rdf-starbase 0.1.0__py3-none-any.whl

rdf_starbase/repository_api.py

@@ -0,0 +1,739 @@
+"""
+Repository Management API Router.
+
+Provides REST endpoints for managing multiple repositories:
+- Create/delete repositories
+- List repositories
+- Get repository info
+- Scoped SPARQL queries per repository
+"""
+
+from pathlib import Path
+from typing import Optional, Union
+import os
+
+from fastapi import APIRouter, HTTPException, Query
+from pydantic import BaseModel, Field
+import polars as pl
+
+from rdf_starbase.repositories import RepositoryManager, RepositoryInfo
+from rdf_starbase import execute_sparql
+
+
+# =============================================================================
+# Pydantic Models
+# =============================================================================
+
+class CreateRepositoryRequest(BaseModel):
+    """Request to create a new repository."""
+    name: str = Field(..., description="Unique repository name (alphanumeric, hyphens, underscores)")
+    description: str = Field(default="", description="Human-readable description")
+    tags: list[str] = Field(default_factory=list, description="Optional tags")
+
+
+class UpdateRepositoryRequest(BaseModel):
+    """Request to update repository metadata."""
+    description: Optional[str] = Field(None, description="New description")
+    tags: Optional[list[str]] = Field(None, description="New tags")
+
+
+class RenameRepositoryRequest(BaseModel):
+    """Request to rename a repository."""
+    new_name: str = Field(..., description="New repository name")
+
+
+class SPARQLQueryRequest(BaseModel):
+    """SPARQL query for a specific repository."""
+    query: str = Field(..., description="SPARQL-Star query string")
+
+
+class RepositoryResponse(BaseModel):
+    """Response containing repository info."""
+    name: str
+    description: str
+    tags: list[str]
+    created_at: str
+    triple_count: int
+    subject_count: int
+    predicate_count: int
+    
+    @classmethod
+    def from_info(cls, info: RepositoryInfo) -> "RepositoryResponse":
+        return cls(
+            name=info.name,
+            description=info.description,
+            tags=info.tags,
+            created_at=info.created_at.isoformat(),
+            triple_count=info.triple_count,
+            subject_count=info.subject_count,
+            predicate_count=info.predicate_count,
+        )
+
+
+def dataframe_to_records(df: pl.DataFrame) -> list[dict]:
+    """Convert Polars DataFrame to list of dicts for JSON serialization."""
+    from datetime import datetime
+    records = []
+    for row in df.iter_rows(named=True):
+        record = {}
+        for k, v in row.items():
+            if isinstance(v, datetime):
+                record[k] = v.isoformat()
+            elif v is None:
+                record[k] = None
+            else:
+                record[k] = v
+        records.append(record)
+    return records
+
+
+def create_repository_router(
+    workspace_path: Optional[str | Path] = None
+) -> tuple[APIRouter, RepositoryManager]:
+    """
+    Create the repository management API router.
+    
+    Args:
+        workspace_path: Path to store repositories (default: ./data/repositories)
+        
+    Returns:
+        Tuple of (router, manager)
+    """
+    # Default workspace path
+    if workspace_path is None:
+        workspace_path = os.environ.get(
+            "RDFSTARBASE_REPOSITORY_PATH",
+            "./data/repositories"
+        )
+    
+    manager = RepositoryManager(workspace_path)
+    router = APIRouter(prefix="/repositories", tags=["Repositories"])
+    
+    # =========================================================================
+    # Repository CRUD
+    # =========================================================================
+    
+    @router.get("")
+    async def list_repositories():
+        """List all repositories."""
+        repos = manager.list_repositories()
+        return {
+            "count": len(repos),
+            "repositories": [RepositoryResponse.from_info(r).model_dump() for r in repos]
+        }
+    
+    @router.post("")
+    async def create_repository(request: CreateRepositoryRequest):
+        """Create a new repository."""
+        try:
+            info = manager.create(
+                name=request.name,
+                description=request.description,
+                tags=request.tags,
+            )
+            # Auto-save after creation
+            manager.save(request.name)
+            return {
+                "success": True,
+                "message": f"Repository '{request.name}' created",
+                "repository": RepositoryResponse.from_info(info).model_dump()
+            }
+        except ValueError as e:
+            raise HTTPException(status_code=400, detail=str(e))
+    
+    @router.get("/{name}")
+    async def get_repository(name: str):
+        """Get repository info."""
+        try:
+            info = manager.get_info(name)
+            return RepositoryResponse.from_info(info).model_dump()
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+    
+    @router.patch("/{name}")
+    async def update_repository(name: str, request: UpdateRepositoryRequest):
+        """Update repository metadata."""
+        try:
+            info = manager.update_info(
+                name=name,
+                description=request.description,
+                tags=request.tags,
+            )
+            return RepositoryResponse.from_info(info).model_dump()
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+    
+    @router.delete("/{name}")
+    async def delete_repository(
+        name: str,
+        force: bool = Query(False, description="Force delete even if repository has data")
+    ):
+        """Delete a repository."""
+        try:
+            manager.delete(name, force=force)
+            return {
+                "success": True,
+                "message": f"Repository '{name}' deleted"
+            }
+        except ValueError as e:
+            if "does not exist" in str(e):
+                raise HTTPException(status_code=404, detail=str(e))
+            raise HTTPException(status_code=400, detail=str(e))
+    
+    @router.post("/{name}/rename")
+    async def rename_repository(name: str, request: RenameRepositoryRequest):
+        """Rename a repository."""
+        try:
+            info = manager.rename(name, request.new_name)
+            return {
+                "success": True,
+                "message": f"Repository renamed from '{name}' to '{request.new_name}'",
+                "repository": RepositoryResponse.from_info(info).model_dump()
+            }
+        except ValueError as e:
+            if "does not exist" in str(e):
+                raise HTTPException(status_code=404, detail=str(e))
+            raise HTTPException(status_code=400, detail=str(e))
+    
+    # =========================================================================
+    # Repository SPARQL
+    # =========================================================================
+    
+    @router.post("/{name}/sparql")
+    async def repository_sparql(name: str, request: SPARQLQueryRequest):
+        """Execute a SPARQL query against a specific repository."""
+        try:
+            store = manager.get_store(name)
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+        
+        try:
+            result = execute_sparql(store, request.query)
+            
+            if isinstance(result, bool):
+                # ASK query
+                return {"type": "ask", "result": result}
+            elif isinstance(result, dict):
+                # UPDATE operation
+                # Auto-save after update
+                manager.save(name)
+                return {
+                    "type": "update",
+                    "operation": result.get("operation", "unknown"),
+                    "count": result.get("count", 0),
+                    "success": True,
+                }
+            elif isinstance(result, pl.DataFrame):
+                # SELECT query
+                return {
+                    "type": "select",
+                    "count": len(result),
+                    "columns": result.columns,
+                    "results": dataframe_to_records(result),
+                }
+            else:
+                return {"type": "unknown", "result": str(result)}
+                
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=f"Query error: {str(e)}")
+    
+    # =========================================================================
+    # Repository Triple Management
+    # =========================================================================
+    
+    @router.get("/{name}/triples")
+    async def get_repository_triples(
+        name: str,
+        subject: Optional[str] = Query(None, description="Filter by subject"),
+        predicate: Optional[str] = Query(None, description="Filter by predicate"),
+        object: Optional[str] = Query(None, description="Filter by object"),
+        limit: int = Query(100, ge=1, le=10000, description="Maximum results"),
+    ):
+        """Get triples from a specific repository."""
+        try:
+            store = manager.get_store(name)
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+        
+        df = store.get_triples(
+            subject=subject,
+            predicate=predicate,
+            obj=object,
+        )
+        
+        df = df.head(limit)
+        
+        return {
+            "count": len(df),
+            "triples": dataframe_to_records(df),
+        }
+    
+    @router.post("/{name}/triples/batch")
+    async def add_repository_triples_batch(
+        name: str,
+        triples: list[dict]
+    ):
+        """Add multiple triples to a specific repository."""
+        try:
+            store = manager.get_store(name)
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+        
+        try:
+            count = store.add_triples_batch(triples)
+            # Auto-save after batch insert
+            manager.save(name)
+            return {
+                "success": True,
+                "count": count,
+                "message": f"Added {count} triples to repository '{name}'",
+            }
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=str(e))
+    
+    # =========================================================================
+    # Repository Stats
+    # =========================================================================
+    
+    @router.get("/{name}/stats")
+    async def get_repository_stats(name: str):
+        """Get detailed statistics for a repository."""
+        try:
+            store = manager.get_store(name)
+            info = manager.get_info(name)
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+        
+        return {
+            "name": name,
+            "description": info.description,
+            "created_at": info.created_at.isoformat(),
+            "stats": store.stats(),
+        }
+    
+    # =========================================================================
+    # Persistence
+    # =========================================================================
+    
+    @router.post("/{name}/save")
+    async def save_repository(name: str):
+        """Explicitly save a repository to disk."""
+        try:
+            manager.save(name)
+            return {
+                "success": True,
+                "message": f"Repository '{name}' saved"
+            }
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+    
+    @router.post("/save-all")
+    async def save_all_repositories():
+        """Save all loaded repositories to disk."""
+        manager.save_all()
+        return {
+            "success": True,
+            "message": "All repositories saved"
+        }
+    
+    # =========================================================================
+    # Example Datasets
+    # =========================================================================
+    
+    @router.get("/examples/datasets")
+    async def list_example_datasets():
+        """List available example datasets."""
+        return {
+            "datasets": [
+                {
+                    "id": "movies",
+                    "name": "Movies & Directors",
+                    "description": "Sample movie data with directors, actors, and genres. Great for learning SPARQL.",
+                    "triple_count": 45,
+                    "tags": ["movies", "entertainment", "schema.org"]
+                },
+                {
+                    "id": "techcorp",
+                    "name": "TechCorp Customer Service",
+                    "description": "Customer service scenario with tickets, products, and customer data. Includes conflicting data from multiple sources.",
+                    "triple_count": 35,
+                    "tags": ["enterprise", "CRM", "support"]
+                },
+                {
+                    "id": "knowledge_graph",
+                    "name": "Simple Knowledge Graph",
+                    "description": "Basic knowledge graph with people, organizations, and relationships.",
+                    "triple_count": 28,
+                    "tags": ["people", "organizations", "relationships"]
+                },
+                {
+                    "id": "rdf_star_demo",
+                    "name": "RDF-Star Demo",
+                    "description": "Demonstrates RDF-Star features with quoted triples, annotations, and provenance metadata.",
+                    "triple_count": 22,
+                    "tags": ["rdf-star", "annotations", "provenance"]
+                }
+            ]
+        }
+    
+    @router.post("/{name}/load-example/{dataset_id}")
+    async def load_example_dataset(name: str, dataset_id: str):
+        """Load an example dataset into a repository."""
+        try:
+            store = manager.get_store(name)
+        except ValueError as e:
+            raise HTTPException(status_code=404, detail=str(e))
+        
+        datasets = {
+            "movies": get_movies_dataset_triples,
+            "techcorp": get_techcorp_dataset_triples,
+            "knowledge_graph": get_knowledge_graph_dataset_triples,
+            "rdf_star_demo": get_rdf_star_demo_dataset_triples,
+        }
+        
+        if dataset_id not in datasets:
+            raise HTTPException(status_code=404, detail=f"Unknown dataset: {dataset_id}")
+        
+        triples = datasets[dataset_id]()
+        
+        try:
+            count = store.add_triples_batch(triples)
+            manager.save(name)
+            return {
+                "success": True,
+                "dataset": dataset_id,
+                "message": f"Loaded example dataset '{dataset_id}' into repository '{name}'",
+                "stats": store.stats()
+            }
+        except Exception as e:
+            raise HTTPException(status_code=400, detail=f"Failed to load dataset: {str(e)}")
+    
+    return router, manager
+
+
+# =============================================================================
+# Example Datasets with Provenance
+# =============================================================================
+
+def get_movies_dataset_triples() -> list[dict]:
+    """Movies & Directors dataset with varied sources and confidence."""
+    RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
+    SCHEMA = "http://schema.org/"
+    EX = "http://example.org/"
+    
+    triples = []
+    
+    # Helper to add triple with provenance
+    def add(s, p, o, source, confidence):
+        triples.append({
+            "subject": s,
+            "predicate": p,
+            "object": o,
+            "source": source,
+            "confidence": confidence
+        })
+    
+    # Directors - from Wikipedia (high confidence)
+    add(f"{EX}person/nolan", RDF_TYPE, f"{SCHEMA}Person", "Wikipedia", 0.98)
+    add(f"{EX}person/nolan", f"{SCHEMA}name", "Christopher Nolan", "Wikipedia", 0.99)
+    add(f"{EX}person/nolan", f"{SCHEMA}birthDate", "1970-07-30", "Wikipedia", 0.95)
+    add(f"{EX}person/nolan", f"{SCHEMA}nationality", "British-American", "Wikipedia", 0.92)
+    
+    add(f"{EX}person/spielberg", RDF_TYPE, f"{SCHEMA}Person", "Wikipedia", 0.98)
+    add(f"{EX}person/spielberg", f"{SCHEMA}name", "Steven Spielberg", "Wikipedia", 0.99)
+    add(f"{EX}person/spielberg", f"{SCHEMA}birthDate", "1946-12-18", "Wikipedia", 0.97)
+    
+    add(f"{EX}person/greta", RDF_TYPE, f"{SCHEMA}Person", "IMDb", 0.94)
+    add(f"{EX}person/greta", f"{SCHEMA}name", "Greta Gerwig", "IMDb", 0.96)
+    add(f"{EX}person/greta", f"{SCHEMA}birthDate", "1983-08-04", "IMDb", 0.90)
+    
+    # Actors - from IMDb (good confidence)
+    add(f"{EX}person/dicaprio", RDF_TYPE, f"{SCHEMA}Person", "IMDb", 0.97)
+    add(f"{EX}person/dicaprio", f"{SCHEMA}name", "Leonardo DiCaprio", "IMDb", 0.99)
+    add(f"{EX}person/dicaprio", f"{SCHEMA}birthDate", "1974-11-11", "IMDb", 0.95)
+    
+    add(f"{EX}person/cillian", RDF_TYPE, f"{SCHEMA}Person", "IMDb", 0.96)
+    add(f"{EX}person/cillian", f"{SCHEMA}name", "Cillian Murphy", "IMDb", 0.98)
+    
+    add(f"{EX}person/margot", RDF_TYPE, f"{SCHEMA}Person", "IMDb", 0.97)
+    add(f"{EX}person/margot", f"{SCHEMA}name", "Margot Robbie", "IMDb", 0.99)
+    
+    # Movies - mixed sources
+    # Inception - from multiple sources
+    add(f"{EX}movie/inception", RDF_TYPE, f"{SCHEMA}Movie", "IMDb", 0.99)
+    add(f"{EX}movie/inception", f"{SCHEMA}name", "Inception", "IMDb", 0.99)
+    add(f"{EX}movie/inception", f"{SCHEMA}datePublished", "2010", "IMDb", 0.98)
+    add(f"{EX}movie/inception", f"{SCHEMA}director", f"{EX}person/nolan", "IMDb", 0.99)
+    add(f"{EX}movie/inception", f"{SCHEMA}actor", f"{EX}person/dicaprio", "IMDb", 0.99)
+    add(f"{EX}movie/inception", f"{SCHEMA}genre", "Sci-Fi", "RottenTomatoes", 0.85)
+    add(f"{EX}movie/inception", f"{SCHEMA}duration", "PT2H28M", "IMDb", 0.97)
+    
+    # Oppenheimer - recent film, high confidence
+    add(f"{EX}movie/oppenheimer", RDF_TYPE, f"{SCHEMA}Movie", "IMDb", 0.99)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}name", "Oppenheimer", "IMDb", 0.99)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}datePublished", "2023", "BoxOfficeMojo", 0.98)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}director", f"{EX}person/nolan", "Wikipedia", 0.99)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}actor", f"{EX}person/cillian", "IMDb", 0.98)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}genre", "Drama", "RottenTomatoes", 0.88)
+    add(f"{EX}movie/oppenheimer", f"{SCHEMA}genre", "Biography", "Wikipedia", 0.82)  # Multiple genres!
+    
+    # Interstellar
+    add(f"{EX}movie/interstellar", RDF_TYPE, f"{SCHEMA}Movie", "IMDb", 0.99)
+    add(f"{EX}movie/interstellar", f"{SCHEMA}name", "Interstellar", "IMDb", 0.99)
+    add(f"{EX}movie/interstellar", f"{SCHEMA}datePublished", "2014", "IMDb", 0.98)
+    add(f"{EX}movie/interstellar", f"{SCHEMA}director", f"{EX}person/nolan", "Wikipedia", 0.99)
+    add(f"{EX}movie/interstellar", f"{SCHEMA}genre", "Sci-Fi", "IMDb", 0.92)
+    
+    # Jurassic Park - classic film
+    add(f"{EX}movie/jurassic_park", RDF_TYPE, f"{SCHEMA}Movie", "Wikipedia", 0.99)
+    add(f"{EX}movie/jurassic_park", f"{SCHEMA}name", "Jurassic Park", "Wikipedia", 0.99)
+    add(f"{EX}movie/jurassic_park", f"{SCHEMA}datePublished", "1993", "Wikipedia", 0.99)
+    add(f"{EX}movie/jurassic_park", f"{SCHEMA}director", f"{EX}person/spielberg", "Wikipedia", 0.99)
+    add(f"{EX}movie/jurassic_park", f"{SCHEMA}genre", "Adventure", "IMDb", 0.90)
+    
+    # Barbie - from entertainment news (slightly lower confidence)
+    add(f"{EX}movie/barbie", RDF_TYPE, f"{SCHEMA}Movie", "BoxOfficeMojo", 0.97)
+    add(f"{EX}movie/barbie", f"{SCHEMA}name", "Barbie", "BoxOfficeMojo", 0.99)
+    add(f"{EX}movie/barbie", f"{SCHEMA}datePublished", "2023", "BoxOfficeMojo", 0.98)
+    add(f"{EX}movie/barbie", f"{SCHEMA}director", f"{EX}person/greta", "IMDb", 0.97)
+    add(f"{EX}movie/barbie", f"{SCHEMA}actor", f"{EX}person/margot", "IMDb", 0.98)
+    add(f"{EX}movie/barbie", f"{SCHEMA}genre", "Comedy", "RottenTomatoes", 0.80)
+    
+    return triples
+
+
+def get_techcorp_dataset_triples() -> list[dict]:
+    """TechCorp Customer Service dataset with conflicting data from multiple sources."""
+    RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
+    TC = "http://techcorp.com/"
+    FOAF = "http://xmlns.com/foaf/0.1/"
+    RDFS = "http://www.w3.org/2000/01/rdf-schema#"
+    
+    triples = []
+    
+    def add(s, p, o, source, confidence):
+        triples.append({
+            "subject": s,
+            "predicate": p,
+            "object": o,
+            "source": source,
+            "confidence": confidence
+        })
+    
+    # Customer C001 - Alice Johnson (CONFLICTING DATA from different systems!)
+    add(f"{TC}customer/C001", RDF_TYPE, f"{TC}Customer", "CRM_System", 0.99)
+    add(f"{TC}customer/C001", f"{FOAF}name", "Alice Johnson", "CRM_System", 0.98)
+    add(f"{TC}customer/C001", f"{TC}email", "alice@example.com", "CRM_System", 0.95)
+    add(f"{TC}customer/C001", f"{TC}tier", "Premium", "CRM_System", 0.90)
+    add(f"{TC}customer/C001", f"{TC}since", "2020-01-15", "CRM_System", 0.99)
+    
+    # CONFLICT: Billing system says different tier!
+    add(f"{TC}customer/C001", f"{TC}tier", "Enterprise", "Billing_System", 0.85)
+    # CONFLICT: Support portal has different email
+    add(f"{TC}customer/C001", f"{TC}email", "a.johnson@corp.example.com", "Support_Portal", 0.75)
+    
+    # Customer C002 - Bob Smith
+    add(f"{TC}customer/C002", RDF_TYPE, f"{TC}Customer", "CRM_System", 0.99)
+    add(f"{TC}customer/C002", f"{FOAF}name", "Bob Smith", "CRM_System", 0.97)
+    add(f"{TC}customer/C002", f"{TC}email", "bob@example.com", "CRM_System", 0.96)
+    add(f"{TC}customer/C002", f"{TC}tier", "Standard", "CRM_System", 0.92)
+    add(f"{TC}customer/C002", f"{TC}since", "2021-06-20", "CRM_System", 0.99)
+    
+    # Customer C003 - Carol White (from legacy system)
+    add(f"{TC}customer/C003", RDF_TYPE, f"{TC}Customer", "Legacy_DB", 0.88)
+    add(f"{TC}customer/C003", f"{FOAF}name", "Carol White", "Legacy_DB", 0.85)
+    add(f"{TC}customer/C003", f"{TC}email", "carol@example.com", "Legacy_DB", 0.80)
+    add(f"{TC}customer/C003", f"{TC}tier", "Enterprise", "Billing_System", 0.95)
+    
+    # Products - from Product Catalog (high confidence)
+    add(f"{TC}product/P001", RDF_TYPE, f"{TC}Product", "Product_Catalog", 0.99)
+    add(f"{TC}product/P001", f"{RDFS}label", "CloudSync Pro", "Product_Catalog", 0.99)
+    add(f"{TC}product/P001", f"{TC}category", "Software", "Product_Catalog", 0.98)
+    add(f"{TC}product/P001", f"{TC}price", "299.99", "Product_Catalog", 0.97)
+    # CONFLICT: Sales team has different price!
+    add(f"{TC}product/P001", f"{TC}price", "279.99", "Sales_Team", 0.70)
+    
+    add(f"{TC}product/P002", RDF_TYPE, f"{TC}Product", "Product_Catalog", 0.99)
+    add(f"{TC}product/P002", f"{RDFS}label", "DataVault", "Product_Catalog", 0.99)
+    add(f"{TC}product/P002", f"{TC}category", "Storage", "Product_Catalog", 0.98)
+    add(f"{TC}product/P002", f"{TC}price", "499.99", "Product_Catalog", 0.97)
+    
+    add(f"{TC}product/P003", RDF_TYPE, f"{TC}Product", "Product_Catalog", 0.99)
+    add(f"{TC}product/P003", f"{RDFS}label", "SecureNet", "Product_Catalog", 0.99)
+    add(f"{TC}product/P003", f"{TC}category", "Security", "Product_Catalog", 0.98)
+    add(f"{TC}product/P003", f"{TC}price", "199.99", "Product_Catalog", 0.97)
+    
+    # Support Tickets - from different support channels
+    add(f"{TC}ticket/T001", RDF_TYPE, f"{TC}SupportTicket", "Support_Portal", 0.99)
+    add(f"{TC}ticket/T001", f"{TC}customer", f"{TC}customer/C001", "Support_Portal", 0.99)
+    add(f"{TC}ticket/T001", f"{TC}product", f"{TC}product/P001", "Support_Portal", 0.98)
+    add(f"{TC}ticket/T001", f"{TC}status", "Open", "Support_Portal", 0.95)
+    add(f"{TC}ticket/T001", f"{TC}priority", "High", "Support_Portal", 0.90)
+    add(f"{TC}ticket/T001", f"{TC}description", "Sync failing intermittently", "Support_Portal", 0.99)
+    
+    add(f"{TC}ticket/T002", RDF_TYPE, f"{TC}SupportTicket", "Email_Integration", 0.95)
+    add(f"{TC}ticket/T002", f"{TC}customer", f"{TC}customer/C002", "Email_Integration", 0.92)
+    add(f"{TC}ticket/T002", f"{TC}product", f"{TC}product/P002", "Email_Integration", 0.90)
+    add(f"{TC}ticket/T002", f"{TC}status", "Resolved", "Support_Portal", 0.98)
+    add(f"{TC}ticket/T002", f"{TC}priority", "Medium", "Email_Integration", 0.85)
+    add(f"{TC}ticket/T002", f"{TC}description", "Storage quota question", "Email_Integration", 0.88)
+    
+    add(f"{TC}ticket/T003", RDF_TYPE, f"{TC}SupportTicket", "Security_Ops", 0.99)
+    add(f"{TC}ticket/T003", f"{TC}customer", f"{TC}customer/C003", "Security_Ops", 0.98)
+    add(f"{TC}ticket/T003", f"{TC}product", f"{TC}product/P003", "Security_Ops", 0.99)
+    add(f"{TC}ticket/T003", f"{TC}status", "Open", "Security_Ops", 0.99)
+    add(f"{TC}ticket/T003", f"{TC}priority", "Critical", "Security_Ops", 0.99)
+    add(f"{TC}ticket/T003", f"{TC}description", "Security alert investigation", "Security_Ops", 0.99)
+    
+    return triples
+
+
+def get_knowledge_graph_dataset_triples() -> list[dict]:
+    """Simple Knowledge Graph with people and organizations."""
+    RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
+    FOAF = "http://xmlns.com/foaf/0.1/"
+    ORG = "http://www.w3.org/ns/org#"
+    EX = "http://example.org/"
+    
+    triples = []
+    
+    def add(s, p, o, source, confidence):
+        triples.append({
+            "subject": s,
+            "predicate": p,
+            "object": o,
+            "source": source,
+            "confidence": confidence
+        })
+    
+    # People - from LinkedIn (professional network)
+    add(f"{EX}person/jane", RDF_TYPE, f"{FOAF}Person", "LinkedIn", 0.95)
+    add(f"{EX}person/jane", f"{FOAF}name", "Jane Doe", "LinkedIn", 0.98)
+    add(f"{EX}person/jane", f"{FOAF}age", "32", "LinkedIn", 0.75)  # Age less reliable
+    add(f"{EX}person/jane", f"{FOAF}knows", f"{EX}person/john", "LinkedIn", 0.90)
+    add(f"{EX}person/jane", f"{FOAF}knows", f"{EX}person/alice", "LinkedIn", 0.92)
+    add(f"{EX}person/jane", f"{ORG}memberOf", f"{EX}org/acme", "LinkedIn", 0.97)
+    
+    add(f"{EX}person/john", RDF_TYPE, f"{FOAF}Person", "LinkedIn", 0.95)
+    add(f"{EX}person/john", f"{FOAF}name", "John Smith", "LinkedIn", 0.97)
+    add(f"{EX}person/john", f"{FOAF}age", "28", "Facebook", 0.70)  # Different source
+    add(f"{EX}person/john", f"{FOAF}knows", f"{EX}person/jane", "LinkedIn", 0.90)
+    add(f"{EX}person/john", f"{ORG}memberOf", f"{EX}org/globex", "CompanyWebsite", 0.99)
+    
+    add(f"{EX}person/alice", RDF_TYPE, f"{FOAF}Person", "LinkedIn", 0.96)
+    add(f"{EX}person/alice", f"{FOAF}name", "Alice Chen", "LinkedIn", 0.98)
+    add(f"{EX}person/alice", f"{FOAF}age", "35", "LinkedIn", 0.72)
+    add(f"{EX}person/alice", f"{FOAF}knows", f"{EX}person/jane", "LinkedIn", 0.92)
+    add(f"{EX}person/alice", f"{FOAF}knows", f"{EX}person/bob", "Email_Analysis", 0.65)
+    add(f"{EX}person/alice", f"{ORG}headOf", f"{EX}org/acme", "CompanyWebsite", 0.99)
+    
+    add(f"{EX}person/bob", RDF_TYPE, f"{FOAF}Person", "HR_System", 0.98)
+    add(f"{EX}person/bob", f"{FOAF}name", "Bob Williams", "HR_System", 0.99)
+    add(f"{EX}person/bob", f"{FOAF}age", "42", "HR_System", 0.95)
+    add(f"{EX}person/bob", f"{ORG}memberOf", f"{EX}org/initech", "HR_System", 0.98)
+    
+    # Organizations - from company registries
+    add(f"{EX}org/acme", RDF_TYPE, f"{ORG}Organization", "SEC_Filings", 0.99)
+    add(f"{EX}org/acme", f"{FOAF}name", "Acme Corp", "SEC_Filings", 0.99)
+    add(f"{EX}org/acme", f"{ORG}hasSite", f"{EX}location/sf", "CompanyWebsite", 0.95)
+    add(f"{EX}org/acme", f"{EX}industry", "Technology", "CrunchBase", 0.88)
+    
+    add(f"{EX}org/globex", RDF_TYPE, f"{ORG}Organization", "SEC_Filings", 0.99)
+    add(f"{EX}org/globex", f"{FOAF}name", "Globex Inc", "SEC_Filings", 0.99)
+    add(f"{EX}org/globex", f"{ORG}hasSite", f"{EX}location/nyc", "CompanyWebsite", 0.96)
+    add(f"{EX}org/globex", f"{EX}industry", "Finance", "Bloomberg", 0.95)
+    
+    add(f"{EX}org/initech", RDF_TYPE, f"{ORG}Organization", "State_Registry", 0.97)
+    add(f"{EX}org/initech", f"{FOAF}name", "Initech", "State_Registry", 0.98)
+    add(f"{EX}org/initech", f"{ORG}hasSite", f"{EX}location/austin", "GoogleMaps", 0.85)
+    add(f"{EX}org/initech", f"{EX}industry", "Software", "LinkedIn", 0.80)
+    
+    # Locations - from geographic databases
+    add(f"{EX}location/sf", RDF_TYPE, f"{EX}Location", "GeoNames", 0.99)
+    add(f"{EX}location/sf", f"{FOAF}name", "San Francisco", "GeoNames", 0.99)
+    add(f"{EX}location/sf", f"{EX}country", "USA", "GeoNames", 0.99)
+    
+    add(f"{EX}location/nyc", RDF_TYPE, f"{EX}Location", "GeoNames", 0.99)
+    add(f"{EX}location/nyc", f"{FOAF}name", "New York City", "GeoNames", 0.99)
+    add(f"{EX}location/nyc", f"{EX}country", "USA", "GeoNames", 0.99)
+    
+    add(f"{EX}location/austin", RDF_TYPE, f"{EX}Location", "GeoNames", 0.99)
+    add(f"{EX}location/austin", f"{FOAF}name", "Austin", "GeoNames", 0.99)
+    add(f"{EX}location/austin", f"{EX}country", "USA", "GeoNames", 0.99)
+    
+    return triples
+
+
+def get_rdf_star_demo_dataset_triples() -> list[dict]:
+    """RDF-Star Demo showing statement-level metadata."""
+    RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
+    EX = "http://example.org/"
+    
+    triples = []
+    
+    def add(s, p, o, source, confidence):
+        triples.append({
+            "subject": s,
+            "predicate": p,
+            "object": o,
+            "source": source,
+            "confidence": confidence
+        })
+    
+    # Employee data from HR System (high confidence)
+    add(f"{EX}alice", f"{EX}name", "Alice", "HR_System", 0.99)
+    add(f"{EX}alice", f"{EX}worksAt", f"{EX}acme", "HR_System", 0.98)
+    add(f"{EX}alice", f"{EX}department", "Engineering", "HR_System", 0.97)
+    add(f"{EX}alice", f"{EX}startDate", "2020-03-15", "HR_System", 0.99)
+    
+    # Salary from Payroll (very high confidence, sensitive)
+    add(f"{EX}alice", f"{EX}salary", "95000", "Payroll_System", 0.999)
+    
+    # Manager relationship
+    add(f"{EX}bob", f"{EX}name", "Bob", "HR_System", 0.99)
+    add(f"{EX}bob", f"{EX}worksAt", f"{EX}acme", "HR_System", 0.98)
+    add(f"{EX}bob", f"{EX}manages", f"{EX}alice", "HR_System", 0.95)
+    add(f"{EX}bob", f"{EX}title", "Engineering Manager", "HR_System", 0.97)
+    
+    # Company info from different sources
+    add(f"{EX}acme", f"{EX}name", "Acme Corporation", "SEC_Filings", 0.99)
+    add(f"{EX}acme", f"{EX}founded", "2010", "CrunchBase", 0.92)
+    add(f"{EX}acme", f"{EX}headquarters", "San Francisco", "CompanyWebsite", 0.95)
+    add(f"{EX}acme", f"{EX}employees", "500", "LinkedIn", 0.75)  # Less reliable
+    add(f"{EX}acme", f"{EX}employees", "487", "SEC_Filings", 0.98)  # Conflicting!
+    
+    # Performance reviews (varying confidence)
+    add(f"{EX}alice", f"{EX}performanceRating", "Exceeds Expectations", "Performance_System", 0.90)
+    add(f"{EX}alice", f"{EX}lastReview", "2024-12-01", "Performance_System", 0.99)
+    
+    # Skills from different sources
+    add(f"{EX}alice", f"{EX}skill", "Python", "LinkedIn", 0.85)
+    add(f"{EX}alice", f"{EX}skill", "Machine Learning", "LinkedIn", 0.80)
+    add(f"{EX}alice", f"{EX}skill", "Distributed Systems", "Manager_Assessment", 0.92)
+    
+    # Project assignments
+    add(f"{EX}project/alpha", f"{EX}name", "Project Alpha", "Jira", 0.99)
+    add(f"{EX}project/alpha", f"{EX}status", "Active", "Jira", 0.98)
+    add(f"{EX}project/alpha", f"{EX}lead", f"{EX}alice", "Jira", 0.97)
+    add(f"{EX}alice", f"{EX}assignedTo", f"{EX}project/alpha", "Jira", 0.96)
+    add(f"{EX}bob", f"{EX}sponsors", f"{EX}project/alpha", "Executive_Dashboard", 0.88)
+    
+    return triples
+
+
+# Keep old SPARQL functions for backwards compatibility (unused)
+def get_movies_dataset() -> str:
+    """Deprecated: Use get_movies_dataset_triples instead."""
+    return ""
+
+def get_techcorp_dataset() -> str:
+    """Deprecated: Use get_techcorp_dataset_triples instead."""
+    return ""
+
+def get_knowledge_graph_dataset() -> str:
+    """Deprecated: Use get_knowledge_graph_dataset_triples instead."""
+    return ""
+
+def get_rdf_star_demo_dataset() -> str:
+    """Deprecated: Use get_rdf_star_demo_dataset_triples instead."""
+    return ""