Skip to content

CLAUDE.md

Latest commit

 

History

History
109 lines (81 loc) · 3.24 KB

CLAUDE.md

File metadata and controls

109 lines (81 loc) · 3.24 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

ghedata is an R data package developed by Global Health Engineering at ETH Zurich for sharing data resources that document the group's work. It provides datasets suitable for research, teaching, and learning purposes.

Development Commands

Essential Commands

# Load development dependencies
library(devtools)
library(usethis)
library(pkgdown)

# Document the package (regenerate man/ files from roxygen comments)
devtools::document()

# Run R CMD check
devtools::check()

# Install the package locally for testing
devtools::install()

# Build the pkgdown website
pkgdown::build_site()

# Run data processing script to update datasets
source("data-raw/data_processing.R")

Testing and Development

# Load the package in development mode
devtools::load_all()

# Test specific dataset
library(ghedata)
data(people)
View(people)

# Check if documentation is properly built
?people

Architecture & Key Components

Data Processing Pipeline

  1. Raw Data Source: Google Sheets (accessed via googlesheets4)
  2. Processing Script: data-raw/data_processing.R
    • Pulls data from configured Google Sheets
    • Performs data cleaning and anonymization
    • Saves processed data to data/ as .rda files
    • Exports to inst/extdata/ as CSV/XLSX

Package Structure

  • R/: Dataset documentation files (e.g., people.R)

    • Each dataset has a corresponding R file with roxygen2 documentation
    • Documents include title, description, format, source, and examples

  • data-raw/: Contains processing scripts and intermediate data

    • Main processing happens in data_processing.R
    • Dictionary files define data schemas

  • inst/extdata/: Exported data files for download

    • Provides CSV and XLSX versions of all datasets
    • Accessible via system.file() calls

Documentation System

  • Uses roxygen2 for function/data documentation
  • pkgdown generates the package website from:
    • README.md (homepage)
    • Function documentation (reference)
    • Vignettes (articles)
    • NEWS.md (changelog)
  • Website configuration in _pkgdown.yml

Key Technical Decisions

  1. Data Anonymization: Names are replaced with unique hash IDs to protect privacy
  2. Multiple Export Formats: Data available as R objects, CSV, and Excel files
  3. Google Sheets Integration: Raw data pulled directly from shared sheets
  4. Versioning: Package version tracks data updates (currently 0.0.5)

Adding New Datasets

  1. Add raw data source to data-raw/data_processing.R
  2. Create documentation file in R/ (e.g., R/newdata.R)
  3. Run processing script to generate data files
  4. Update package documentation with devtools::document()
  5. Export to CSV/XLSX in processing script
  6. Update README with new dataset information
  7. Rebuild pkgdown site

Code Style

  • 2 spaces for indentation (no tabs)
  • Auto-append newlines to files
  • Strip trailing whitespace
  • Follow tidyverse style guide for R code

R Code Rules

  • Do not display results by using print() or message() functions

Markdown Rules

  • When writing markdown text, always add an empty row between a heading and the first paragraph or the first bullet of a list