json-to-schema

json-to-schema is a Python tool that infers a JSON Schema from JSON data. Use it to generate JSON Schema definitions quickly from real payloads. It outputs schemas compatible with JSON Schema draft 2020-12.

Features

• Infers types for objects, arrays, and scalars.
• Merges array item types across elements.
• Marks object properties as required based on the input instance.
• Optionally infers bounds for selected fields with --infer-bounds.
• Optionally infers enum values for selected fields with --infer-enum.
• Supports global inference with --infer-all-bounds and --infer-all-enum.
• Sets root schema metadata with --schema-id, --schema-title, and --schema-description.
• Sets field-level metadata with --field-title and --field-description.
• Validates input JSON against a provided schema file with --validate.
• Emits schemas compliant with JSON Schema draft 2020-12.

Requirements

• Python 3.8+

Install

pip install json-to-schema

Generate JSON Schema from a JSON file (CLI)

Provide an input JSON file (defaults to file.json) and print the inferred schema to stdout:

json-to-schema

Specify a custom input file and write output to a schema file:

json-to-schema -i input.json -o schema.json

Print compact/minified output:

json-to-schema -i input.json --minify

Infer bounds:

json-to-schema -i input.json --infer-bounds age price tags

Infer enum values:

json-to-schema -i input.json --infer-enum status type

Infer bounds for all applicable fields:

json-to-schema -i input.json --infer-all-bounds

Infer enum values for all applicable fields:

json-to-schema -i input.json --infer-all-enum

Set root schema metadata:

json-to-schema -i input.json \
  --schema-id https://example.com/schemas/product \
  --schema-title "Product Schema" \
  --schema-description "Schema for product payloads"

Set field-level metadata (dot paths, with [] for array items):

json-to-schema -i input.json \
  --field-title user.name="Full name" \
  --field-description tags[]="Tag value"

Validate a payload file against an existing schema:

json-to-schema -i payload.json --validate schema.json

Validate piped stdin against a schema:

echo '{"name":"Widget","price":12.5}' | json-to-schema --validate schema.json

You can also pipe JSON directly into stdin:

echo '{"name":"Widget","price":12.5}' | json-to-schema

Infer JSON Schema in Python code (library usage)

You can also import and use this package directly in Python applications:

import json
from json_to_schema import infer_schema, SCHEMA_DRAFT

data = {
  "name": "Widget",
  "price": 12.5,
  "tags": ["sale", "featured"],
  "in_stock": True
}

schema = {
  "$schema": SCHEMA_DRAFT,
  **infer_schema(data)
}

print(json.dumps(schema, indent=2))

Example: convert sample JSON to schema

Input (input.json):

{
  "name": "Widget",
  "price": 12.5,
  "tags": ["sale", "featured"],
  "in_stock": true
}

Run:

json-to-schema -i input.json -o schema.json

Output (schema.json):

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "type": "object",
  "properties": {
    "name": { "type": "string" },
    "price": { "type": "number" },
    "tags": {
      "type": "array",
      "items": { "type": "string" }
    },
    "in_stock": { "type": "boolean" }
  },
  "required": ["in_stock", "name", "price", "tags"],
  "additionalProperties": false
}

Testing

python -m unittest discover -s tests