How mlambardi/trunclcdist works
Project Type and Purpose
• This is an open-source R package providing non-uniform random variate generation for truncated log-concave distributions.
• The core functionality is a generic acceptance–rejection sampler (rtruncLC()) plus built-in “plugins” for common distributions (normal, Poisson, binomial, etc.) and the ability for users to register their own.
High-Level Components
A. User API Layer
• Public functions (in R/):
– rtruncLC() – central entry point
– Distribution-specific wrappers (rtruncLCcontinuous.R, rtruncLCdiscrete.R)
– Utility helpers (e.g. truncLCdistoptions.R, inverseTransform.R)
• Vignettes (in vignettes/) – usage guides, examples
B. Core Engine
• Acceptance–rejection algorithm (Devroye’s method) implemented in R
• Interfaces to PMF/PDF (f), CDF (F), log-derivative (d log f), mode (m)
C. Distribution Modules
• Built-in specs under R/ (e.g. "norm", "pois", etc.)
• Extensibility point: user-supplied distribution definitions
D. Packaging & Documentation
• DESCRIPTION, NAMESPACE – CRAN metadata
• man/ – Rd documentation files auto-generated from roxygen comments
• inst/ – CITATION and REFERENCES.bib
E. Testing & CI
• tests/testthat/ – unit tests for core functions
• .github/ISSUE_TEMPLATE and pull-request workflows – contribution templates
F. Build Workflow
• R CMD build/check, devtools, roxygen2, R Markdown (vignettes)
Relationships & Data Flow
1. User calls rtruncLC(n, spec, …, a, b)
2. API layer parses arguments, sets up truncation bounds
3. Core engine loads distribution module (built-in or user-provided)
4. Core invokes PMF/PDF, CDF, log-derivative, mode functions in a loop
5. Acceptance–rejection logic generates variates until count n met
6. Results returned to user environment
7. (Offline) Documentation and vignettes guide the user
8. Tests validate each component; CI runs on push/PR
Architectural Patterns & Principles
• Plugin/Strategy pattern: each distribution implements the same interface (f, F, d log f, m).
• Modular layering: clear separation between API, core engine, and distribution implementations.
• Convention-over-configuration: CRAN standard layout (R/, man/, tests/, vignettes/, inst/).
Technologies & Frameworks
• R language & base packages
• testthat for unit testing
• roxygen2 for documentation
• R Markdown for vignettes
• GitHub Actions (workflow templates under .github/)
Essential Elements Checklist
• Notes on architectural patterns (plugin strategy)
• External systems (R console, CRAN, GitHub Actions)