Read or plot saved results

Start here when the analysis already exists. Read a full saved object, RDS, workbook or delivery directory before requesting another fit.

Read an existing delivery

readback <- iobrcausal(
  request = list(operation = "read"),
  inputs = list(path = "results/my_delivery")
)
print(readback)

A direct R reader is also available:

bundle <- iobr_causal_read_result_bundle("results/my_delivery")
iobr_causal_readout(bundle)

Directory discovery can be a bounded preview. Use the matching full numeric table and companion files before reporting a specific feature’s estimate.

Draw a figure from saved numbers

When saved_table is a single numeric effect table:

figure <- iobrcausal(
  request = list(operation = "plot", parameters = list(
    plot_type = "forest", formats = c("png", "pdf", "svg")
  )),
  inputs = list(result = saved_table), output_dir = "figures"
)

For a result object containing several endpoint tables, specify the matching table_path. The plot operation uses saved estimates and intervals. A survival curve additionally requires a saved curve cache; raw OS columns do not stand in for cached curves.

Check the exported source table and plot_provenance.csv alongside each image. Explicit Log-OR and Log-HR values may be exponentiated with their own intervals; additive effects remain additive. A display subset retains the original family.

Reuse and reproduction

Reuse a compatible saved result when the inputs and request are unchanged. An explicit reproduce request is different: it can execute a fresh association run using verified original inputs, seed and settings, then compare saved scientific tables. Frozen OS mediation has its own reuse rules; see advanced analysis.