How-to guides¶

Recipes for specific tasks. They assume you have done the getting started tutorial; every example runs as shown (the test suite executes them).

Connect and evaluate

• Connect your data: write a DataProvider for your storage, including the optional open_keys, table_group, operation and earlier_report hooks.
• Evaluate a rule file: run a rule set against a report, in Python or with dpm-xl eval.
• Explain failures in a UI: sentences, filled formulas and the cells to highlight.
• From an XBRL report to explained failures: read a report with dpm-xbrl, bind it to a provider, evaluate and explain.
• Toolkit tour: generate a report with dpm-synth, read it with dpm-xbrl, then evaluate and explain its rules.

Scale up

• Evaluate large reports: the columnar engine for millions of open-table rows, with the reference fallback.
• Run rules in threads: what is safe to share.
• Compare with earlier reports: time_shift and annualise.

Build tooling

• Check rule syntax in CI: dpm-xl check and a pytest test.
• Build editor support: tokens, error ranges and the TextMate grammar.
• Analyse rules statically: the tree, the selections and the tables a rule reads.