link-server-class v1.0.1

link-server-class

Provides a class that runs an http server with paths addressing processes that run search queries, many of which may be common database type of search queries.

Query execution and path parameters provide the basics for paging data.

The class is highly configurable. In its default form, it can handle queries made in a narrow, easily-parsed syntax. It then stores its data, query data, aggregation results, etc. in common JavaScript objects in application memory. In its configured form, it can extend or replace the query language and use any sort of backend storage that provides basic container methods.

Much of the flexibility stems from the use of copious-little-searcher and copious-registry.

The finall class, provided in this module, provides the constructor that takes in the configuration, provides a initialization method that prepars data classes and the API paths for HTTP, and provides method for starting and stopping the application server.

Usage

npm install -s link-server-class

Typical Application Pattern

//
const GeneralDataServer = require('link-server-class')
//


// ---- ---- ---- ---- CONSTRUCT  ---- ---- ---- ---- ---- ---- ---- ----

if ( g_conf.application_searcher === undefined || g_conf.application_searcher === false ) {
    g_conf.application_searcher = './application_searching.js'
}

let g_data_server = new GeneralDataServer(g_conf)

// ---- ---- ---- ---- START  ---- ---- ---- ---- ---- ---- ---- ----
//
g_data_server.initialize(TIMEOUT_THRESHHOLD_INTERVAL)
//
g_data_server.start()
//

// ---- ---- ---- ---- SHUTDOWN  ---- ---- ---- ---- ---- ---- ---- ----

// Handle ^C
process.on('SIGINT', () => {
    console.log("shutting down")
    try {
        g_data_server.shutdown()
    } catch(e) {
        process.exit(0)
    }
});

Web API paths

• /
• /:uid/:query/:bcount/:offset
• /custom/:owner/:query/:bcount/:offset
• /custom/:op/:owner
• /cycle/:halt -- (requires admin permission)
• /reload -- (requires admin permission)
• /persistence/add-publisher/:plink -- (requires admin permission)

Principle Classes

The module export three princpile classes:

• GeneralLinkServer
• RecordSearchApp
• ServiceEntryWatcher

The GeneralLinkServer uses RecordSearchApp or its descendants to respond to web requests that it initializes. In turn, RecordSearchApp acts as a go-between data arrival handlers and searching processes.

Data arrival handlers are network and directory interfaces that take in new items to store for future searching. The ServiceEntryWatcher looks for new objects in a directory by deriving these process from DirWatcherHandler found in copious-little-searcher. The ServiceEntryWatcher also subscribes to publication process when it is commanded to do so. The links to the publishers are added in response to /persistence/add-publisher/:plink an API endpoint of GeneralLinkServer. The ServiceEntryWatcher passes new publications onto the RecordSearchApp which then installs them into searching.

The conigured conf.application_searcher should be a module accessible in the file system and it should provide a class that is a descendant of searching in copious-little-searcher.

This module does not provide a default for conf.application_searcher. Notice that in the above example code (Typical Application Pattern) the application provides it.

The provided application code should descend from Searching provided by copious-little-searcher.