dpm-synth 1.0.0

Generate synthetic EBA reports that satisfy DPM-XL validation rules.

Generate synthetic EBA xBRL-CSV report packages that satisfy a module's DPM-XL validation rules. Give dpm-synth a template with a module's tables, cells and rules; it fills every cell with a plausible value, repairs the values until the rules pass, and writes a report package that reads and validates like a real submission. Use it to test loaders, data warehouses and validators for EBA reports, for any module, without real supervisory data.

dpm-synth is an independent project, not affiliated with or endorsed by the European Banking Authority.

Installation

pip install dpm-synth   # Python 3.11+; installs dpm-xl and dpm-xbrl

Quickstart

from dpmxbrl import write_xbrl_csv

from dpmsynth import generate, load_template

template = load_template("minimal.json")  # a module's tables, cells and rules
result = generate(template, seed=7)
print(result.passed, "of", result.rules, "rules pass; failing:", result.failed)
write_xbrl_csv(result.report, "report.zip")

2 of 2 rules pass; failing: []

From the command line, with an exit status of 1 if some rules still fail:

$ dpm-synth generate minimal.json -o report.zip --seed 7
rules: 2, passed: 2, failed: 0, not evaluated: 0, not applicable: 0
rounds: 2 (failing rules per round: 1, 0)
wrote report.zip (3 facts)

The sample template minimal.json is in docs/samples; templates for real EBA modules are exported from a DPM release by any tool that writes the template format.

Features

• Any module: the dpm-synth-template/1 format describes tables, open keys, cells, data types, allowed items and DPM-XL rules.
• Rules satisfied: an evaluate-and-repair loop over the rules, with dpm-xl, repairs sums, differences, products, sign rules, isnull, in and match. On EBA DPM 4.3 COREP_OF, 999 of the 1,051 rules that apply pass, in 15 rounds; the result names the rest.
• Valid packages: written by dpm-xbrl, with a fictional LEI, and read back without broken EBA filing rules other than the file name.
• Reproducible: the same template, seed and versions give the same bytes.
• Large: scaled packages stream open tables with millions of rows.

Toolkit

Part of the dpm toolkit, three Python packages that work together:

Package	What it does
	Parse and evaluate DPM-XL validation rules
	Read, write and check xBRL-CSV and XBRL-XML reports
	Generate test reports that satisfy DPM-XL rules (this package)

The toolkit tour takes one report through all three: generated with dpm-synth, read with dpm-xbrl, then evaluated and explained with dpm-xl.

Documentation

Documentation: https://luukhopman.github.io/dpm-synth/

• Getting started
• How-to guides: write a template, build one from a DPM release, load-test, reproducible seeds, validate the output with dpm-xbrl and dpm-xl
• Concepts: the evaluate-and-repair loop and its limits
• Reference: the template format, the CLI and the Python API
• For AI agents: llms.txt and llms-full.txt
• Changelog, stability policy and security policy

Development

uv sync
uv run pytest
uv run ruff check
uv run ruff format --check
uv run ty check
uv run python scripts/build_docs.py

See contributing.

Licence

MIT. See LICENSE.