Impress

This repository is for code connected with the Impress trial.

Running the pipeline after cloning

1. Add the raw data

The pipeline reads raw data from a single timestamped export folder from Viedoc, placed under data/raw/. Either:

• copy the existing _20260626_083406 export folder into data/raw/, so you end up with data/raw/_20260626_083406/, or
• download a fresh export from Viedoc and place it under data/raw/ the same way.

If you use a new export, update the export: value in config/cfg.yml to match the new folder's name (e.g. export: "_20260626_083406"), since that value tells the pipeline which subfolder of data/raw/ to read.

The biomarker Excel files used by the adlb target also need to be present under data/raw/biomarkers/.

2. Install dependencies with renv

This project pins package versions with renv (R 4.6.0). From the project root in R:

renv::restore()

This installs every package listed in renv.lock, including targets and tarchetypes, which run the pipeline itself.

3. Run the pipeline with targets

The pipeline is defined in _targets.R. To run it end-to-end:

targets::tar_make()

This builds every target in dependency order — importing and cleaning the raw data, building the ADaM/RDaM datasets, and rendering the final reports: reports/impress_statistical_analysis.docx, reports/impress_statistical_analysis.pdf, and the model-fit reports under reports/model fit/.

To build only a specific report (and its dependencies), pass its target name, e.g. targets::tar_make(report_pdf). To inspect the pipeline before running it, use targets::tar_visnetwork().