## Status: experimental, shall be included in v1.1
JSON Schema is a draft standard for describing format of JSON. The schema itself is also a JSON. By validating a JSON with JSON Schema, your code can safely access the JSON without manually checking types, or whether a key exists, etc.
RapidJSON implemented a JSON Schema validator for [JSON Schema Draft v4](http://json-schema.org/documentation.html). If you do not familiar with JSON Schema, you may refer to [Understanding JSON Schema](http://spacetelescope.github.io/understanding-json-schema/).
## Basic Usage
First of all, you need to parse a JSON Schema into `Document`, and then compile the `Document` into `SchemaDocument`.
Secondly, construct a `SchemaValidator` with the `SchedmaDocument`. It is similar to a `Writer` in the sense of handling SAX events. So, you can use `document.Accept(validator)` to validate a document, and then check the validity.
~~~cpp
#include "rapidjson/schema.h"
// ...
Documentsd;
if(!sd.Parse(schemaJson)){
// the schema is not a valid JSON.
// ...
}
SchemaDocumentschema(sd);// Compile a Document to SchemaDocument
* One `SchemaDocment` can be referenced by multiple `SchemaValidator`s. It will not be modified by `SchemaValidator`s.
* A `SchemaValidator` may be reused to validate multiple documents. To run it for other documents, call `validator.Reset()` first.
## Validation during parsing/serialization
Differ to most JSON Schema validator implementations, RapidJSON provides a SAX-based schema validator. Therefore, you can parse a JSON from a stream while validating it on the fly. If the validator encounters a JSON value that invalidates the supplied schema, the parsing will be terminated immediately. This design is especially useful for parsing large JSON files.
### DOM parsing
For using DOM in parsing, `Document` needs some preparation and finailizaing tasks, in addition to recieving SAX events, thus it needs some work to route the reader, validator and the document. `SchemaValidatingReader` is a helper class that doing such work.
~~~cpp
#include "rapidjson/filereadstream.h"
// ...
SchemaDocumentschema(sd);// Compile a Document to SchemaDocument
// Use reader to parse the JSON
FILE*fp=fopen("big.json","r");
FileReadStreamis(fp,buffer,sizeof(buffer));
// Parse JSON from reader, validate the SAX events, and store in d.
For using SAX in parsing, it is much simpler. If it only need to validate the JSON without further processing, it is simply:
~~~
SchemaValidator validator(schema);
Reader reader;
if (!reader.Parse(stream, validator)) {
if (!validator.IsValid()) {
// ...
}
}
~~~
This is exactly the method used in [schemavalidator](example/schemavalidator/schemavalidator.cpp) example. The distinct advantage is low memory usage, no matter how big the JSON was (the memory usage depends on the complexity of the schema).
If you need to handle the SAX events further, then you need to use the template class `GenericSchemaValidator` to set the output handler of the validator:
// Some problem during Accept(), it may be validation or encoding issues.
if (!validator.IsValid()) {
// ...
}
}
~~~
Of course, if your appication only needs SAX-style serialization, it can simply send SAX events to `SchemaValidator` instead of `Writer`.
## Remote Schema
JSON Schema supports [`$ref` keyword](http://spacetelescope.github.io/understanding-json-schema/structuring.html), which is a [JSON pointer](pointer.md) referencing to a local or remote schema. Local pointer is prefixed with `#`, while remote pointer is an relative or absolute URI. For example:
~~~js
{"$ref":"definitions.json#/address"}
~~~
As `SchemaValidator` does not know how to resolve such URI, it needs a user-provided `IRemoteSchemaDocumentProvider` instance to do so.
~~~
class MyRemoteSchemaDocumentProvider : public IRemoteSchemaDocumentProvider {
// Resolve the uri and returns a pointer to that schema.
}
};
// ...
MyRemoteSchemaDocumentProvider provider;
SchemaValidator validator(schema, &provider);
~~~
## Conformance
RapidJSON passed 262 out of 263 tests in [JSON Schema Test Suite](https://github.com/json-schema/JSON-Schema-Test-Suite)(Json Schema draft 4).
The failed test is "changed scope ref invalid" of "change resolution scope" in `refRemote.json`. It is due to that `id` schema keyword and URI combining function are not implemented.
Besides, the `format` schema keyword for string values is ignored, since it is not required by the specification.
### Regular Expression
The schema keyword `pattern` and `patternProperties` uses regular expression to match the required pattern.
RapidJSON implemented a simple DFA regular expression engine, which is used by default. It supports the following syntax.
|Syntax|Description|
|------|-----------|
|`ab` | Concatenation
|`a|b` | Alternation
|`a?` | Zero or one
|`a*` | Zero or more
|`a+` | One or more
|`a{3}` | Exactly 3 times
|`a{3,}` | At least 3 times
|`a{3,5}`| 3 to 5 times
|`(ab)` | Grouping
|`^a` | At the beginning
|`a$` | At the end
|`.` | Any character
|`[abc]` | Character classes
|`[a-c]` | Character class range
|`[a-z0-9_]` | Character class combination
|`[^abc]` | Negated character classes
|`[^a-c]` | Negated character class range
|`[\b]` | Backspace (U+0008)
|``\|`, `\\`, ... | Escape characters
|`\f` | Form feed (U+000C)
|`\n` | Line feed (U+000A)
|`\r` | Carriage return (U+000D)
|`\t` | Tab (U+0009)
|`\v` | Vertical tab (U+000B)
For C++11 compiler, it is also possible to use the `std::regex` by defining `RAPIDJSON_SCHEMA_USE_INTERNALREGEX=0` and `RAPIDJSON_SCHEMA_USE_STDREGEX=1`. If your schemas do not need `pattern` and `patternProperties`, you can set both macros to zero to disable this feature, which will reduce some code size.
## Performance
Most C++ JSON libraries have not yet supporting JSON Schema. So we tried to evluate the performance of RapidJSON's JSON Schema validator according to [json-schema-benchmark](https://github.com/ebdrup/json-schema-benchmark), which tests 11 JavaScript libraries running on Node.js.
That benchmark runs validations on [JSON Schema Test Suite](https://github.com/json-schema/JSON-Schema-Test-Suite), in which some test suites and tests are excluded.
On a Mac Book Pro (2.8 GHz Intel Core i7), the following results are gathered.
|Validator|Relative speed|Number of test runs per second|
