pycuf¶

Read, check and analyse CUF-XML construction cost estimates in Python.

CUF-XML (Calculatie Uitwissel Formaat) is the Dutch exchange format for construction cost estimates (begrotingen). Estimating software such as Ibis, Kraan, Bakker & Spees CIVIEL and Matrix writes it; ERP systems such as AFAS, Exact and 4PS read it. pycuf reads every CUF-XML 4.x file into one typed model, computes its costs under an explicit, configurable policy, tells you precisely what is wrong with a file, and exports normalized tables.

import pycuf

cuf = pycuf.read("begroting.xml")
print(cuf.project.name, cuf.totals().estimate.total)
for finding in cuf.validate().findings:
    print(finding)

Features¶

• Zero runtime dependencies. The core uses only the standard library (pyexpat, decimal, csv, json) and is pure, fully typed Python. Arrow, polars, pandas, Parquet and the command line are optional extras.
• Made for real files. Windows-1252 and UTF-8, namespaces that are not valid URIs, empty required attributes, Dutch dates, decimal commas, vendor attributes (Ibis, Bakker & Spees) and missing sections: pycuf reads them and reports each one as a coded finding instead of refusing the file.
• Exact numbers. Every number is a decimal.Decimal with its original text kept next to it; computations run at 60 significant digits, so totals are exact.
• Explicit calculation rules. CUF-XML's rules are ambiguous in places. A policy makes every interpretation explicit, with presets for the official usage rules, the literal schema and ERP practice.
• Honest checks. Real files often carry wrong control totals. pycuf recomputes every bundle and compares, with tolerances you choose. See Validation.
• Normalized tables for CSV, JSONL, Arrow, polars, pandas, DuckDB and Parquet. See Interoperability.
• Safe with untrusted files. No DTDs, no entity expansion, no network access, size and depth limits. See Security.

Installation¶

pip install pycuf                 # core, no dependencies
pip install "pycuf[cli,polars]"   # with the command line and polars

Where to next¶

• Getting started: a ten-minute tutorial with the sample files.
• Reading CUF files: the data model, the raw layer, encodings, lenient parsing.
• Calculating: totals, multipliers, resource lines and policies.
• Validation and the finding codes.
• Interoperability and the command line.
• The CUF-XML format: history, the two specification texts and vendor dialects.
• Glossary: every Dutch element and attribute with its English name.

pycuf is an independent open-source project. It is not affiliated with or endorsed by Ketenstandaard Bouw en Techniek, Forum Systeemhuizen Bouw or any software vendor.