name: import-to-ontology description: Intelligently parse and move content from source markdown files to appropriate ontology locations with validation, enrichment, and web content integration

Import to Ontology Skill

Intelligently processes ~200 source markdown files and moves content blocks to the most suitable locations in the canonical ontology structure. Features semantic targeting, assertion validation, stub enrichment, and async web content integration.

Overview

This skill:

Parses large source markdown files into content blocks (batch mode)
Targets optimal ontology file locations using semantic index
Validates OWL2 compliance before and after content moves (ontology-core integration)
Handles image asset references (preserves paths to shared assets/ folder)
Updates assertions/claims in content
Enriches WikiLink stubs and isolated URLs with web summaries
Moves content destructively (removes from source)
Rolls back on validation failures
Cleans up empty source files automatically
Tracks progress across batches with resume capability

⚠️ DESTRUCTIVE OPERATION: Content is moved from source files, not copied. Source files are deleted when empty. NO BACKUPS ARE CREATED - ensure external backups exist before running.

Quick Start

# From project root
cd /home/devuser/workspace/project/Metaverse-Ontology

# ⚠️ IMPORTANT: Set up shared assets folder first
# Both source and target directories should reference the same assets/
mkdir -p assets/
mkdir -p /path/to/sources/assets/  # If not already present

# Dry run (analyze without moving)
claude-code "Use import-to-ontology skill with dry-run on source-file.md"

# Import single file (DESTRUCTIVE - moves content)
claude-code "Use import-to-ontology skill to process source-file.md"

# Import small batch (5 files at a time)
claude-code "Use import-to-ontology skill to process batch of 5 files from /sources/"

# Import all files (one at a time with progress tracking)
claude-code "Use import-to-ontology skill to process all files in /path/to/sources/"

⚠️ WARNING: This is a DESTRUCTIVE operation. Content is MOVED from source files and source files are DELETED when empty. NO BACKUPS ARE CREATED.

Ontology-Core Integration

This skill integrates with the ontology-core library for OWL2 validation:

Validation Bridge (Node.js ↔ Python)

// src/validation_bridge.js
const { validateOntologyFile } = require("./src/validation_bridge");

// Validates OWL2 compliance for target files
const result = await validateOntologyFile(targetFile);

if (!result.is_valid) {
  console.error(`Validation failed: ${result.errors.length} errors`);
  // Rollback content move
}

Validation Workflow

Pre-Move Validation: Check target file OWL2 compliance before modification
Content Migration: Insert content block into target file
Post-Move Validation: Re-validate target file with new content
Rollback on Failure: Restore source file if validation fails

OWL2 Checks (via ontology-core)

Class declarations and SubClassOf axioms
Property declarations (ObjectProperty, DataProperty)
Restrictions (ObjectSomeValuesFrom, etc.)
Namespace consistency (ai:, bc:, mv:, rb:, dt:)
Parentheses balance
Annotation format
Naming conventions

Rollback Strategy

// If post-move validation fails:
// 1. Restore source file from backup
// 2. Remove added content from target (if possible)
// 3. Log failure reason
// 4. Continue with next block

Skill Architecture

Phase 1: Analysis & Planning

Input: Source markdown file path(s) Output: Import plan with semantic targeting

interface ImportPlan {
  sourceFile: string;
  blocks: ContentBlock[];
  targets: TargetMapping[];
  enrichments: EnrichmentTask[];
  validations: ValidationTask[];
  estimatedTime: number;
}

interface ContentBlock {
  id: string;
  type: "heading" | "paragraph" | "list" | "code" | "quote";
  content: string;
  metadata: {
    keywords: string[];
    wikiLinks: string[];
    urls: string[];
    assertions: Assertion[];
  };
  startLine: number;
  endLine: number;
}

interface TargetMapping {
  blockId: string;
  targetFile: string;
  targetConcept: string;
  insertionPoint: "about" | "description" | "use-cases" | "examples" | "references";
  confidence: number;
  reasoning: string;
}

Phase 2: Content Processing

Actions:

Extract content blocks from source
Identify keywords, WikiLinks, URLs, assertions
Query semantic index for target concepts
Generate insertion plan with confidence scores

Phase 3: Enrichment (Async)

Web Content Integration:

interface EnrichmentTask {
  blockId: string;
  url: string;
  type: "stub-expansion" | "citation-enrichment" | "context-addition";
  priority: "high" | "medium" | "low";
  status: "pending" | "processing" | "completed" | "failed";
}

// Uses web-summary skill (async, ~3-10s per URL)
async function enrichUrl(url: string): Promise<EnrichedContent> {
  // Call web-summary skill
  const summary = await executeSkill("web-summary", { url });

  return {
    url,
    title: summary.title,
    summary: summary.summary,
    keyPoints: summary.keyPoints,
    relevantLogseqLinks: summary.semanticLinks, // Auto-generated [[WikiLinks]]
    citations: summary.citations,
  };
}

Phase 4: Validation & Updates

Assertion Validation:

interface Assertion {
  text: string;
  type: "claim" | "definition" | "statistic" | "example";
  needsValidation: boolean;
  updatedText?: string;
  confidence: number;
}

function validateAssertions(block: ContentBlock): ValidationResult {
  const assertions = extractAssertions(block.content);

  return assertions.map((assertion) => {
    // Check against current ontology knowledge
    const validation = checkAgainstOntology(assertion);

    // Suggest updates if outdated
    if (validation.outdated) {
      return {
        original: assertion.text,
        updated: validation.suggestedUpdate,
        reason: validation.reason,
        confidence: validation.confidence,
      };
    }

    return { valid: true, assertion };
  });
}

Phase 5: Content Migration with Validation

Safe Move Strategy with OWL2 Compliance:

interface MigrationResult {
  sourceFile: string;
  targetFile: string;
  blocksMoved: number;
  blocksEnriched: number;
  assertionsUpdated: number;
  backupPath: string;
  success: boolean;
  errors: Error[];
}

async function migrateContent(plan: ImportPlan): Promise<MigrationResult> {
  // 1. Create backups
  const backup = await createBackup(plan.sourceFile);

  // 2. Process each block with validation
  for (const block of plan.blocks) {
    const target = plan.targets.find((t) => t.blockId === block.id);

    // 3. Pre-move OWL2 validation
    const preValidation = await validateOntologyFile(target.targetFile);
    if (!preValidation.is_valid && !preValidation.new_file) {
      console.error(`Pre-move validation failed: ${target.targetFile}`);
      continue; // Skip this block
    }

    // 4. Enrich content (async web summaries)
    const enriched = await enrichBlock(block, plan.enrichments);

    // 5. Validate and update assertions
    const validated = await validateBlock(enriched);

    // 6. Insert into target file
    await insertContent(target.targetFile, target.insertionPoint, validated);

    // 7. Post-move OWL2 validation
    const postValidation = await validateOntologyFile(target.targetFile);
    if (!postValidation.is_valid) {
      console.error(`Post-move validation failed: ${target.targetFile}`);

      // Rollback: restore source file
      await restoreFromBackup(backup, plan.sourceFile);

      // Log validation errors
      postValidation.errors.forEach((err) => {
        console.error(`  Line ${err.line_number}: ${err.message}`);
      });

      continue; // Skip this block
    }

    // 8. Remove from source (only if validation passed)
    await removeBlockFromSource(plan.sourceFile, block);

    // 9. Log progress
    logProgress(block, target);
  }

  // 10. Remove source file if empty
  await archiveSourceFile(plan.sourceFile);

  return {
    sourceFile: plan.sourceFile,
    targetFile: target.targetFile,
    blocksMoved: plan.blocks.length,
    success: true,
  };
}

Semantic Targeting System

Uses the in-memory ontology index for intelligent placement:

// Load index once
const INDEX = JSON.parse(fs.readFileSync(".cache/ontology-index.json", "utf-8"));

function findTargetConcept(block: ContentBlock): TargetMapping {
  // 1. Extract semantic features
  const keywords = extractKeywords(block.content);
  const wikiLinks = extractWikiLinks(block.content);

  // 2. Score all concepts by relevance
  const scored = Object.values(INDEX.concepts.concepts)
    .map((concept) => {
      let score = 0;

      // Keyword overlap
      const keywordMatch = keywords.filter((k) =>
        concept.keywords.some((ck) => ck.includes(k) || k.includes(ck))
      ).length;
      score += keywordMatch * 0.4;

      // WikiLink overlap
      const linkMatch = wikiLinks.filter(
        (link) => concept.linksTo.includes(link) || concept.linkedFrom.includes(link)
      ).length;
      score += linkMatch * 0.6;

      return { concept, score };
    })
    .filter((s) => s.score > 0)
    .sort((a, b) => b.score - a.score);

  if (scored.length === 0) {
    // Fallback: use domain detection
    return detectDomainAndSuggest(block);
  }

  const best = scored[0];

  return {
    blockId: block.id,
    targetFile: best.concept.file,
    targetConcept: best.concept.preferredTerm,
    insertionPoint: selectInsertionPoint(block),
    confidence: Math.min(best.score, 0.95),
    reasoning: `Matched ${keywordMatch} keywords and ${linkMatch} links`,
  };
}

function selectInsertionPoint(block: ContentBlock): InsertionPoint {
  // Heuristics for where to insert content
  if (block.type === "heading" && block.content.includes("Definition")) {
    return "description";
  }
  if (block.type === "heading" && block.content.includes("Example")) {
    return "examples";
  }
  if (block.type === "heading" && block.content.includes("Use Case")) {
    return "use-cases";
  }
  if (block.content.includes("http") || block.content.includes("[[")) {
    return "references";
  }

  // Default: append to About section
  return "about";
}

WikiLink & URL Detection

interface DetectedStub {
  type: "wikilink" | "url";
  value: string;
  context: string; // Surrounding text
  line: number;
  enrichmentNeeded: boolean;
}

function detectStubs(content: string): DetectedStub[] {
  const stubs: DetectedStub[] = [];

  // 1. Find WikiLinks
  const wikilinkRegex = /\[\[([^\]]+)\]\]/g;
  let match;

  while ((match = wikilinkRegex.exec(content)) !== null) {
    const wikilink = match[1];

    // Check if it's a stub (broken link or minimal context)
    const isStub = !INDEX.wikilinks.valid[`[[${wikilink}]]`] || !hasContext(content, match.index);

    if (isStub) {
      stubs.push({
        type: "wikilink",
        value: wikilink,
        context: extractContext(content, match.index, 100),
        line: getLineNumber(content, match.index),
        enrichmentNeeded: true,
      });
    }
  }

  // 2. Find isolated URLs
  const urlRegex = /(https?:\/\/[^\s\)]+)/g;

  while ((match = urlRegex.exec(content)) !== null) {
    const url = match[1];

    // Check if URL has description nearby
    const hasDescription = hasContext(content, match.index, 50);

    if (!hasDescription) {
      stubs.push({
        type: "url",
        value: url,
        context: extractContext(content, match.index, 100),
        line: getLineNumber(content, match.index),
        enrichmentNeeded: true,
      });
    }
  }

  return stubs;
}

function hasContext(content: string, index: number, minLength = 30): boolean {
  // Check if there's meaningful text around the link/URL
  const before = content.substring(Math.max(0, index - 50), index);
  const after = content.substring(index, Math.min(content.length, index + 50));

  const contextText = before + after;
  const words = contextText.split(/\s+/).filter((w) => w.length > 3);

  return words.length >= 5 && contextText.length >= minLength;
}

Web Summary Integration

interface WebSummarySkillCall {
  url: string;
  options: {
    maxLength: number;
    includeSemanticLinks: boolean;
    format: "logseq" | "markdown";
  };
}

async function enrichStubsWithWebContent(stubs: DetectedStub[]): Promise<EnrichedStub[]> {
  const enriched: EnrichedStub[] = [];

  // Process in parallel with concurrency limit (5 at a time)
  const CONCURRENCY = 5;

  for (let i = 0; i < stubs.length; i += CONCURRENCY) {
    const batch = stubs.slice(i, i + CONCURRENCY);

    const results = await Promise.all(
      batch.map(async (stub) => {
        if (stub.type === "url") {
          try {
            // Call web-summary skill (async, 3-10s)
            console.log(`Enriching URL: ${stub.value}`);

            const summary = await executeSkill("web-summary", {
              url: stub.value,
              options: {
                maxLength: 300,
                includeSemanticLinks: true,
                format: "logseq",
              },
            });

            return {
              ...stub,
              enrichedContent: formatEnrichedContent(summary, stub),
              status: "completed",
            };
          } catch (error) {
            console.warn(`Failed to enrich ${stub.value}: ${error.message}`);
            return {
              ...stub,
              enrichedContent: null,
              status: "failed",
              error: error.message,
            };
          }
        } else if (stub.type === "wikilink") {
          // Check if we can create the concept
          const suggestion = suggestConceptCreation(stub.value);

          return {
            ...stub,
            enrichedContent: suggestion,
            status: "suggest-creation",
          };
        }

        return stub;
      })
    );

    enriched.push(...results);

    // Progress update
    console.log(`Enriched ${i + results.length}/${stubs.length} stubs`);
  }

  return enriched;
}

function formatEnrichedContent(summary: WebSummarySummary, stub: DetectedStub): string {
  // Format as Logseq-compatible content
  return `
- **Source**:
  - ${summary.summary}
  - **Key Points**:
${summary.keyPoints.map((pt) => `    - ${pt}`).join("\n")}
  - **Related Concepts**: ${summary.semanticLinks.join(", ")}
  - **Retrieved**: ${new Date().toISOString().split("T")[0]}
`.trim();
}

// Execute skill helper (calls Claude Code skill system)
async function executeSkill(skillName: string, params: any): Promise<any> {
  // This calls the web-summary skill from your skill database
  // Implementation depends on Claude Code skill execution API

  // For now, placeholder that would be replaced with actual skill call
  const result = await fetch(`/skill/${skillName}`, {
    method: "POST",
    body: JSON.stringify(params),
  });

  return result.json();
}

Content Block Parser

interface ParsedContent {
  blocks: ContentBlock[];
  metadata: {
    totalBlocks: number;
    totalLines: number;
    hasWikiLinks: boolean;
    hasUrls: boolean;
    estimatedAssertions: number;
  };
}

function parseSourceFile(filePath: string): ParsedContent {
  const content = fs.readFileSync(filePath, "utf-8");
  const lines = content.split("\n");

  const blocks: ContentBlock[] = [];
  let currentBlock: Partial<ContentBlock> | null = null;
  let blockId = 1;

  for (let i = 0; i < lines.length; i++) {
    const line = lines[i];

    // Detect block boundaries
    if (line.startsWith("#")) {
      // New heading - start new block
      if (currentBlock) {
        blocks.push(completeBlock(currentBlock));
      }

      currentBlock = {
        id: `block-${blockId++}`,
        type: "heading",
        content: line,
        startLine: i,
        metadata: {
          keywords: [],
          wikiLinks: [],
          urls: [],
          assertions: [],
        },
      };
    } else if (line.startsWith("```")) {
      // Code block
      if (currentBlock) {
        blocks.push(completeBlock(currentBlock));
      }

      // Find end of code block
      let endLine = i + 1;
      while (endLine < lines.length && !lines[endLine].startsWith("```")) {
        endLine++;
      }

      currentBlock = {
        id: `block-${blockId++}`,
        type: "code",
        content: lines.slice(i, endLine + 1).join("\n"),
        startLine: i,
        endLine: endLine,
      };

      i = endLine; // Skip to end
    } else if (currentBlock) {
      // Continuation of current block
      currentBlock.content += "\n" + line;
    } else if (line.trim()) {
      // Start new paragraph block
      currentBlock = {
        id: `block-${blockId++}`,
        type: "paragraph",
        content: line,
        startLine: i,
      };
    }
  }

  // Complete final block
  if (currentBlock) {
    blocks.push(completeBlock(currentBlock));
  }

  // Extract metadata for each block
  blocks.forEach((block) => {
    block.metadata = {
      keywords: extractKeywords(block.content),
      wikiLinks: extractWikiLinks(block.content),
      urls: extractUrls(block.content),
      assertions: extractAssertions(block.content),
    };
  });

  return {
    blocks,
    metadata: {
      totalBlocks: blocks.length,
      totalLines: lines.length,
      hasWikiLinks: blocks.some((b) => b.metadata.wikiLinks.length > 0),
      hasUrls: blocks.some((b) => b.metadata.urls.length > 0),
      estimatedAssertions: blocks.reduce((sum, b) => sum + b.metadata.assertions.length, 0),
    },
  };
}

function extractUrls(content: string): string[] {
  const urlRegex = /(https?:\/\/[^\s\)]+)/g;
  const matches = content.match(urlRegex);
  return matches ? [...new Set(matches)] : [];
}

function extractAssertions(content: string): Assertion[] {
  const assertions: Assertion[] = [];

  // Patterns that indicate assertions
  const patterns = [
    /is defined as (.+?)\./gi,
    /refers to (.+?)\./gi,
    /(\d+%|\d+ percent)/gi,
    /according to (.+?),/gi,
    /enables (.+?)\./gi,
    /provides (.+?)\./gi,
  ];

  for (const pattern of patterns) {
    let match;
    while ((match = pattern.exec(content)) !== null) {
      assertions.push({
        text: match[0],
        type: determineAssertionType(match[0]),
        needsValidation: true,
        confidence: 0.7,
      });
    }
  }

  return assertions;
}

function determineAssertionType(text: string): Assertion["type"] {
  if (text.includes("%") || text.includes("percent")) return "statistic";
  if (text.includes("defined as") || text.includes("refers to")) return "definition";
  if (text.includes("for example") || text.includes("such as")) return "example";
  return "claim";
}

Dry Run Mode

interface DryRunReport {
  sourceFile: string;
  analysisDate: string;
  summary: {
    totalBlocks: number;
    targetedFiles: string[];
    urlsToEnrich: number;
    assertionsToValidate: number;
    estimatedTimeSeconds: number;
  };
  plan: ImportPlan;
  warnings: string[];
  recommendations: string[];
}

function dryRun(sourceFile: string): DryRunReport {
  const parsed = parseSourceFile(sourceFile);
  const plan = createImportPlan(parsed);

  const urlCount = parsed.blocks.reduce((sum, b) => sum + b.metadata.urls.length, 0);

  const estimatedTime =
    parsed.blocks.length * 2 + // 2s per block
    urlCount * 5; // 5s per URL (web-summary avg)

  return {
    sourceFile,
    analysisDate: new Date().toISOString(),
    summary: {
      totalBlocks: parsed.blocks.length,
      targetedFiles: [...new Set(plan.targets.map((t) => t.targetFile))],
      urlsToEnrich: urlCount,
      assertionsToValidate: parsed.metadata.estimatedAssertions,
      estimatedTimeSeconds: estimatedTime,
    },
    plan,
    warnings: generateWarnings(plan),
    recommendations: generateRecommendations(plan),
  };
}

function generateWarnings(plan: ImportPlan): string[] {
  const warnings: string[] = [];

  // Low confidence targets
  const lowConfidence = plan.targets.filter((t) => t.confidence < 0.5);
  if (lowConfidence.length > 0) {
    warnings.push(`${lowConfidence.length} blocks have low confidence targeting (<50%)`);
  }

  // Many URLs to enrich
  if (plan.enrichments.length > 20) {
    warnings.push(`${plan.enrichments.length} URLs to enrich - this will be slow (~${plan.enrichments.length * 5}s)`);
  }

  return warnings;
}

function generateRecommendations(plan: ImportPlan): string[] {
  const recs: string[] = [];

  // Suggest manual review for low confidence
  const veryLowConf = plan.targets.filter((t) => t.confidence < 0.3);
  if (veryLowConf.length > 0) {
    recs.push(`Recommend manual review for ${veryLowConf.length} low-confidence blocks`);
  }

  // Suggest creating missing concepts
  const missingConcepts = new Set(
    plan.blocks.flatMap((b) => b.metadata.wikiLinks).filter((link) => !INDEX.wikilinks.valid[`[[${link}]]`])
  );

  if (missingConcepts.size > 0) {
    recs.push(
      `Consider creating ${missingConcepts.size} missing concepts: ${Array.from(missingConcepts).slice(0, 5).join(", ")}...`
    );
  }

  return recs;
}

Progress Tracking

interface ImportProgress {
  sessionId: string;
  startTime: string;
  currentFile: string;
  filesProcessed: number;
  totalFiles: number;
  blocksProcessed: number;
  totalBlocks: number;
  urlsEnriched: number;
  totalUrls: number;
  assertionsValidated: number;
  errors: ImportError[];
  estimatedTimeRemaining: number;
}

class ImportTracker {
  private progress: ImportProgress;
  private logFile: string;

  constructor(totalFiles: number) {
    this.progress = {
      sessionId: generateSessionId(),
      startTime: new Date().toISOString(),
      currentFile: "",
      filesProcessed: 0,
      totalFiles,
      blocksProcessed: 0,
      totalBlocks: 0,
      urlsEnriched: 0,
      totalUrls: 0,
      assertionsValidated: 0,
      errors: [],
      estimatedTimeRemaining: 0,
    };

    this.logFile = `/tmp/import-ontology-${this.progress.sessionId}.log`;
  }

  updateProgress(update: Partial<ImportProgress>) {
    Object.assign(this.progress, update);

    // Calculate estimated time remaining
    const elapsed = Date.now() - new Date(this.progress.startTime).getTime();
    const rate = this.progress.blocksProcessed / (elapsed / 1000);
    const remaining = this.progress.totalBlocks - this.progress.blocksProcessed;
    this.progress.estimatedTimeRemaining = remaining / rate;

    // Log to file
    this.log(JSON.stringify(this.progress, null, 2));

    // Console update
    this.printProgress();
  }

  log(message: string) {
    fs.appendFileSync(this.logFile, `${new Date().toISOString()} - ${message}\n`);
  }

  printProgress() {
    const pct = ((this.progress.blocksProcessed / this.progress.totalBlocks) * 100).toFixed(1);
    const eta = Math.ceil(this.progress.estimatedTimeRemaining / 60);

    console.log(
      `
📊 Import Progress: ${pct}%
   Files: ${this.progress.filesProcessed}/${this.progress.totalFiles}
   Blocks: ${this.progress.blocksProcessed}/${this.progress.totalBlocks}
   URLs Enriched: ${this.progress.urlsEnriched}/${this.progress.totalUrls}
   ETA: ${eta} minutes
   Errors: ${this.progress.errors.length}
    `.trim()
    );
  }

  addError(error: ImportError) {
    this.progress.errors.push(error);
    this.log(`ERROR: ${JSON.stringify(error)}`);
  }

  getReport(): ImportProgress {
    return { ...this.progress };
  }
}

Validation Bridge Usage

Standalone Validation

# Validate single file
node src/validation_bridge.js /path/to/ontology-file.md

# Batch validate multiple files
node src/validation_bridge.js file1.md file2.md file3.md

# From import-engine
const { validateOntologyFile } = require('./src/validation_bridge');

async function checkFile(filePath) {
  try {
    const result = await validateOntologyFile(filePath);

    if (result.is_valid) {
      console.log(`✅ Valid: ${result.total_axioms} axioms`);
    } else {
      console.log(`❌ Invalid: ${result.errors.length} errors`);
      result.errors.forEach(err => {
        console.log(`  Line ${err.line_number}: ${err.message}`);
        if (err.fix_suggestion) {
          console.log(`    Fix: ${err.fix_suggestion}`);
        }
      });
    }
  } catch (error) {
    console.error(`Validation failed: ${error.message}`);
  }
}

Integration Pattern

// Before destructive move
const preValidation = await validateOntologyFile(targetFile);
if (!preValidation.is_valid) {
  console.error(`Target has errors - aborting move`);
  return { success: false, reason: "pre-validation-failed" };
}

// Perform move
moveContentBlock(source, target);

// After move - re-validate
const postValidation = await validateOntologyFile(targetFile);
if (!postValidation.is_valid) {
  // Rollback
  restoreFromBackup(source, target);
  return { success: false, reason: "post-validation-failed" };
}

Usage Examples

Example 1: Import Single File with Dry Run and Validation

# Analyze before importing
claude-code "Use import-to-ontology skill to dry-run source-notes.md"

# Review report, then import
claude-code "Use import-to-ontology skill to import source-notes.md"

Example 2: Batch Import Directory

claude-code "Use import-to-ontology skill to import all files from /sources/research-notes/"

Example 3: Import with Manual Target Override

claude-code "Use import-to-ontology skill to import blockchain-notes.md targeting BC-0001-blockchain.md"

Configuration

Create .import-ontology.config.json in project root:

{
  "sourceDirectory": "/home/devuser/workspace/project/sources",
  "ontologyDirectory": "/home/devuser/workspace/project/Metaverse-Ontology/logseq/pages",
  "backupDirectory": "/home/devuser/workspace/project/.backups",
  "indexPath": ".cache/ontology-index.json",

  "webSummary": {
    "enabled": true,
    "concurrency": 5,
    "timeout": 10000,
    "maxLength": 300
  },

  "validation": {
    "enabled": true,
    "autoFix": false,
    "requireManualReview": true,
    "owl2Compliance": true,
    "rollbackOnFailure": true
  },

  "targeting": {
    "minConfidence": 0.4,
    "requireManualReview": 0.7
  },

  "safety": {
    "createBackups": true,
    "dryRunFirst": true,
    "archiveSourceAfterImport": true
  }
}

import-to-ontology

$ Instalar