Skip to content

Latest commit

 

History

History
899 lines (646 loc) · 27.8 KB

README.md

File metadata and controls

899 lines (646 loc) · 27.8 KB
Raw

json-schema-library

Customizable and hackable json-validator and json-schema utilities for traversal, data generation and validation

This package offers tools and utilities to work with json-schema, create and validate data. Unfortunately, most packages, editors or validators do not care to expose basic json-schema functionality. Instead of small memory footprint or high performance, this package focuses on exposing utilities for browser and node environments and lessens the pain to build custom tools around json-schema.

draft methods | draft extensions | draft customization | breaking changes

install

yarn add json-schema-library

Npm package version CI

This library currently supports draft4, draft6 and draft7 features @see benchmark

usage

json-schema-library exposes a Draft for most json-schema drafts. Each Draft can be instantiated and offers a common set of actions working on the specified json-schema version. Example:

import { Draft04, Draft06, Draft07, Draft, JSONError } from "json-schema-library";
import myJsonSchema from "./myJsonSchema.json";
import myData from "./myData.json";

const jsonSchema: Draft = new Draft07(myJsonSchema);
const errors: JSONError[] = jsonSchema.validate(myData);

What follows is a description of the main draft methods.

Draft methods

validate

validate is a complete json-schema validator for your input data. Calling validate will return a list of validation errors for the passed data.

const jsonSchema = new Draft07(myJsonSchema);
const errors: JSONError[] = jsonSchema.validate(myData);

Additionally, you can validate a sub-schema and its data. Doing this, the intial schema will be used as rootSchema (for example, to resolve $ref from definitions).

const jsonSchema = new Draft07(myJsonSchema);
const errors: JSONError[] = jsonSchema.validate("my-string", { type: "number" });

To prevent some errors when using helper methods with an independent sub-schema, please use compileSchema if it is not retrieved from a draft-method directly (which was compiled by passing it to Draft). Specifically, if the schema contains a $ref you need to use compileSchema. More details in compileSchema.

About type JSONError

In json-schema-library all errors are in the format of a JSONError:

type JSONError = {
    type: "error";
    name: string;
    code: string;
    message: string;
    data?: { [p: string]: any };
};

In almost all cases, a json-pointer is given on error.data.pointer, which points to the source within data where the error occured.

For more details on how to work with errors, refer to section custom errors.

Example 
import { Draft07, JSONSchema, JSONError } from "json-schema-library";

const myJsonSchema: JSONSchema = {
    type: "object",
    additionalProperties: false
};

const jsonSchema = new Draft07(myJsonSchema);
const errors: JSONError[] = jsonSchema.validate({ name: "my-data" });

expect(errors).to.deep.equal([
    {
        type: "error",
        name: "NoAdditionalPropertiesError",
        code: "no-additional-properties-error",
        message: "Additional property `name` in `#` is not allowed",
        data: { property: "name", properties: [], pointer: "#" }
    }
]);
Example with separate schema 
import { Draft07, JSONSchema, JSONError } from "json-schema-library";

const myJsonSchema: JSONSchema = {
    type: "object",
    additionalProperties: false
};

const jsonSchema = new Draft07(myJsonSchema);
const mySchema = jsonSchema.compileSchema({ type: "number" });
const errors: JSONError[] = jsonSchema.validate("my-string", mySchema);

expect(errors).to.deep.equal([
    {
        type: "error",
        name: "TypeError",
        code: "type-error",
        message: "Expected `my-string` (string) in `#` to be of type `number`",
        data: {
            received: "string",
            expected: "number",
            value: "my-string",
            pointer: "#"
        }
    }
]);

isValid

isValid will return true if the given json-data is valid against the json-schema.

const jsonSchema = new Draft07(myJsonSchema);
const isValid: boolean = jsonSchema.isValid(myData);

validateAsync

This method is not yet exposed by a draft directly as the API of this is yet unsatisfactory. Nonetheless, this function is in production and can be used reliably.

Optional support for onError helper, which is invoked for each error (after being resolved):

import { Draft07, JSONError, validateAsync } from "json-schema-library";

const draft = new Core(mySchema);

validateAsync(draft, "", { onError: (err: JSONError) => {}, schema: draft.getSchema() })
    .then((allErrors: JSONError[]) => {});

getTemplate

getTemplate creates input data from a json-schema that is valid to the schema. Where possible, the json-schema default property will be used to initially setup input data. Otherwise, the first values encountered (enum values, initial values, etc.) are user to build up the json-data.

const jsonSchema = new Draft07(myJsonSchema);
const myData = jsonSchema.getTemplate();

Additionally, you can pass input data. getTemplate will then complement any missing values from the schema, while keeping the initial values.

const jsonSchema = new Draft07(myJsonSchema);
const myData = jsonSchema.getTemplate({ name: "input-data" });

Note: If you are using references in your schema, getTemplate will only resolve the first $ref in each path, ensuring no inifinte data structures are created. In case the limit of 1 $ref resolution is too low, you can modify the value globally one by adjusting the json-schema-library settings:

import { settings } from "json-schema-library";
settings..GET_TEMPLATE_RECURSION_LIMIT = 5;
Example 
import { Draft07, JSONSchema } from 'json-schema-library';

const myJsonSchema: JSONSchema = {
  type: 'object',
  properties: {
    name: { type: 'string' },
    option: {
      type: 'string',
      enum: ['first-option', 'second-option']
    },
    list: {
      type: 'array',
      items: {
        type: 'string',
        default: 'new item'
      },
      minItems: 1
    }
  }
};

const jsonSchema = new Draft07(myJsonSchema);
const myData = jsonSchema.getTemplate();

expect(myData).to.deep.equal({
  name: ',
  option: 'first-option',
  list: ['new item']
});
Example with input data 
import { Draft07, JSONSchema } from "json-schema-library";

const myJsonSchema: JSONSchema = {
    type: "object",
    properties: {
        name: { type: "string" },
        option: {
            type: "string",
            enum: ["first-option", "second-option"]
        },
        list: {
            type: "array",
            items: {
                type: "string",
                default: "new item"
            },
            minItems: 1
        }
    }
};

const jsonSchema = new Draft07(myJsonSchema);
const myData = jsonSchema.getTemplate({ name: "input-data", list: [] });

expect(myData).to.deep.equal({
    name: "input-data",
    option: "first-option",
    list: ["new item"]
});

each

each iterates over each data-item (object, array and value) and emits the data-item, schema and location to a callback.

const jsonSchema = new Draft07(mySchema);
const myCallback = (schema: JSONSchema, value: unknown, pointer: JSONPointer) => {
    console.log(schema, value, pointer);
};
jsonSchema.each(myData, myCallback);
Example 
import { Draft07, JSONSchema, JSONPointer } from "json-schema-library";

const mySchema: JSONSchema = {
  type: "array",
  items: [{ type: "number" }, { type: "string" }]
};

const jsonSchema = new Draft07(mySchema);
const calls = [];
const myCallback = (schema: JSONSchema, value: unknown, pointer: JSONPointer) => {
  calls.push({ schema, value, pointer });
};

jsonSchema.each([5, "nine"], myCallback);

expect(calls).to.deep.equal([
  { schema: mySchema, value: [5, "nine"], pointer: "#" },
  { schema: { type: "number" }, value: 5, pointer: "#/0" },
  { schema: { type: "string" }, value: "nine", pointer: "#/1" }
]);

eachSchema

eachSchema emits each sub-schema definition to a callback.

const jsonSchema = new Draft07(mySchema);
const myCallback = (schema: JSONSchema) => {
    console.log(schema);
};
jsonSchema.eachSchema(myCallback);
Example 
import { Draft07, JSONSchema } from "json-schema-library";

const mySchema: JSONSchema = {
  type: "array",
  items: {
    oneOf: [{ type: "number" }, { $ref: "#/$defs/value" }]
  },
  $defs: {
    value: { type: "string" },
    object: { type: "object" }
  }
};

const jsonSchema = new Draft07(mySchema);
const calls = [];
const myCallback = (schema: JSONSchema) => {
  calls.push(schema);
};

jsonSchema.eachSchema(myCallback);

expect(calls).to.deep.equal([
  mySchema,
  { oneOf: [{ type: "number" }, { $ref: "#/$defs/value" }] },
  { type: "number" },
  { $ref: "#/$defs/value" },
  { type: "string" },
  { type: "object" }
]);

getSchema

getSchema retrieves the json-schema of a specific location in data. The location in data is given by a json-pointer. In many cases the schema can be retrieved without passing the actual data, but in situations where the schema is dynamic (e.g., in oneOf, dependencies, etc.), the data is required or will return a JSONError if the location cannot be found.

const jsonSchema = new Draft07(mySchema);
let schemaOfName: JSONSchema | JSONError;
schemaOfName = jsonSchema.getSchema("/list/1/name", myData);
Example 
import { Draft07, JSONSchema, JSONError } from "json-schema-library";

const mySchema = {
  type: "object",
  properties: {
    list: {
      type: "array",
      items: {
        oneOf: [
          {
            type: "object",
            required: ["name"],
            properties: {
              name: {
                type: "string",
                title: "name of item"
              }
            }
          },
          {
            type: "object",
            required: ["description"],
            properties: {
              description: {
                type: "string",
                title: "description of item"
              }
            }
          }
        ]
      }
    }
  }
};

const jsonSchema = new Draft07(mySchema);
let schemaOfItem: JSONSchema | JSONError;
schemaOfItem = jsonSchema.getSchema("/list/1", {
    list: [{ description: "..." }, { name: "my-item" }]
});

expect(schemaOfItem).to.deep.equal({
    type: "object",
    required: ["name"],
    properties: {
        name: {
            type: "string",
            title: "name of item"
        }
    }
});
About JSONPointer

JSON-Pointer defines a string syntax for identifying a specific value within a JSON document and is supported by JSON-Schema. Given a JSON document, it behaves similar to a lodash path (a[0].b.c), which follows JS-syntax, but instead uses / separators (e.g., a/0/b/c). In the end, you describe a path into the JSON data to a specific point.

getChildSchemaSelection

getChildSchemaSelection returns a list of available sub-schemas for the given property. In many cases, a single schema will be returned. For oneOf-schemas, a list of possible options is returned.

This helper always returns a list of schemas.

Note: This helper currenly supports a subset of json-schema for multiple results, mainly oneOf-definitions

const jsonSchema = new Draft07(mySchema);
const schemas: JSONSchema[] = jsonSchema.getChildSchemaSelection("content", localSchema);
Example 
import { Draft07, JSONSchema, JSONError } from "json-schema-library";

const jsonSchema = new Draft07();
const localSchema = {
  type: "object",
  properties: {
    content: {
      oneOf: [
        { type: "string" },
        { type: "number" }
      ]
    }
  }
}

const schemas: JSONSchema[] = jsonSchema.getChildSchemaSelection("content", localSchema);

expect(schemas).to.deep.equal([
  { type: "string" },
  { type: "number" }
])

step

step retrieves the json-schema of a child property or index. Using step it is possible to incrementally go through the data, retrieving the schema for each next item.

const jsonSchema = new Draft07(mySchema);
const localSchema = { type: "object", properties: { title: { type: "string" } } };
const localData = { title: "value" };
const schemaOfTitle = jsonSchema.step("title", localSchema, localData);
Example 
import { Draft07, JSONSchema, JSONError } from "json-schema-library";

const jsonSchema = new Draft07(mySchema);
const localSchema = {
  oneOf: [
    {
      type: "object",
      properties: { title: { type: "string" } }
    },
    {
      type: "object",
      properties: { title: { type: "number" } }
    }
  ]
};
const localData = { title: 4 }
const schemaOfTitle = jsonSchema.step("title", localSchema, localData);

expect(res).to.deep.eq({ type: "number" });

addRemoteSchema

addRemoteSchema lets you add additional schemas that can be referenced by an URL using $ref. Use this to combine multiple schemas without changing the actual schema.

const jsonSchema = new Draft07({
  $ref: "http://json-schema.org/draft-07/schema#definitions/nonNegativeInteger"
});
jsonSchema.addRemoteSchema("http://json-schema.org/draft-07/schema", draft07Schema);
Example 
import { Draft07 } from "json-schema-library";

const jsonSchema = new Draft07({
    $ref: "http://drafts.com/7/schema#definitions/nonNegativeInteger"
});
jsonSchema.addRemoteSchema("http://drafts.com/7/schema", {
  definitions: {
    nonNegativeInteger: { type: "integer", minimum: 0 }
  }
});

const errors = jsonSchema.validate(1);
expect(errors).to.deep.equal([]);

const schema = jsonSchema.getSchema("#");
expect(schema).to.deep.equal({ type: "integer", minimum: 0 });

createSchemaOf

createSchemaOf returns a simple json-schema of the input data.

const jsonSchema = new Draft07(mySchema);
const schema: JSONSchema = jsonSchema.createSchemaOf({ title: "initial value" });

compileSchema

compileSchema adds $ref resolution support to a json-schema. Internally, each draft compiles a passed schema on its own, but when passing additional schemas to individual functions, compileSchema has to be called manually for json-schemas containing $ref-references.

const jsonSchema = new Draft07(mySchema);
const compiledSchema = jsonSchema.compileSchema({ $ref: "/$defs/table" });
const tableSchema = compiledSchema.getRef();

Note: that draft.compileSchema compiles a schema under the current rootSchema. That is, definitions from root schema will be copied to the local schema, to enable $ref resolutions.

Draft extensions

For each draft json-schema-library supports the following custom properties:

patternExample

For error generation, an attribute patternExample may be set for a pattern validation. Instead of the regular expression, the example will be printed in the error message.

oneOfProperty

For oneOf resolution, json-schema states that data is valid if it validates against exactly one of those sub-schemas. In some scenarios this is unwanted behaviour, as the actual oneOf schema is known and only validation errors of this exact sub-schema should be returned.

For an explicit oneOf resolution, the json-schema may be extended by a property oneOfProperty. This will always associate an entry with a matching value (instead of schema validation) and return only this schema or validation errors, depending on the current task. For example:

const schema = {
  oneOfProperty: "id",
  oneOf: [
    {
      type: "object",
      properties: { id: { const: "1" }, title: { type: "number" } }
    },
    {
      type: "object",
      properties: { id: { const: "2" }, title: { type: "number" } }
    },
    {
      type: "object",
      properties: { id: { const: "3" }, title: { type: "number" } }
    }
  ]
};

const resolvedSchema = jsonSchema.resolveOneOf({ id: "2", title: "not a number" }, schema);

// will always return (even if invalid)
expect(resolvedSchema).to.deep.eq({
  type: "object",
  properties: { id: { const: "2" }, title: { type: "number" } }
});

Draft customization

custom resolvers | custom validators | custom errors

Each Draft in json-schema-library is build around a DraftConfig. A DraftConfig holds all functions and configurations for each json-schema drafts. The DraftConfig is your main way to alter or extend behaviour for json-schema-library. You can either create your own draftConfig or adjust any existing draftConfig. For the current drafts (4-7), each draftConfig is exposed along with its actual class. For example:

import { Draft, Draft07, draft07Config } from 'json-schema-library';
// the following calls are identical:
new Draft(draft07Config, mySchema);
new Draft07(mySchema, draft07Config);
new Draft07(mySchema, {});
new Draft07(mySchema);

All draft configurations for specific Draft classes accept a partial configuration that lets you overwrite default behaviour:

replace the strict resolveOneOf behaviour to use fuzzy search instead:

import { Draft07, draft07Config, resolveOneOfFuzzy } from 'json-schema-library';
// the following calls are identical:
new Draft07(mySchema, { resolveOneOf: resolveOneOfFuzzy });
new Draft({ ...draft07Config, resolveOneOf: resolveOneOfFuzzy }, mySchema);

custom resolvers

A resolver is a simple method implementing a specific feature of json-schema to retrieve a sub-schema. Implementing the signature of each resolver you can create and pass your own resolvers.

resolveRef with merge

The default json-schema behaviour for $ref resolution is to replace the schema where a $ref is defined. In some scenarios you what to add context-specific information (e.g., a specific title). For this, a modified $ref-resolver is exposed by json-schema-library:

import { Draft07, draft07Config, resolveRefMerge } from 'json-schema-library';
const jsonSchema = new Draft07(mySchema, { resolveRef: resolveRefMerge });

resolveRefMerge performs a shallow merge (first level of properties), adding the local schemas properties last.

Caution: With this resolver, it is possible to overwrite json-schema behavioural properties. Treat with care.

Example 
import { Draft07, draft07Config, resolveRefMerge } from 'json-schema-library';
const mySchema = {
  type: "object",
  properties: {
    subHeader: {
      $ref: "#/$defs/header",
      title: "sub header"
    }
  },
  "$defs": {
    header: {
      title: "header",
      type: "string"
    }
  }
}

const jsonSchema = new Draft07(mySchema, { resolveRef: resolveRefMerge });
const subHeaderSchema = jsonSchema.getSchema("#/subHeader");

expect(subHeaderSchema).to.eq({
  title: "sub header",
  type: "string"
})

resolveOneOf fuzzy search

The default json-schema behaviour for oneOf resolution is to validate all contained oneOf-schemas and return the one schema that validates against the given input data. If no item validates completely an error returned, containing all validation errors of all schemas. When you are interested in the actual error (rather than simply determining “Is the data is valid or not?”), this is behaviour is not very helpful as the result is hard to read.

json-schema-library exposes a method resolveOneOfFuzzy, which will return a single schema in cases where no valid schema could be resolved. resolveOneOfFuzzy uses a simple scoring mechanism to return the best fitting schema for the given input data. Thus, resolveOneOfFuzzy may return schemas that do not validate a given input data.

import { Draft07, draft07Config, resolveOneOfFuzzy } from 'json-schema-library';
const jsonSchema = new Draft07(mySchema, { resolveOneOf: resolveOneOfFuzzy });

custom validators

All json-schema validation is done using validator functions for keywords and formats.

keyword validators are called for each keyword defined on a json-schema. For example, the following schema will run two keyword-validators (one for items and one of minItems) which are defined in draft.validateKeyword.items and draft.validateKeyword.minItems.

{ type: "object", items: {}, minItems: 1 }

Since valid json-schema keywords vary by their type an additional mapping registers, which keyword should be tested per schema-type. This mapping is defined in draft.typeKeywords:

import { draft07Config } from "json-schema-library";

console.log(draft07Config.typeKeywords.array);
// ["enum", "contains", "items", "minItems", "maxItems", "uniqueItems", "not", "if"];

The keyword format is also registered in draft.validateKeyword.format, but each actual format validation is defined as follows:

format validators are called on each occurrence of a property format in a json-schema. In the next example, the schema will run the email-validator given in draft.validateFormat.email:

{ type: "string", format: "email" }

add custom keyword validator

To add or overwrite a keyword validator, you must add a validator function on your draft config in validateKeyword.

Using specific Draft configuration, where draft configuration objects will be merged:

import { Draft07, JSONValidator } from "json-schema-library";

const jsonSchema = new Draft07(mySchema, {
  validateKeyword: {
    customKeyword: myCustomKeywordValidator as JSONValidator
  },
  // in case for new keywords, or if keyword should be supported on other types
  // add keyword-name to typeKeywords
  typeKeywords: {
    object: draft07Config.typeKeywords.object.concat("customKeyword")
  }
});

Example: Manually extending draft configuration:

import { Draft, draft07Config, JSONValidator } from "json-schema-library";

const myDraftConfiguration = {
  ...draft07Config,
  draft07Config,
  validateKeyword: {
    ...draft07Config.validateKeyword,
    customKeyword: myCustomKeywordValidator as JSONValidator
  },
  // in case for new keywords, or if keyword should be supported on other types
  // add keyword-name to typeKeywords
  typeKeywords: {
    ...draft07Config.typeKeywords,
    object: draft07Config.typeKeywords.object.concat("customKeyword")
  }
};

const jsonSchema = new Draft(myDraftConfiguration, mySchema);

add custom format validator

To add or overwrite a format validator you must add a validator function on your draft config in validateFormat.

import { Draft07, JSONValidator } from "json-schema-library";

const jsonSchema = new Draft07(mySchema, {
  validateFormat: {
    customFormat: myCustomFormatValidator as JSONValidator
  }
});

custom errors

json-schema-library exposes available errors on its draft configurations on DraftConfig.errors and uses a global configuration config.strings for error messages. Both can be extended or modified.

import { draft07Config, createCustomError, config } from "json-schema-library";

draft07Config.errors.enumError;
// (data: { pointer: JSONPointer } & Record<string, unknown>) => JSONError

console.log(config.strings.EnumError)
// "Expected given value `{{value}}` in `{{pointer}}` to be one of `{{values}}`"

Each error message in config.strings receives the data-property of an error. These properties can be referenced in handlebars brackets and will be replaced automatically. To demonstrate the behaviour:

import { render } from "json-schema-library";

render(
  "Expected given value `{{value}}` in `{{pointer}}` to be one of `{{values}}`",
  { pointer: "[A]", value: "[B]" }
);
// "Expected given value `[B]` in `[A]` to be one of ``"

customize error messages

import { config } from "json-schema-library";

config.strings.EnumError = "Invalid enum value {{value}} in {{pointer}}`";

customize errors

import { draft07Config, ErrorData, JSONError } from "json-schema-library";

draft07Config.errors.EnumError = (data: ErrorData): JSONError => {
  return {
    type: "error",
    code: "my-error",
    name: "MyError",
    message: `something went wrong at ${data.pointer}`,
    data
  };
};

create errors

import { draft07Config, createError } from "json-schema-library";
// will use the string from `config.strings.EnumError` as message
const error: JSONError = createError("EnumError", { data: { pointer: "#/location" }});

Breaking Changes

v7.0.0

With version v7.0.0, library export and Draft API has changed heavily. The API is now more consistent across draft-versions and offers a simple and consistent configuration interface for existing and custom drafts. In addition, most standalone functions are no longer exposed separately, but under its current draftConfigs and mainly on each draft-instance. This will help to reduce confusion when consuming this API.

The above documentation reflects all these changes. Just reach out if you have troubles migrating to the latest version.

Details of breaking changes
  • replaced Core interface by new Draft interface
  • changed export of Interface to Draft
  • renamed addSchema to addRemoteSchema
  • changed API of compileSchema to have an additional schema-parameter for rootSchema reference
  • changed compileSchema and addRemoteSchema to work on instance state, instead of global state
  • addRemoteSchema, compileSchema now requires draft instance as first parameter
  • removed direct export of following functions: addValidator, compileSchema, createSchemaOf, each, eachSchema, getChildSchemaSelection, getSchema, getTemplate, isValid, step, validate. They are still accessible under the draftConfigs of each draft-version
  • changed draft version of JsonEditor to draft07

v6.0.0

With version v6.0.0 supported json schema drafts are exported directly as Draft04, Draft06, Draft07. Example use: import { Draft07 } from "json-schema-library".

v5.0.0

With version v5.0.0 the API has changed to es6 modules, where there is no default export, only named exports. Additionally all code has been rewritten in TypeScript. When directly accessing files, switch to dist/module/*.js-files for plain js-modules.

v4.0.0

With version v4.0.0 the API has changed in order to use the defined (root) schema in draft as default where possible. This means most methods have a changed signature, where data is passed before an optional schema argument.