edifactlib

A Python package for parsing, validating, and inspecting EDIFACT interchanges. It comes both with a CLI tool for inspecting EDIFACT files directly from the terminal, and with a regular Python library for use in code.

Installation

pip install edifactlib

Usage

As a Python library

from edifactlib import Parser, InterchangeResolver

message = """
UNA:+.? '
UNB+UNOC:3+5412345000013:14+4012345000006:14+260704:1200+REF00001'
UNH+1+ORDERS:D:24A:UN'
BGM+220+PO123456+9'
DTM+137:20260704:102'
NAD+BY+5412345000013::9'
NAD+SU+4012345000006::9'
LIN+1++4712345:EN'
QTY+21:100'
LIN+2++4798765:EN'
QTY+21:50'
UNS+S'
UNT+11+1'
UNZ+1+REF00001'
"""

parser = Parser()
interchange = parser.parse(message)

resolver = InterchangeResolver()
resolver.resolve(interchange)

Parser turns the raw EDIFACT message as an Interchange model (segments, data elements, components) and validates it against the corresponding EDIFACT directory by default (pass validate=False to parser.parse(...) to skip validation). InterchangeResolver then enriches the parsed data with the names/descriptions from the directory.

Command line

After installation, the edifact command is available.

Read and inspect an EDIFACT file:

edifact /path/to/message.edi

Alternatively, pass the message directly as a string:

edifact --from-string "UNA:+.? 'UNB+UNOC:3+...'"

The interchange header, contained messages, and functional groups are printed to the console in a formatted view:

Pass --print-json to print the parsed interchange as JSON instead, e.g. for piping into other tools:

edifact /path/to/message.edi --print-json

Supported EDIFACT versions and directories

EDIFACT version

Currently version 2 and 3 are supported. An implementation for version 4 is planned.

Directories

The following directories are currently supported:

• D.24A
• D.23A
• D.22B
• D.22A
• D.21B
• D.21A

Older directories will follow.

Transparency regarding AI-generated content

This project values transparency about which content was hand-written and which was created with AI assistance:

• Source code under src/ is mainly hand-written. Any AI-generated portions are explicitly marked as such.
• Exception: docstrings. Docstrings under src/ may be AI-generated and are not individually marked.
• Exception: the tests folder. Test code is created with AI assistance.
• Any AI-generated code within the tests folder is marked as such with a corresponding comment in the code.
• If a file in the tests folder does not contain such a marker comment, its content was written manually.
• The README.MD file was mostly generated by AI.

Development

Code is formatted with Black (line length 120) and imports are sorted with isort (Black profile). Install the dev dependencies with:

pip install -e ".[dev]"

Then format the code with:

black .
isort .

Status

This project is under active development. The API may still change before the first stable release.