glam/frontend/src/hooks/useGraphData.ts

/**
 * React Hook for Graph Data Management
 * Processes parsed RDF data for D3.js visualization
 */

import { useState, useCallback, useMemo } from 'react';
import {
  extractNodeTypes,
  extractPredicates,
  getLinkCounts,
  getNodeCounts,
} from '@/lib/rdf/graph-utils';
import type { GraphData, GraphNode, GraphLink, SemanticCategory } from '@/types/rdf';
import { getSemanticCategory } from '@/types/rdf';

export interface GraphFilters {
  nodeTypes: Set<string>;
  predicates: Set<string>;
  searchTerm: string;
  /** Hide literal nodes (strings, numbers, dates) from the graph */
  hideLiterals: boolean;
  /** Semantic categories to show (empty = show all) */
  semanticCategories: Set<SemanticCategory>;
}

interface UseGraphDataReturn {
  // Data
  nodes: GraphNode[];
  links: GraphLink[];
  filteredNodes: GraphNode[];
  filteredLinks: GraphLink[];
  
  // Metadata
  nodeTypes: string[];
  predicates: string[];
  linkCounts: Record<string, number>;
  nodeCounts: Record<string, number>;
  
  // Filters
  filters: GraphFilters;
  setFilters: (filters: Partial<GraphFilters>) => void;
  resetFilters: () => void;
  
  // Selection
  selectedNode: GraphNode | null;
  setSelectedNode: (node: GraphNode | null) => void;
  
  // Data loading
  loadGraphData: (data: GraphData) => void;
  clearGraphData: () => void;
  
  // Stats
  stats: {
    totalNodes: number;
    totalLinks: number;
    filteredNodeCount: number;
    filteredLinkCount: number;
  };
}

const DEFAULT_FILTERS: GraphFilters = {
  nodeTypes: new Set(),
  predicates: new Set(),
  searchTerm: '',
  hideLiterals: true, // Default to hiding literals for cleaner graphs
  semanticCategories: new Set(), // Empty = show all categories
};

export function useGraphData(): UseGraphDataReturn {
  const [graphData, setGraphData] = useState<GraphData | null>(null);
  const [filters, setFiltersState] = useState<GraphFilters>(DEFAULT_FILTERS);
  const [selectedNode, setSelectedNode] = useState<GraphNode | null>(null);

  // Extract metadata from graph data
  const nodeTypes = useMemo(
    () => (graphData ? extractNodeTypes(graphData) : []),
    [graphData]
  );

  const predicates = useMemo(
    () => (graphData ? extractPredicates(graphData) : []),
    [graphData]
  );

  const linkCounts = useMemo(
    () => (graphData ? getLinkCounts(graphData) : {}),
    [graphData]
  );

  const nodeCounts = useMemo(
    () => (graphData ? getNodeCounts(graphData) : {}),
    [graphData]
  );

  // Get raw nodes and links
  const nodes = graphData?.nodes || [];
  const links = graphData?.links || [];

  // Apply filters to nodes
  const filteredNodes = useMemo(() => {
    if (!graphData) return [];

    let result = nodes;

    // Filter by semantic categories (if any selected)
    if (filters.semanticCategories.size > 0) {
      result = result.filter((node) => {
        const category = getSemanticCategory(node.type);
        return filters.semanticCategories.has(category);
      });
    }

    // Filter out literals if hideLiterals is enabled
    if (filters.hideLiterals) {
      result = result.filter((node) => {
        const category = getSemanticCategory(node.type);
        return category !== 'literals';
      });
    }

    // Filter by node type
    if (filters.nodeTypes.size > 0) {
      result = result.filter((node) => filters.nodeTypes.has(node.type));
    }

    // Filter by search term
    if (filters.searchTerm) {
      const term = filters.searchTerm.toLowerCase();
      result = result.filter(
        (node) =>
          node.id.toLowerCase().includes(term) ||
          node.label?.toLowerCase().includes(term)
      );
    }

    return result;
  }, [graphData, nodes, filters]);

  // Apply filters to links and expand to include connected nodes
  const { filteredLinks, expandedNodes } = useMemo(() => {
    if (!graphData) return { filteredLinks: [], expandedNodes: [] as GraphNode[] };

    // Start with the filtered nodes based on type/search
    const primaryNodeIds = new Set(filteredNodes.map((n) => n.id));
    
    // Find links where at least ONE end is a primary node
    // This ensures we see connections FROM the filtered nodes
    let relevantLinks = links.filter((link) => {
      const sourceId = typeof link.source === 'string' ? link.source : link.source.id;
      const targetId = typeof link.target === 'string' ? link.target : link.target.id;
      return primaryNodeIds.has(sourceId) || primaryNodeIds.has(targetId);
    });

    // Filter by predicate if specified
    if (filters.predicates.size > 0) {
      relevantLinks = relevantLinks.filter((link) => filters.predicates.has(link.predicate));
    }
    
    // Collect all nodes involved in the links
    const expandedNodeIds = new Set(primaryNodeIds);
    for (const link of relevantLinks) {
      const sourceId = typeof link.source === 'string' ? link.source : link.source.id;
      const targetId = typeof link.target === 'string' ? link.target : link.target.id;
      expandedNodeIds.add(sourceId);
      expandedNodeIds.add(targetId);
    }
    
    // Return expanded set of nodes (primary + connected)
    const expanded = nodes.filter((n) => expandedNodeIds.has(n.id));
    
    return { filteredLinks: relevantLinks, expandedNodes: expanded };
  }, [graphData, nodes, links, filteredNodes, filters.predicates]);

  // Use expandedNodes instead of filteredNodes for display
  const displayNodes = expandedNodes.length > 0 ? expandedNodes : filteredNodes;

  // Load graph data
  const loadGraphData = useCallback((data: GraphData) => {
    setGraphData(data);
    setSelectedNode(null);
    // Reset filters when loading new data
    setFiltersState(DEFAULT_FILTERS);
  }, []);

  // Clear graph data
  const clearGraphData = useCallback(() => {
    setGraphData(null);
    setSelectedNode(null);
    setFiltersState(DEFAULT_FILTERS);
  }, []);

  // Update filters
  const setFilters = useCallback((newFilters: Partial<GraphFilters>) => {
    setFiltersState((prev) => ({
      ...prev,
      ...newFilters,
    }));
  }, []);

  // Reset filters
  const resetFilters = useCallback(() => {
    setFiltersState(DEFAULT_FILTERS);
  }, []);

  // Stats
  const stats = useMemo(
    () => ({
      totalNodes: nodes.length,
      totalLinks: links.length,
      filteredNodeCount: displayNodes.length,
      filteredLinkCount: filteredLinks.length,
    }),
    [nodes.length, links.length, displayNodes.length, filteredLinks.length]
  );

  return {
    nodes,
    links,
    filteredNodes: displayNodes,
    filteredLinks,
    nodeTypes,
    predicates,
    linkCounts,
    nodeCounts,
    filters,
    setFilters,
    resetFilters,
    selectedNode,
    setSelectedNode,
    loadGraphData,
    clearGraphData,
    stats,
  };
}