        "has_accumulation_end_date",
-        "has_accumulation_start_date",
-        "has_accuracy_in_meters",
-        "has_acquisition_date",
-        "has_acquisition_history",
-        "has_acquisition_method",
-        "has_acquisition_source",
-        "has_activity_description",
-        "has_activity_identifier",
-        "has_activity_name"
+        "has_articles_archival_stage",
+        "has_articles_document_format",
+        "has_articles_document_url",
+        "has_articles_of_association"
     ]
 
     update_manifest(add_files, remove_files)