ReactiveBase is a container component that wraps all the ReactiveSearch
components together. It binds the backend app (data source) with the UI view components (elements wrapped within ReactiveBase), allowing a UI component to be reactively updated every time there is a change in the data source or in other UI components.
This is the first component you will need to add when using ReactiveSearch
.
Usage
<ReactiveBase
app="appname"
credentials="abcdef123:abcdef12-ab12-ab12-ab12-abcdef123456"
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
Props
app
Type | Optional |
---|---|
String |
Yes |
app name as it appears on the dashboard. Refers to an index if you're using your own Elasticsearch cluster. (Multiple indexes can be connected to by specifying comma-separated index names)
type
Type | Optional |
---|---|
String |
Yes |
types on which the queries should run on. Multiple types can be passed as comma-separated values. The default behavior here is to search on all the app types.
credentials
Type | Optional |
---|---|
String |
Yes |
app credentials as they appear on the dashboard. It should be a string of the format "username:password" and is used for authenticating the app. If you are not using an appbase.io app, credentials may not be necessary - although having an open-access Elasticsearch cluster is not recommended.
url
Type | Optional |
---|---|
String |
Yes |
URL where the Elasticsearch cluster is hosted, only needed if your app uses a non-appbase.io URL.
headers
Type | Optional |
---|---|
Object |
Yes |
set custom headers to be sent with each server request as key/value pairs. For example:
<ReactiveBase
app="appname"
credentials="abcdef123:abcdef12-ab12-ab12-ab12-abcdef123456"
headers={{
secret: 'reactivesearch-is-awesome'
}}
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
endpoint
Type | Optional |
---|---|
Object |
Yes |
endpoint prop provides the ability to query a user-defined backend service for ReactiveBase and its children components. This service is typically a ReactiveSearch backend pipeline or any other API handler that works with the ReactiveSearch API references.
Accepts the following properties:
- url
String
[Required] URL where the data cluster is hosted. - headers
Object
[optional]
set custom headers to be sent with each server request as key/value pairs. - method
String
[optional]
set method of the API request. - body
Object
[optional]
request body of the API request. When body isn't set and method is POST, the request body is set based on the component's configured props.
Top level props -
url
,app
andcredentials
are optional and overridden in the final request whenendpoint
prop is defined.
<ReactiveBase
endpoint={{
url:"https://appbase-demo-ansible-abxiydt-arc.searchbase.io/recipes-demo/_reactivesearch.v3", //mandatory
headers: {
// put relevant headers
}
method: 'POST'
}}
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
reactivesearchAPIConfig
Type | Optional |
---|---|
Object |
Yes |
allows you to customize the analytics experience when appbase.io is used as a backend. Read more about it over here.
initialQueriesSyncTime
Type | Optional |
---|---|
Number |
Yes |
allows you to define a wait time in milliseconds. We wait for initialQueriesSyncTime
time to combine the individual component queries to a single network request at initial load. This prop is helpful to optimize the performance when you have a lot of filters on the search page, using a wait time of 100-200
milliseconds would merge the multiple requests into a single request.
httpRequestTimeout
Type | Optional |
---|---|
Number |
Yes |
allows you to define a timeout in milliseconds to timeout a fired HTTP request. If the request takes longer than the specified time, an error will be thrown. You can handle this error at the component level using the onError callback prop. Defaults to 30
seconds.
as
Type | Optional |
---|---|
String |
Yes |
allows to use the custom html element tag, defaults to div
.
theme
Type | Optional |
---|---|
Object |
Yes |
allows over-writing of default styles by providing the respective key/values. You can read more about its usage here
themePreset
Type | Optional |
---|---|
String |
Yes |
allows over-writing of default styles by providing a preset value. Supported values are light
(default) and dark
. You can read more about its usage here
getSearchParams
Type | Optional |
---|---|
Function |
Yes |
Enables you to customize the evaluation of query-params-string from the url (or) any other source. If this function is not set, the library will use window.location.search
as the search query-params-string for parsing selected-values. This can come handy if the URL is using hash values.
setSearchParams
Type | Optional |
---|---|
Function |
Yes |
Enables you to customize setting of the query params string in the url by providing the updated query-params-string as the function parameter. If this function is not set, the library will set the window.history
via pushState
method.
transformRequest
Type | Optional |
---|---|
Function |
Yes |
Enables transformation of network request before execution. This function will give you the request object as the param and expect an updated request in return, for execution. Note that this is an experimental API and will likely change in the future.
Note:
From v3.0.1 it is possible to define
transformRequest
as anasync
method which will return a promise which resolves the modified request options.
If you need to include credentials (credentials are cookies, authorization headers or TLS client certificates), you can do it this way:
<ReactiveBase
app="appname"
transformRequest={props => ({
...props,
credentials: "include",
})}
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
You can also modify the request URL
in that way:
<ReactiveBase
app="appname"
transformRequest={props => ({
...props,
url: props.url.replace('_msearch', '_search')
})}
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
The above example will change the default _msearch
request to _search
request.
graphQLUrl
Type | Optional |
---|---|
String |
Yes |
Allows user to query from GraphqQL server instead of Elasticsearch
REST api. graphql-compose-elasticsearch helps in transforming GraphQL
queries into Elasticsearch
rest api. Here is an example of GraphQL
server which acts as proxy for Elasticsearch
.
transformResponse
Type | Optional |
---|---|
Function |
Yes |
Enables transformation of search network response before rendering them. This asynchronous function will give you elasticsearch response object and componentId as params and expects an updated response in return in the similar structure of elasticsearch. You can use componentId
to conditionally transform response for a particular reactivesearch component only.
<ReactiveBase
app="appname"
credentials="abcdef123:abcdef12-ab12-ab12-ab12-abcdef123456"
headers={{
secret: 'reactivesearch-is-awesome'
}}
transformResponse={async (elasticsearchResponse, componentId) => { const ids = elasticsearchResponse.responses[0].hits.hits.map( item => item._id ); const extraInformation = await getExatraInformation(ids); const hits = elasticsearchResponse.responses[0].hits.hits.map( (item) => { const extraInformationItem = extraInformation.find( otherItem => otherItem._id === item._id ); return { ...item, ...extraInformationItem }; } ); return { response: [ { ...elasticsearchResponse.responses[0], hits: { ...elasticsearchResponse.responses[0].hits, hits } } ] }; }}
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
Note
transformResponse
function is expected to return data in following structure.
{
response: [
{
hits: {
hits: [...],
total: 100
},
took: 1
}
]
}
💡 Use below props when using ReactiveMaps 🗺️ library.
mapKey
Type | Optional |
---|---|
String |
Yes |
accepts the secret key to load google script.
mapLibraries
Type | Optional |
---|---|
Array<String> |
Yes |
accepts the list of addiotional maps libraries to load. The following are available according to Google Docs:
drawing
geometry
localContext
places
visualization
<ReactiveBase
mapKey="<YOUR_MAP_KEY>"
mapLibraries={['visualization', 'places']}
// ...other props
/>
Mandatorily pass
mapLibraries={['places']}
when using either or both of GeoDistanceDropdown/ GeoDistanceSlider components from ReactiveMaps 🗺️ .
preferences
Type | Optional |
---|---|
Object |
Yes |
accepts an object to configure the search settings for components. The preferences
property allows configuring the search settings for your components in one place. The preferences object must follow the following structure:
{
componentSettings: {
[componentId]: {
// If disabled, the component would not get rendered and would not constitute the search query
enabled: true,
rsConfig: {
// props supported by reactivesearch components
}
// can have additional keys to store meta data for components
// the preference object for a component can be accessed using the
// `SearchPreferencesContext` context.
// For e.g custom property to control the collapsible property for a facet
isCollapsible: true,
}
}
}
To connect a ReactiveSearch component to a preference, use the componentId
prop. The following example has defined the preferences
object for bookSearch
component, the DataSearch
component is using the same componentId (bookSearch
).
<ReactiveBase preferences={{
componentSettings: {
bookSearch: {
rsConfig: {
dataField: 'original_title',
title: 'Search for Books',
size: 5,
}
}
}
}}>
SearchBox componentId="bookSearch" />
</ReactiveBase>
Additionally, the ReactiveSearch components support preferencesPath
prop which is helpful to define the path of preference object for a component. It is helpful when you have to use conflicting component Ids. The following example defines the preferences for home
and search
pages, components have defined the preferencesPath
prop to connect to preferences.
<ReactiveBase preferences={{
pages: {
home: {
bookSearch: {
rsConfig: {
dataField: 'original_title',
title: 'Search for Books',
size: 5,
}
}
},
search: {
bookSearch: {
rsConfig: {
dataField: ['original_title', 'authors', 'publishers'],
title: 'Search for Books, Authors, Publishers',
size: 10,
}
}
}
}
}}>
{/** home page */}
SearchBox
preferencesPath="pages.home.bookSearch"
componentId="bookSearch"
/>
{/** search page */}
SearchBox
preferencesPath="pages.search.bookSearch"
componentId="bookSearch"
/>
</ReactiveBase>
Note:
Preferences is meant to be a one time configuration for components. We don't recommend to mutate it as it can cause performance issues.
Connect to Elasticsearch
Note
An app within ReactiveSearch's context refers to an index in Elasticsearch.
ReactiveSearch works out of the box with an Elasticsearch index hosted anywhere. You can use the url
prop of the ReactiveBase component to connect the child ReactiveSearch components to your own index. For example,
<ReactiveBase
app="your-elasticsearch-index"
url="http://your-elasticsearch-cluster"
>
<Component1 .. />
<Component2 .. />
</ReactiveBase>
It's also possible to secure your Elasticsearch cluster's access with a middleware proxy server that is connected to ReactiveSearch. This allows you to set up custom authorization rules, prevent misuse, only pass back non-sensitive data, etc. Here's an example app where we show this using a Node.JS / Express middleware:
Note
If you are using the url prop for an Elasticsearch cluster, ensure that your ReactiveSearch app can access the cluster. Typically, you will need to configure CORS in elasticsearch.yml to enable access.
http.cors.enabled: true
http.cors.allow-credentials: true
http.cors.allow-origin: 'http://reactive-search-app-domain.com:port'
http.cors.allow-headers: X-Requested-With, X-Auth-Token, Content-Type, Content-Length, Authorization, Access-Control-Allow-Headers, Accept
Note
If you are using Elasticsearch on AWS, then the recommended approach is to connect via the middleware proxy as they don't allow setting the Elasticsearch configurations.
Next Steps
Once you have added the ReactiveBase component, you can get started with adding other components as its children.
- List specific components can be found here.
- Range specific components can be found here.
- Search specific components can be found here.
- Result specific components can be found here.
You can read more about when to use which components in the overview guide here.