The ibm_opensearch package includes a data store conforming to the dojo.data Read API. The ibm_opensearch package is used to interface with a server with open search capabilities. Typically, server-side support consists of hosting an open search description document that defines the URLs used to query the server. The server can return the data in any number of formats, with the return type defined in the URL element in the description document. Typically, the return type is either X/HTML or an Atom or RSS feed.
This data store is instantiated using a URL to a open search description document. Unless you have a proxy setup, such as the AjaxProxy component, this document and all search endpoints must be on the same server (same domain) as the page that the data store is created in, because of cross-domain limitations of browsers. When the store is instantiated, it parses the description document, determines the best URL element to use, if there are multiple elements, based on the ease of parsing the results. The types, in order, are Atom, RSS, HTML. Then, when results are fetched from the data store, it replaces the searchTerms parameter and any other supplied parameters, based on the template and what is provided in the request. The store queries the server to retrieve the results. Again, due to cross-domain limitations, the URL endpoints must be on the same domain as the page on which the datastore resides.
See the Reference and Usage Examples pages.
The OpenSearch data store is distributed in two different packages.
The package is laid out in the following manner: