High-Level System Architecture

Relevant source files

The following files were used as context for generating this wiki page:

• routes/index.js
• public/js/p3/app/p3app.js
• public/js/p3/widget/app/AppBase.js
• public/js/p3/widget/WorkspaceBrowser.js
• public/js/p3/widget/viewer/SystemStatus.js

The DXKB (Data and Knowledge Base) web application is a complex bioinformatics platform built on a distributed service-oriented architecture. It integrates multiple backend bio-services, a centralized data API, and a workspace management system to provide genomic analysis tools, data visualization, and collaborative research environments. The frontend is primarily a Single Page Application (SPA) driven by the Dojo Toolkit, utilizing a custom routing system to navigate between specialized analysis widgets and data viewers.

Core System Components

The DXKB architecture consists of a web server layer, a specialized API gateway layer, and various backend analytical services. The system relies on a token-based authentication mechanism to secure user data across these distributed components.

Backend Infrastructure and APIs

The system communicates with several specialized backend services, which are monitored for health and availability. Key services include:

Service Name	Purpose	Configuration Key
Auth	User registration, login, and token management.	userServiceURL
Data API	Centralized access to public and private PATRIC data.	dataAPI
Workspace	Management of user-created folders and analysis objects.	workspaceServiceURL
App Service	Handling of asynchronous genomic analysis job submissions.	serviceAPI
Homology	Specialized service for BLAST queries.	homologyServiceURL
Shock	Storage and retrieval of large data files.	shockServiceURL

Sources: public/js/p3/widget/viewer/SystemStatus.js:15-55, public/js/p3/app/p3app.js:410-435

Client-Side Application Structure

The application logic is encapsulated in the p3app module, which extends a base App class. It manages the application lifecycle, including startup, login verification, and global topic subscriptions for inter-widget communication.

Routing and Navigation

DXKB uses a custom client-side router to manage state transitions without full page reloads. The router maps URL patterns to specific Dojo widgets (viewers or apps) and handles complex state parameters passed via URL segments, query strings, and hashes.

Route Registration Examples

The application registers various routes that determine which widget class to instantiate based on the URL path:

URL Pattern	Target Widget Class	Purpose
/job(/.*)	p3/widget/JobManager	Monitoring analysis job status.
/workspace(/.*)	p3/widget/WorkspaceManager	Browsing and managing files.
/view/(.*)	p3/widget/viewer/[type]	Data-specific visualizations.
/app/(.*)	p3/widget/app/[type]	Analytical tool submission forms.

Sources: public/js/p3/app/p3app.js:180-380

State Management

The getState and populateState functions parse the current URL to build a state object that includes the pathname, search parameters, and hash parameters. This object is used to initialize the target widget with the correct context.

Sources: public/js/p3/app/p3app.js:115-160

Analysis Tool Architecture (AppBase)

Analysis tools (e.g., MSA, Primer Design, Comprehensive Genome Analysis) derive from AppBase.js. This base class provides a standardized framework for handling job submissions to the App Service.

Job Submission Flow

1. Validation: Widgets perform local validation of user inputs (e.g., checking FASTA formats).
2. Parameter Mapping: Form values are converted into a JSON payload compatible with the backend service.
3. Authentication: Requests are signed with the authorizationToken stored in local storage.
4. Submission: The doSubmit function calls the AppService.start_app2 RPC method.

Workspace and Data Interaction

The WorkspaceBrowser acts as a central hub for interacting with data stored in the user's workspace. It utilizes a ContainerActionBar for global actions and an ActionBar for context-sensitive operations based on selected items.

Contextual Actions

Actions are dynamically available based on the validTypes of the selected workspace objects. For example:

• View: Navigates to a specific viewer based on file type (e.g., .nwk files trigger PhylogeneticTree2).
• MSA: Only valid for feature_group types; it triggers a Multiple Sequence Alignment rerun.
• Download: Handles single file downloads or zipping multiple files via the WorkspaceManager.

Sources: public/js/p3/widget/WorkspaceBrowser.js:180-250, public/js/p3/widget/WorkspaceBrowser.js:400-450, public/js/p3/widget/WorkspaceBrowser.js:1050-1150

Authentication and Security

DXKB implements a persistent authentication state using browser localStorage.

• Token Persistence: The tokenstring and userProfile are stored locally.
• Session Monitoring: The application performs a periodic check (localStorageCheckInterval) to ensure the token hasn't been cleared in another tab or expired.
• Automatic Refresh: If the mouse is active or an upload is in progress, the application automatically calls the /authenticate/refresh/ endpoint to extend the session.
• Admin Elevation: Support for "Super User" (SU) mode allows administrators to switch identities, tracked via Atokenstring and Aauth storage keys.

Sources: public/js/p3/app/p3app.js:460-520, public/js/p3/app/p3app.js:560-585, public/js/p3/app/p3app.js:630-650

Conclusion

The DXKB web application architecture is a modular, service-oriented system that decouples UI components from analytical logic. By utilizing a central application controller (p3app.js) and standardized base classes (AppBase.js), the project maintains a consistent interface for diverse genomic tools while effectively managing distributed backend resources and secure user sessions.