Skip to content

CHANGELOG.md

Latest commit

 

History

History
87 lines (66 loc) · 3.24 KB

CHANGELOG.md

File metadata and controls

87 lines (66 loc) · 3.24 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.2.3]

  • Update dependencies (recommendations from dependabot)
  • migrate toolchain to uv

[0.2.2] - 2026-04-10

Fixed

  • Fix regression when no segments.xml is present for the used EDIFACT directory. Print a warning instead of raising an exception.

[0.2.1] - 2026-03-28

Fixed

  • don't write trailing empty elements (colons) at end of composite data elements

[0.2.0] - UNRELEASED

CHANGES

  • BREAKING CHANGE: Remove SegmentCollection
  • BREAKING CHANGE: Remove FileSourcableMixin
  • BREAKING CHANGE: Segment() calls must provide the tag name now as the first parameter.
  • BREAKING CHANGE: Parser cannot override syntax_version and syntax_identifier anymore. It is determined from the UNB header.
  • change plugin system from metaclass to __init_subclass__
  • raise ValidationError, not AssertionError, when ServiceAdviceString is not 6 chars long
  • Drop support for Python versions < 3.10 (3.8 was EOL 2024-10-07, 3.9 not much used in Distros.)

ADDED

  • EDED, EDCD, EDMD, EDSD, UNCL, UNSL parsers and zip extraction logic for automated EDI schema generation
  • massively improve type annotations (thanks Tammo Ippen)
  • use code-based structures to validate segments and data elements
  • improve error output when EDIFACT syntax errors are found
  • Implement automatic detection and validation of EDIFACT syntax version (1 - 4) by examining the UNB segment for proper version control and adherence to version-specific syntax rules.
  • provide line numbers/columns in EdiFactSyntaxError exceptions

[0.1.9]

CHANGES

  • BREAKING CHANGE (possibly): Characters can be set up with reserved character now. If anyone used args instead of kwargs, this could lead to problems as the 2nd last parameter now is "reserved". But I suppose this will not harm anyone's implementations.

FIXES

  • Token creation with the wrong type fixed
  • raise Error when "escaped newline" occurs

[0.1.8]

  • allow parsing EDI files created with SAGE COALA (with a "header" before the UNA segment)

[0.1.7]

  • TODO

[0.1.6] - 2023-04-22

  • Fix Parser drops UNA Segments while converting tokens to segments #56
  • Apply fixes for Segmentfactory and Pluginmount #54

[0.1.5] - 2022-01-24

  • fix compatibility with Python <3.9

[0.1.4] - 2021-12-14

  • fix compatibility with Python 3.10

[0.1.3] - 2021-11-03

  • Fix multi-message interchange parsing (bug #43)
  • Fix correct handling of UNT segments (bug #44)
  • Add validation to AbstractSegmentsCollection

[0.1.2] - 2021-10-25

  • Fix parsing of colon at line endings
  • Fix multi-message interchange parsing

[0.1.1] - 2021-06-27

  • fix missing pydifact.syntax package

[0.1] - 2021-06-27

  • Add a split_by(tag_name) helper to manually split segment collections

[0.0.6] - 2021-04-06

  • Deprecate SegmentCollection in favour of Interchange
  • Add Interchange and Message semantic and featured classes
  • Add an optional predicate filter to get_segment and get_segments methods
  • Allow to access AbstractSegmentsContainer elements using container[i]

[0.0.5] - 2020-04-29

  • corrected pypi details

[0.0.4] - 2020-04-27

  • First release