Triage open/closed issues: parser fixes + opt-in transaction grouping

[wolph]

Goes through the open issues/PRs (and stale-closed ones) and implements everything safely fixable, without breaking backwards compatibility. Every parser change was verified against the full fixture set (parse → JSONEncoder → diff): zero change to existing output. New behaviour that would otherwise change output is opt-in.

Bug fixes

• Wrong entry_date year when :61: entry date crosses previous year #121 — entry_date now resolves across year boundaries (e.g. value date Jan 2026 + entry date Dec → 2025-12-30) instead of staying in the value-date year. guessed_entry_date kept as an alias.
• parser error #111 (GLS / Atruvia) — customer_reference may exceed 16 chars when a // bank reference follows; the 16x cap is kept when there's no //, so banks that pack same-line extra data (e.g. Rabobank) are unaffected.
• Unable parse Wise MT940 file #117 (Wise) — extra_details may exceed 34 chars.
• Minor parsing errors #109 — strip a dangling BIC/ IBAN label (a label with no value) from detail segments; the old '28D' branch was dead code (segment keys are always 2 chars). IBAN handling now actually works.

Robustness

• AttributeError: 'NoneType' object has no attribute 'encoding' #106 — regression test: parsing works when sys.stdout/sys.stderr are None (sandboxed/serverless). The dependency was already removed by the modernization; this locks it in.

New (opt-in, BC-safe) feature

• New transaction is only created on :61: tag (statement) and not :20: tag (transaction reference number) #110 — transaction_boundary option on mt940.parse() / Transactions(...): an iterable of tag slugs that each start a new transaction (e.g. {'transaction_reference_number'} to start one per :20:). Defaults to today's behaviour (only :61: starts a transaction). Documented in the README. This is the configurable "dialect"-style hook requested in the issue, rather than a default change (which previously broke ABN AMRO, transaction_reference only present in the first transaction. #26).

Verification

• Full tox green: py310–313, lint, pyright, mypy, pyrefly, docs, coverage 100%, codespell, repo-review.
• Fixture parse-diff harness: no change to any existing fixture output.

Fixes #121
Fixes #111
Fixes #117
Fixes #109
Fixes #110
Fixes #106

[gemini-code-assist]

Code Review

This pull request introduces several improvements and bug fixes to the MT940 parser, including an opt-in transaction_boundary feature to customize transaction grouping, support for longer customer references and supplementary details, and year-boundary correction for entry dates. It also fixes an issue where dangling 'BIC' or 'IBAN' labels are left in transaction details, and ensures parsing works in environments without stdout/stderr. The review feedback highlights two key improvements: first, ensuring that matching a boundary tag updates self.data with the tag's parsed result so subsequent transactions do not lose their references; second, adding a defensive check to handle a single string passed to transaction_boundary to prevent character-by-character iteration.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3600304003

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[Copilot]

Pull request overview

This PR addresses several MT940 parsing edge cases reported in issues (#121, #111, #117, #109, #106) and introduces an opt-in mechanism to alter transaction grouping via configurable “boundary” tag slugs (issue #110), while aiming to preserve existing default behavior.

Changes:

• Fix year-boundary resolution for entry_date in :61: parsing while keeping guessed_entry_date as a backward-compatible alias.
• Relax :61: parsing constraints to support long customer_reference (when followed by //) and long extra_details, plus strip dangling BIC/ IBAN labels from detail segments.
• Add transaction_boundary option to mt940.parse() / Transactions(...) and expand regression test coverage for the reported issues.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
README.md	Documents the new opt-in transaction grouping via transaction_boundary.
mt940/tags.py	Adjusts :61: regex and resolves entry_date across year boundaries.
mt940/processors.py	Strips dangling BIC / IBAN labels from detail segments.
mt940/parser.py	Plumbs transaction_boundary through the public parse() API.
mt940/models.py	Implements opt-in transaction boundaries during tag processing.
mt940_tests/test_processors_issue109.py	Adds end-to-end regression coverage for issue #109 segment handling.
mt940_tests/test_issues/test_long_references.py	Adds regression tests for issues #111 and #117 plus Rabobank BC behavior.
mt940_tests/test_issues/test_issue110_boundary.py	Adds tests for opt-in transaction boundary behavior.
mt940_tests/test_issues/test_issue106_streams.py	Adds regression test for parsing when sys.stdout/sys.stderr are None.
mt940_tests/test_entry_dates.py	Expands entry-date tests to cover forward/backward year boundary cases (#121).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.