v2.1.0 — Ecosystem Networks, i18n, Production Deployment

What's New in v2.1.0

Ecosystem Co-abundance Network (Major Feature)

• Interactive species co-abundance network built from the abundance matrix using pairwise Spearman correlations
• Louvain community detection partitions the network into ecological modules with modularity score
• Taxonomic coloring — hierarchical phylum→family color palette inspired by the SCAPIS ecosystem approach
• 4 layout algorithms: Organic (Fruchterman-Reingold with simulated annealing), Force-directed, Circle, Radial
• 3 color modes: Taxonomy, Module, Enrichment
• FBM overlay — annotate network nodes with Family of Best Models prevalence and coefficient direction
• Configurable prevalence and correlation thresholds, class filtering, module highlight on click

PCoA Analysis

• Principal Coordinates Analysis with Bray-Curtis, Euclidean, Jaccard, Cosine distance metrics
• 95% confidence ellipses per class with PERMANOVA (F, R², p-value, 999 permutations)
• FBM feature filtering from completed jobs

Internationalization (i18n)

• Full English and French translations (~450+ keys)
• All views translated: Settings, Data, Parameters, Results, Admin, Auth, User, Sharing

Production Deployment

• Docker Compose production stack with NGINX reverse proxy, SSL/TLS (Let's Encrypt), automated backups
• Helm chart for Kubernetes deployment
• API documentation exposed at /api/docs (Swagger) and /api/redoc

Real-time Progress & Notifications

• WebSocket streaming of console output with SVG sparkline progress chart
• Browser desktop notifications on job completion/failure

Additional Features

• Project archiving — archive/unarchive projects with visual indicator and filtering
• scitq integration — distributed computing via scitq task queue (configurable from Admin panel)
• Dataset metadata scanning — automatic feature detection with SVG icons and class name annotation
• Animated landing page — 4-step animated workflow with flowing dots, staggered reveal, hover glow
• Credits & Citations page — funding acknowledgments (ANR DeepIntegrOmics, GMT Science) and contributor list

Bug Fixes

• Fix public sharing endpoint tuple unpacking
• Regenerate frontend package-lock.json for vue-i18n compatibility

---

Full Changelog: v2.0.0...v2.1.0

v2.0.0 — Full-Featured Release

PredomicsApp v2.0.0

Major release with 26 features across 3 phases, comprehensive test suite, and full documentation.

Highlights

• Co-presence analysis — Pairwise feature co-occurrence with hypergeometric significance testing, Viridis heatmaps, network visualization (4 layout algorithms), and functional annotations from biobanks.gmt.bio
• Feature importance visualization — Waterfall charts, coefficient direction bars, per-sample contribution heatmaps
• Batch parameter sweeps — Launch up to 50 jobs with cartesian product of parameter grids
• Real-time job logs — WebSocket live streaming with HTTP polling fallback
• Multi-user security — API keys, webhooks, audit logging, password reset, rate limiting

Phase 1: Core Features (13 items)

• Export & Reports — CSV exports, HTML reports, Python/R notebook generation
• Job Comparison — Side-by-side metrics tables with config diff highlighting
• Real-time Progress — Console panel with ANSI-rendered logs, progress bar, generation tracking
• Notebook Integration — Pre-filled Python (.ipynb) and R (.Rmd) notebooks from completed jobs
• Onboarding Tour — 6-step interactive walkthrough for first-time users
• Browser Notifications — Desktop alerts when jobs complete or fail
• Batch Runs — Parameter sweep grid builder with up to 50 jobs per batch
• Dataset Tagging & Search — Tag-based organization with search and filter
• Animated Workflow Diagram — CSS-animated pipeline on landing page
• Use Case Examples — 4 real-world research examples with publication references
• Test Coverage — 215 backend + 59 frontend tests (~86% backend coverage)
• API Documentation — Swagger UI at /docs, ReDoc at /redoc
• Deployment Guide — Docker Compose, NGINX, SSL, Kubernetes, backup/restore

Phase 2: Enhancements (5 items)

• Database Backup & Import — Portable tar.gz archives with replace/merge restore modes
• WebSocket Live Logs — Real-time log streaming with JWT auth
• Dataset Preview — In-browser file preview with stats
• Performance — TTL cache, GZip compression, lazy-loaded Plotly.js
• Error Handling — Structured error responses, toast notifications, retry utility

Phase 3: Security & Integrations (8 items)

• Feature Importance Visualization — Waterfall, coefficient bars, per-sample heatmap
• Project Templates — Admin-managed parameter presets
• Audit Log — 14 tracked actions across 6 routers, admin-queryable
• Password Reset — Self-service with optional SMTP email
• API Key Management — Programmatic access with bcrypt-hashed keys
• Webhook Notifications — HMAC-SHA256 signed HTTP callbacks on job events
• Dataset Versioning — Auto-snapshots on file changes with restore
• Rate Limiting — Per-user/IP limits via slowapi (no Redis required)

Co-presence Analysis (new in this release)

• Feature prevalence across model population (FBM-filtered)
• Pairwise co-occurrence with Fisher's exact (hypergeometric) test
• Configurable significance threshold (p ≤ 0.01 to 0.20)
• Viridis-colored co-occurrence heatmap
• Network visualization with 4 layout algorithms (force-directed, circle, grid, radial)
• Functional annotations: butyrate production, inflammation, transit speed, oral origin
• MSP taxonomic annotations from biobanks.gmt.bio with concurrent fetching (20 threads)

Documentation

• Comprehensive README with 11 screenshots covering all app pages
• Playwright-based screenshot automation script
• Full API endpoint reference
• Architecture diagram and technology stack

Testing

• 215 backend tests (pytest) — ~86% coverage
• 59 frontend tests (Vitest + Vue Test Utils)
• CI/CD with GitHub Actions

Database Migrations

v2–v15: dataset library, composite datasets, admin flag, job name/user/config hash/disk size, dataset tags, batch_id, audit logs, password reset tokens, API keys, webhooks, dataset versions

Full Changelog: v1.3.0...v2.0.0