This project extends react-jsonschema-form with conditional logic, which allow to have more complicated logic expressed and controlled with JSON schema. This is primarily useful for complicated schemas with extended business logic, which are suspect to changes and need to be manageable and changeable without modifying running application.
This version supports @rjsf/core v2 and above
if you need to support an earlier version of react-jsonschema-form
use https://github.com/RxNT/react-jsonschema-form-conditionals#readme
- Support for Json Rules Engine and json-rules-engine-simplified
- Extensible action mechanism
- Configuration over coding
- Lightweight and extensible
Install rjsf-conditionals
by running:
npm i rjsf-conditionals
The simplest example of using rjsf-conditionals
import applyRules from 'rjsf-conditionals';
import Engine from 'json-rules-engine-simplified';
import Form from "@rjsf/core";
// ...
const rules = [{
// ...
}];
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine)(Form);
ReactDOM.render(<FormWithConditionals />, document.querySelector('#app'));
To show case uses for this library we'll be using simple registration schema example
import applyRules from 'rjsf-conditionals';
import Form from "@rjsf/core";
let schema = {
definitions: {
hobby: {
type: "object",
properties: {
name: { type: "string" },
durationInMonth: { "type": "integer" },
}
}
},
title: "A registration form",
description: "A simple form example.",
type: "object",
required: [
"firstName",
"lastName"
],
properties: {
firstName: {
type: "string",
title: "First name"
},
lastName: {
type: "string",
title: "Last name"
},
age: {
type: "integer",
title: "Age",
},
bio: {
type: "string",
title: "Bio",
},
country: {
type: "string",
title: "Country"
},
state: {
type: "string",
title: "State"
},
zip: {
type: "string",
title: "ZIP"
},
password: {
type: "string",
title: "Password",
minLength: 3
},
telephone: {
type: "string",
title: "Telephone",
minLength: 10
},
hobbies: {
type: "array",
items: { "$ref": "#/definitions/hobby" }
}
}
}
let rules = [{
...
}]
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine)(Form);
render((
<FormWithConditionals />
), document.getElementById("app"));
Conditionals functionality is build using 2 things
- Rules engine (Json Rules Engine or Simplified Json Rules Engine)
- Schema action mechanism
Rules engine responsibility is to trigger events, action mechanism performs needed actions on the requests.
Project supports 2 rules engines out of the box:
In order to use either of those, you need to specify Engine
in applyRules
configuration.
For example:
To use Simplified Json Rules Engine, you can do following:
import applyRules from 'rjsf-conditionals';
import Form from "@rjsf/core";
import Engine from 'json-rules-engine-simplified';
...
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine)(Form);
ReactDOM.render(
<FormWithConditionals />,
document.querySelector('#app')
);
To use Json Rules Engine, is almost the same:
import applyRules from 'rjsf-conditionals';
import Engine from 'json-rules-engine';
import Form from "@rjsf/core";
// ...
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine)(Form);
ReactDOM.render(
<FormWithConditionals />,
document.querySelector('#app')
);
If non of the provided engines satisfies, your needs, you can
implement your own Engine
which should
comply to following:
class Engine {
constructor(rules, schema) {
}
addRule = (rule) => {
}
run = (formData) => {
return Promise[Event]
}
}
Original rules
and schema
is used as a parameter for a factory call,
in order to be able to have additional functionality, such as rules to schema compliance validation,
like it's done in Simplified Json Rules Engine](https://github.com/RxNT/json-rules-engine-simplified)
Rules engine emits events, which are expected to have a type
and params
field,
type
is used to distinguish action that is needed, params
are used as input for that action:
{
"type": "remove",
"params": {
"field": "name"
}
}
By default action mechanism defines a supported set of rules, which you can extend as needed:
remove
removes a field or set of fields fromschema
anduiSchema
require
makes a field or set of fields required
If you want to remove a field, your configuration should look like this:
{
"conditions": { },
"event": {
"type": "remove",
"params": {
"field": "password"
}
}
}
When condition
is met, password
will be removed from both schema
and uiSchema
.
In case you want to remove multiple fields name
, password
, rule should look like this:
{
"conditions": { },
"event": {
"type": "remove",
"params": {
"field": [ "name", "password" ]
}
}
}
To remove nested schema properties, use json dot notation. e.g. For schema object:
{
"type": "object",
"properties": {
"someParentWrapper": {
"type": "object",
"properties": {
"booleanValA": {
"type": "boolean",
"title": "Some boolean input"
},
"booleanValB": {
"type": "boolean",
"title": "Another boolean input"
}
}
}
}
}
You can remove the nested booleanValA or booleanValB like so:
{
"conditions": { },
"event": {
"type": "remove",
"params": {
"field": "someParentWrapper.booleanValA"
}
}
}
The same convention goes for require
action
For a single field:
{
"conditions": { },
"event": {
"type": "require",
"params": {
"field": "password"
}
}
}
For multiple fields:
{
"conditions": { },
"event": {
"type": "require",
"params": {
"field": [ "name", "password"]
}
}
}
API defines a set of actions, that you can take on uiSchema
, they cover most of the
uiAppend
appendsuiSchema
specified in params with an originaluiSchema
uiOverride
replaces field in originaluiSchema
with fields inparams
, keeping unrelated entriesuiRepalce
replaces wholeuiSchema
with a conf schema
To show case, let's take a simple schema
{
"properties": {
"lastName": { "type": "string" },
"firstName": { "type": "string" },
"nickName": { "type": "string"}
}
}
and uiSchema
{
"ui:order": ["firstName"],
"lastName": {
"classNames": "col-md-1",
},
"firstName": {
"ui:disabled": false,
"num": 23
},
"nickName": {
"classNames": "col-md-12"
}
}
with event params
something like this
{
"ui:order": [ "lastName" ],
"lastName": {
"classNames": "has-error"
},
"firstName" : {
"classNames": "col-md-6",
"ui:disabled": true,
"num": 22
}
}
And look at different results depend on the choosen action.
UiAppend can handle arrays
and string
, with fallback to uiOverride
behavior for all other fields.
So the expected result uiSchema
will be:
{
"ui:order": ["firstName", "lastName"],
"lastName": {
"classNames": "col-md-1 has-error"
},
"firstName": {
"classNames": "col-md-6",
"ui:disabled": true,
"num": 22
},
"nickName": {
"classNames": "col-md-12"
}
}
In this case it
- added
lastName
toui:order
array, - appended
has-error
toclassNames
inlastName
field - added
classNames
and enabledfirstName
- as for the
num
infirstName
it just overrode it
This is useful for example if you want to add some additional markup in your code, without touching layout that you've defined.
uiOverride
behaves similar to append, but instead of appending it completely replaces overlapping values
So the expected result uiSchema
will be:
{
"ui:order": [ "lastName" ],
"lastName": {
"classNames": "has-error"
},
"firstName": {
"classNames": "col-md-6",
"ui:disabled": true,
"num": 22
},
"nickName": {
"classNames": "col-md-12"
}
}
In this case it
ui:order
was replaced with configured valueclassName
for thelastName
was replaced withhas-error
- added
classNames
and enabledfirstName
- as for the
num
infirstName
it just overrode it
uiReplace
just replaces all fields in uiSchema
with params
fields, leaving unrelated fields untouched.
So the result uiSchema
will be
{
"ui:order": [ "lastName" ],
"lastName": {
"classNames": "has-error"
},
"firstName" : {
"classNames": "col-md-6",
"ui:disabled": true,
"num": 22
},
"nickName": {
"classNames": "col-md-12"
}
}
You can extend existing actions list, by specifying extraActions
on the form.
Let's say we need to introduce replaceClassNames
action, that
would just specify classNames
col-md-4
for all fields except for ignore
d one.
We also want to trigger it only when password
is empty
.
This is how we can do this:
import applyRules from 'rjsf-conditionals';
import Engine from 'json-rules-engine-simplified';
import Form from "@rjsf/core";
...
const rules = [
{
conditons: {
password: "empty"
},
event: {
type: "replaceClassNames",
params: {
classNames: "col-md-4",
ignore: [ "password" ]
}
}
}
];
let extraActions = {
replaceClassNames: function(params, schema, uiSchema, formData) {
Object.keys(schema.properties).forEach((field) => {
if (uiSchema[field] === undefined) {
uiSchema[field] = {}
}
uiSchema[field].classNames = params.classNames;
}
}
};
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine, extraActions)(Form);
ReactDOM.render(
<FormWithConditionals/>,
document.querySelector('#app')
);
Provided snippet does just that.
In case you need to calculate value, based on other field values, you can also do that.
Let's say we want to have schema with a
, b
and sum
fields
import applyRules from 'rjsf-conditionals';
import Engine from 'json-rules-engine-simplified';
import Form from "@rjsf/core";
...
const rules = [
{
conditons: {
a: { not: "empty" },
b: { not: "empty" }
},
event: {
type: "updateSum"
}
}
];
let extraActions = {
updateSum: function(params, schema, uiSchema, formData) {
formData.sum = formData.a + formData.b;
}
};
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine, extraActions)(Form);
ReactDOM.render(
<FormWithConditionals/>,
document.querySelector('#app')
);
This is how you can do that.
WARNING!!! You need to be careful with a rules order, when using calculated values. Put calculation rules at the top of your rules specification.
For example, let's say you want to mark sum
field, if you have sum greater
than 10
. The rule would look something like this:
{
"conditions": {
"sum": { "greater" : 10 }
},
"event": {
"type": "appendClass",
"classNames": "has-success"
}
}
But it will work only if you put it after updateSum
rule, like this
[
{
"conditons": {
"a": { "not": "empty" },
"b": { "not": "empty" }
},
"event": {
"type": "updateSum"
}
},
{
"conditions": {
"sum": { "greater" : 10 }
},
"event": {
"type": "appendClass",
"classNames": "has-success"
}
}
];
Otherwise it will work with old sum
values and therefor show incorrect value.
Originally actions performed in sequence defined in the array. If you have interdependent rules, that you need to run in order
you can specify order
on a rule, so that it would be executed first. Rules are executed based on order from lowest to highest with
rules without order executed last.
For example to make updateSum work regardless the order rules were added, you can do following:
[
{
"conditions": {
"sum": { "greater" : 10 }
},
"order": 1,
"event": {
"type": "appendClass",
"classNames": "has-success"
}
},
{
"conditons": {
"a": { "not": "empty" },
"b": { "not": "empty" }
},
"order": 0,
"event": {
"type": "updateSum"
}
}
]
Here although updateSum
comes after appendClass
, it will be executed first, since it has a lower order.
All default actions are validated by default, checking that field exists in the schema, to save you some headaches. There are 2 levels of validation
propTypes
validation, using FBprop-types
package- explicit validation
You can define those validations in your actions as well, to improve actions usability.
All validation is disabled in production.
This is reuse of familiar prop-types
validation used with React components, and it's used in the same way:
In case of require
it can look like this:
require.propTypes = {
field: PropTypes.oneOfType([
PropTypes.string,
PropTypes.arrayOf(PropTypes.string),
]).isRequired,
};
The rest is magic.
WARNING, the default behavior of prop-types
is to send errors to console,
which you need to have running in order to see them.
For our replaceClassNames
action, it can look like this:
replaceClassNames.propTypes = {
classNames: PropTypes.string.isRequired,
ignore: PropTypes.arrayOf(PropTypes.string)
};
In order to provide more granular validation, you can specify validate function on
your action, that will receive params
, schema
and uiSchema
so you could provide appropriate validation.
For example, validation for require
can be done like this:
require.validate = function({ field }, schema, uiSchema) {
if (Array.isArray(field)) {
field
.filter(f => schema && schema.properties && schema.properties[f] === undefined)
.forEach(f => console.error(`Field "${f}" is missing from schema on "require"`));
} else if (
schema &&
schema.properties &&
schema.properties[field] === undefined
) {
console.error(`Field "${field}" is missing from schema on "require"`);
}
};
Validation is not mandatory, and will be done only if field is provided.
For our replaceClassNames
action, it would look similar:
replaceClassNames.validate = function({ ignore }, schema, uiSchema) {
if (Array.isArray(field)) {
ignore
.filter(f => schema && schema.properties && schema.properties[f] === undefined)
.forEach(f => console.error(`Field "${f}" is missing from schema on "replaceClassNames"`));
} else if (
schema &&
schema.properties &&
schema.properties[ignore] === undefined
) {
console.error(`Field "${ignore}" is missing from schema on "replaceClassNames"`);
}
};
In order to listen for configuration changes you can specify onSchemaConfChange
, which will be notified every time schema
or uiSchema
changes it's value.
let FormWithConditionals = applyRules(schema, uiSchema, rules, Engine, extraActions)(Form);
ReactDOM.render(
<FormWithConditionals onSchemaConfChange = {({ schema, uiSchema }) => { console.log("configuration changed") }}/>,
document.querySelector('#app')
);
This repo is using Github Actions as a CI environment which is used to publish releases to NPM.
- Change version in
package.json
and inpackage-lock.json
to2.3.4
or another semver one - Create a new tag and push it
git tag v2.3.4 && git push --tags
- Create a new release on Github and select the tag to release against, i.e.
v2.3.4
- Add a title and notes. Notes usually are extracted by running
git log --oneline
on a dev machine - Click to publish a release and Github Action will pick this event up (see .github/workflows/npmpublish.yml)
- Issue Tracker: https://github.com/ivarprudnikov/rjsf-conditionals/issues
- Source Code: https://github.com/ivarprudnikov/rjsf-conditionals
If you are having issues, please let us know.
The project is licensed under the Apache-2.0 license.