fast-schema-traverse

High-performance JSON Schema traversal library - 2x faster drop-in replacement for json-schema-traverse.

Features

• 2-2.5x faster than json-schema-traverse
• 50-70% less memory usage
• 100% API compatible - drop-in replacement
• Iterative traversal (no recursion limits)
• Optimized keyword lookup with Map
• Object pooling for reduced GC pressure
• String caching for JSON Pointers

Installation

npm install fast-schema-traverse

Usage

const traverse = require('fast-schema-traverse');

const schema = {
  type: 'object',
  properties: {
    foo: { type: 'string' },
    bar: { type: 'number' }
  }
};

// Simple callback
traverse(schema, (schema, jsonPtr, rootSchema, parentJsonPtr, parentKeyword, parentSchema, keyIndex) => {
  console.log('Schema at', jsonPtr);
});

// With options
traverse(schema, {
  cb: (schema) => console.log(schema),
  allKeys: false  // Only traverse schema keywords (default)
});

// Pre/post order traversal
traverse(schema, {
  cb: {
    pre: (schema, jsonPtr) => console.log('Entering:', jsonPtr),
    post: (schema, jsonPtr) => console.log('Exiting:', jsonPtr)
  }
});

API

traverse(schema, options, callback?)

• schema - JSON Schema to traverse
• options - Object or callback function
  • cb - Callback function or {pre, post} object
  • allKeys - Traverse all object properties (default: false)
• callback - Optional callback (for legacy API)

Callback Parameters

All callbacks receive 7 parameters:

1. schema - Current schema object
2. jsonPtr - JSON Pointer path (e.g., "/properties/foo")
3. rootSchema - Root schema reference
4. parentJsonPtr - Parent's JSON Pointer
5. parentKeyword - Parent keyword (e.g., "properties")
6. parentSchema - Parent schema object
7. keyIndex - Property name or array index

Performance

Benchmarked against json-schema-traverse v1.0.0:

| Schema Type | Speedup | Memory Reduction | |-------------|---------|------------------| | Deep (20+ levels) | 2.5x | 70% | | Wide (100+ properties) | 2.0x | 50% | | Mixed (real-world) | 2.2x | 60% |

Optimizations

• Iterative traversal - Eliminates recursion overhead and call stack limits
• Fast keyword lookup - Map-based O(1) keyword classification
• Object pooling - Reuses stack frame objects to reduce GC pressure
• String caching - Caches escaped JSON Pointer segments

Compatibility

100% compatible with json-schema-traverse v1.0.0. Supports:

• All JSON Schema keywords (draft-2019-09 and earlier)
• Legacy v0.3.1 API
• Pre/post order traversal
• allKeys option for custom keywords
• Static properties: keywords, arrayKeywords, propsKeywords, skipKeywords

License

MIT

Credits

API design and test suite from json-schema-traverse by Evgeny Poberezkin.