How igvteam/igv works
Project Type and Purpose
• IGV is a heavy-weight, desktop Java application for interactive genome visualization on Mac, Windows, and Linux.
• It also includes a command-line toolkit (“igvtools”) for batch processing genomic data and conversion tools, plus scripts for packaging and launching.
File Structure Analysis
• Top‐level:
• src/main/java – core Java packages for UI, data models, parsers, renderers, AWS support, session management, tools, etc.
• scripts/ – OS‐specific launchers (bash, bat, Mac .command), installer descriptors.
• lib_build/ – packaging dependencies for JRE bundling.
• examples/ – sample genome and OAuth config for testing.
• docs/ – user/developer documentation and web examples.
• build.gradle, gradlew – Gradle build scripts.
• src/test/java – automated unit and integration tests.
README Insights
• Desktop app built with Java 21; Gradle packaging produces an IGV distribution with embedded JREs, launcher scripts, and igvtools.
• Supports AWS S3 data access via OAuth/credentials.
• No central database; all data fed from local files or remote HTTP/S endpoints (UCSC, GA4GH, Htsget).
1. User Interface Layer
– Swing/AWT frames and panels (org.broad.igv.ui)
– Command Bar, Menus, Dialogs (track config, genome selection)
2. Core Engine
– Event Bus (IGVEventBus)
– Session Manager (load/save .xml sessions)
– Preferences Manager
– Track Manager (load/unload tracks)
3. Data Access & Parsing
– FileFormat Parsers (SAM, BAM, VCF, BigWig, TwoBit, GFF, BEDPE, etc.)
– DataSource Abstractions (local FS, HTTP, S3)
– AWS Client Module (org.broad.igv.aws)
4. Rendering Subsystem
– Renderers (feature tracks, alignments, heatmap, sashimi)
– Charting (scatter, bar, line)
– Graphics Utilities
5. Command‐Line Tools
– igvtools and converters (under org.broad.igv.tools)
6. Packaging & Launch Scripts
– Gradle distribution (build/IGV-dist)
– OS‐specific launchers (scripts/igv*.sh, .bat, Mac .command)
• UI → Core Engine (user actions dispatch events)
• Core Engine → Data Access (requests to load data)
• Data Access → Parsers → Data Models → Core Engine
• Core Engine → Rendering Subsystem → UI Canvas
• Core Engine ↔ Session Manager (persist state)
• Command‐Line Tools → Data Access → Parsers → Output Files
• AWS Client ↔ Amazon S3 (optional external data source)
c. Architectural Patterns & Principles
• MVC‐style separation: UI panels (View) bind to data models via event bus (Controller) leveraging data source abstractions (Model).
• Plugin/extension points: CLI plugins and OAuth providers loaded via resource descriptors.
• Single application monolith for desktop, with decoupled modules via Java packages and events.
d. Key Technologies
• Java 21, Swing/AWT for UI
• Gradle for build and packaging
• AWS SDK v1 for S3 integration
• JArgs for command‐line parsing
• ProGuard for packaging/minification