new_sdk_wu/www/new_sdk/.understand-anything/.trash-20260630093035/tmp/ua-graph-validate.js

const fs = require('fs');
const path = require('path');

const VALID_NODE_TYPES = new Set([
  'file','function','class','module','concept','config','document',
  'service','table','endpoint','pipeline','schema','resource','domain','flow','step'
]);

const VALID_EDGE_TYPES = new Set([
  'imports','exports','contains','inherits','implements','calls','subscribes',
  'publishes','middleware','reads_from','writes_to','transforms','validates',
  'depends_on','tested_by','configures','related','similar_to','deploys','serves',
  'migrates','documents','provisions','routes','defines_schema','triggers',
  'contains_flow','flow_step','cross_domain'
]);

const VALID_COMPLEXITY = new Set(['simple','moderate','complex']);
const VALID_DIRECTION = new Set(['forward','backward','bidirectional']);

const FILE_LEVEL_TYPES = new Set(['file','config','document','service','pipeline','table','schema','resource','endpoint']);

const graphPath = process.argv[2];
const outPath = process.argv[3];

const issues = [];
const warnings = [];

function addIssue(msg) { issues.push(msg); }
function addWarning(msg) { warnings.push(msg); }

try {
  const raw = fs.readFileSync(graphPath, 'utf8');
  const graph = JSON.parse(raw);

  const nodes = graph.nodes || [];
  const edges = graph.edges || [];
  const layers = graph.layers || [];
  const tour = graph.tour || [];

  // Build lookup sets
  const nodeIds = new Set(nodes.map(n => n.id));
  const nodeIdsInLayers = new Set();
  const nodeIdsInTour = new Set();

  // Stats
  const nodeTypeCounts = {};
  const edgeTypeCounts = {};
  for (const n of nodes) { nodeTypeCounts[n.type] = (nodeTypeCounts[n.type] || 0) + 1; }
  for (const e of edges) { edgeTypeCounts[e.type] = (edgeTypeCounts[e.type] || 0) + 1; }

  // ========== Check 1: Schema Validation ==========

  // Nodes
  for (const n of nodes) {
    if (!n.id || typeof n.id !== 'string' || n.id.trim() === '') addIssue(`Node missing/empty id: ${JSON.stringify(n)}`);
    if (!n.type || !VALID_NODE_TYPES.has(n.type)) addIssue(`Node "${n.id}" has invalid type "${n.type}"`);
    if (!n.name || typeof n.name !== 'string' || n.name.trim() === '') addIssue(`Node "${n.id}" missing/empty name`);
    if (!n.summary || typeof n.summary !== 'string' || n.summary.trim() === '') addIssue(`Node "${n.id}" missing/empty summary`);
    if (n.summary && n.name && n.summary.trim() === n.name.trim()) addWarning(`Node "${n.id}" summary equals name`);
    if (!Array.isArray(n.tags) || n.tags.length === 0) addIssue(`Node "${n.id}" missing/empty tags`);
    else {
      for (const t of n.tags) {
        if (typeof t !== 'string') addIssue(`Node "${n.id}" has non-string tag`);
      }
    }
    if (!n.complexity || !VALID_COMPLEXITY.has(n.complexity)) addIssue(`Node "${n.id}" has invalid complexity "${n.complexity}"`);

    // ID prefix check (Check 9)
    if (n.id && n.type) {
      const prefix = n.id.split(':')[0];
      if (prefix !== n.type) addWarning(`Node "${n.id}" type "${n.type}" does not match ID prefix "${prefix}"`);
    }
  }

  // Edges
  for (let i = 0; i < edges.length; i++) {
    const e = edges[i];
    if (!e.source || typeof e.source !== 'string') addIssue(`Edge[${i}] missing/empty source`);
    if (!e.target || typeof e.target !== 'string') addIssue(`Edge[${i}] missing/empty target`);
    if (!e.type || !VALID_EDGE_TYPES.has(e.type)) addIssue(`Edge[${i}] (${e.source} -> ${e.target}) has invalid type "${e.type}"`);
    if (!e.direction || !VALID_DIRECTION.has(e.direction)) addIssue(`Edge[${i}] (${e.source} -> ${e.target}) has invalid direction "${e.direction}"`);
    if (typeof e.weight !== 'number' || e.weight < 0 || e.weight > 1) addIssue(`Edge[${i}] (${e.source} -> ${e.target}) has invalid weight "${e.weight}"`);
  }

  // ========== Check 2: Referential Integrity ==========

  for (let i = 0; i < edges.length; i++) {
    const e = edges[i];
    if (e.source && !nodeIds.has(e.source)) addIssue(`Edge[${i}] source "${e.source}" references non-existent node`);
    if (e.target && !nodeIds.has(e.target)) addIssue(`Edge[${i}] target "${e.target}" references non-existent node`);
  }

  for (const layer of layers) {
    if (Array.isArray(layer.nodeIds)) {
      for (const nid of layer.nodeIds) {
        nodeIdsInLayers.add(nid);
        if (!nodeIds.has(nid)) addIssue(`Layer "${layer.id}" nodeIds references non-existent node "${nid}"`);
      }
    }
  }

  for (const step of tour) {
    if (Array.isArray(step.nodeIds)) {
      for (const nid of step.nodeIds) {
        nodeIdsInTour.add(nid);
        if (!nodeIds.has(nid)) addIssue(`Tour step ${step.order} references non-existent node "${nid}"`);
      }
    }
  }

  // ========== Check 3: Completeness ==========

  if (nodes.length < 1) addIssue('No nodes in graph');
  if (edges.length < 1) addIssue('No edges in graph');
  if (layers.length < 1) addIssue('No layers in graph');
  if (tour.length < 1) addIssue('No tour steps in graph');

  // ========== Check 4: Layer Coverage ==========

  const fileLevelNodes = nodes.filter(n => FILE_LEVEL_TYPES.has(n.type));
  const fileLevelNodeIds = new Set(fileLevelNodes.map(n => n.id));
  const coveredInLayers = new Set();

  for (const layer of layers) {
    if (!Array.isArray(layer.nodeIds) || layer.nodeIds.length === 0) {
      addIssue(`Layer "${layer.id}" has empty nodeIds`);
    }
    for (const nid of (layer.nodeIds || [])) {
      if (fileLevelNodeIds.has(nid)) {
        if (coveredInLayers.has(nid)) {
          addIssue(`File-level node "${nid}" appears in multiple layers`);
        } else {
          coveredInLayers.add(nid);
        }
      }
    }
  }

  for (const fnid of fileLevelNodeIds) {
    if (!coveredInLayers.has(fnid)) {
      addIssue(`File-level node "${fnid}" is not in any layer`);
    }
  }

  // ========== Check 5: Uniqueness ==========

  const seenIds = new Set();
  for (const n of nodes) {
    if (seenIds.has(n.id)) addIssue(`Duplicate node ID: "${n.id}"`);
    seenIds.add(n.id);
  }

  // ========== Check 6: Tour Validation ==========

  if (tour.length < 5 || tour.length > 15) {
    addWarning(`Tour has ${tour.length} steps (expected 5-15)`);
  }
  for (let i = 0; i < tour.length; i++) {
    const step = tour[i];
    if (step.order !== i + 1) addWarning(`Tour step index ${i} has order ${step.order}, expected ${i + 1}`);
    if (!Array.isArray(step.nodeIds) || step.nodeIds.length === 0) addWarning(`Tour step ${step.order} has no nodeIds`);
  }

  // ========== Check 7: Quality Checks ==========

  // Build edge connectivity set
  const connectedNodes = new Set();
  for (const e of edges) {
    connectedNodes.add(e.source);
    connectedNodes.add(e.target);
  }

  for (const n of nodes) {
    if (!connectedNodes.has(n.id)) addWarning(`Orphan node: "${n.id}" has no edges`);
  }

  for (let i = 0; i < edges.length; i++) {
    const e = edges[i];
    if (e.source === e.target) addWarning(`Self-referencing edge[${i}]: "${e.source}"`);
  }

  // ========== Check 8: Non-Code Node Quality ==========

  // Build edge index by source and target
  const edgesBySource = {};
  const edgesByTarget = {};
  for (const e of edges) {
    if (!edgesBySource[e.source]) edgesBySource[e.source] = [];
    edgesBySource[e.source].push(e);
    if (!edgesByTarget[e.target]) edgesByTarget[e.target] = [];
    edgesByTarget[e.target].push(e);
  }

  function nodeHasEdgeType(nodeId, edgeType) {
    const outEdges = edgesBySource[nodeId] || [];
    const inEdges = edgesByTarget[nodeId] || [];
    return outEdges.some(e => e.type === edgeType) || inEdges.some(e => e.type === edgeType);
  }

  for (const n of nodes) {
    const allEdges = [...(edgesBySource[n.id] || []), ...(edgesByTarget[n.id] || [])];
    if (n.type === 'document' && !nodeHasEdgeType(n.id, 'documents')) {
      addWarning(`Document node "${n.id}" has no "documents" edge`);
    }
    if (n.type === 'service' && !(nodeHasEdgeType(n.id, 'deploys') || nodeHasEdgeType(n.id, 'depends_on'))) {
      addWarning(`Service node "${n.id}" has no "deploys" or "depends_on" edge`);
    }
    if (n.type === 'pipeline' && !nodeHasEdgeType(n.id, 'triggers')) {
      addWarning(`Pipeline node "${n.id}" has no "triggers" edge`);
    }
    if (n.type === 'table' && !(nodeHasEdgeType(n.id, 'migrates') || nodeHasEdgeType(n.id, 'defines_schema'))) {
      addWarning(`Table node "${n.id}" has no "migrates" or "defines_schema" edge`);
    }
    if (n.type === 'schema' && !nodeHasEdgeType(n.id, 'defines_schema')) {
      addWarning(`Schema node "${n.id}" has no "defines_schema" edge`);
    }
  }

  // ========== Output ==========

  const result = {
    scriptCompleted: true,
    issues,
    warnings,
    stats: {
      totalNodes: nodes.length,
      totalEdges: edges.length,
      totalLayers: layers.length,
      tourSteps: tour.length,
      nodeTypes: nodeTypeCounts,
      edgeTypes: edgeTypeCounts
    }
  };

  fs.writeFileSync(outPath, JSON.stringify(result, null, 2), 'utf8');
  console.log(JSON.stringify(result, null, 2));
  process.exit(0);

} catch (err) {
  console.error('Script crash:', err.message);
  process.exit(1);
}