Donate to Postulo/postulo

• https://buymeacoffee.com/tiagoagueda

---

See the documentation for more details on donation links.

Check and link company register numbers per country, offline (#935) #1022

Merged

tiagoagueda merged 2 commits from 935-register-links into main 2026-10-20 17:02:50 +00:00

Conversation 0 Commits 2 Files changed 91 +576 -196

tiagoagueda commented 2026-10-19 19:18:44 +00:00

Owner

• Copy link
• Copy Markdown

Refs #935 (the online half stays open, tracked in #1021)

What changed

The Company register number scheme (register), inside the identifiers module and still offline:

• Check digits per country, replacing the bare prefix check: France SIREN and SIRET (Luhn, with La Poste's exception), Czechia IČO, Norway organisasjonsnummer, Finland Y-tunnus, Estonia registrikood, Poland KRS (ten digits, shape only). Any other country is held to its two-letter prefix, as before. A failure is refused with a clear message.
• A link per country to the register's own page: FR (annuaire-entreprises, SIREN or SIRET), NO (Brønnøysund), EE (e-Äriregister), GB (Companies House), NL (KVK search), DK (Virk CVR), BE (KBO Public Search). Every other EU/EEA country leads to the EU business registers search (BRIS); a country BRIS does not cover (CH, US) has no link. Every number is reduced to digits (or letters and digits) before it goes into an address. Links are drawn by the existing identifier badge (rel="noopener noreferrer external").
• Scheme gains link_by, a function from the value to an address, for a link that depends on the value; other schemes are unchanged.
• Docs: the Tracking applications wiki page (committed locally to the wiki repo, not pushed).

Not in this PR: the LEI cross-check. An LEI's first four characters name its issuing LOU and not a country, so nothing sound can be said offline; it needs GLEIF's registeredAs and belongs to #1021.

URLs checked against the live registers: NO, EE, BE (page titles read), GB, NL, FR (reachable). Dropped as unverifiable: Finland (tietopalvelu.ytj.fi deep link not documented anywhere I could find), Czechia (the ARES web ?ico= parameter is undocumented; its REST API is not a page), Germany and Switzerland (no stable per-company address). DK is kept on the strength of the documented datacvr.virk.dk/enhed/virksomhed/<cvr> pattern; the site answers 403 to a script, so it was not opened here.

Library check

python-stdnum (LGPL-2.1+, pure Python, maintained) covers FR, CZ, NO, FI and EE, but not Poland's KRS, and ships some two hundred formats for six short sums. I wrote the six as small pure functions with their sources named, and ran them against python-stdnum on 20,000 random numbers each: they agree (the Estonian first-digit rule and the SIRET's SIREN rule were taken from it after the first run disagreed). Czech IČO differs only on numbers starting with 9, which stdnum rejects as a DIČ rule and the IČO algorithm does not.

Tests

tests/test_register_numbers.py: valid and invalid numbers per country (including wrong length and the La Poste and Estonian second-pass cases), every link, the BRIS fallback, no link outside it, hostile numbers never reaching an address, the badge markup, linked(). Full unit suite green (19460 passed).

Model use

Written with Claude Code, reviewed by the person who opens the PR.

🤖 Generated with Claude Code

Refs #935 (the online half stays open, tracked in #1021) ## What changed The *Company register number* scheme (`register`), inside the identifiers module and still offline: - **Check digits per country**, replacing the bare prefix check: France SIREN and SIRET (Luhn, with La Poste's exception), Czechia IČO, Norway organisasjonsnummer, Finland Y-tunnus, Estonia registrikood, Poland KRS (ten digits, shape only). Any other country is held to its two-letter prefix, as before. A failure is refused with a clear message. - **A link per country** to the register's own page: FR (annuaire-entreprises, SIREN or SIRET), NO (Brønnøysund), EE (e-Äriregister), GB (Companies House), NL (KVK search), DK (Virk CVR), BE (KBO Public Search). Every other EU/EEA country leads to the EU business registers search (BRIS); a country BRIS does not cover (CH, US) has no link. Every number is reduced to digits (or letters and digits) before it goes into an address. Links are drawn by the existing identifier badge (`rel="noopener noreferrer external"`). - `Scheme` gains `link_by`, a function from the value to an address, for a link that depends on the value; other schemes are unchanged. - Docs: the *Tracking applications* wiki page (committed locally to the wiki repo, not pushed). Not in this PR: the LEI cross-check. An LEI's first four characters name its issuing LOU and not a country, so nothing sound can be said offline; it needs GLEIF's `registeredAs` and belongs to #1021. URLs checked against the live registers: NO, EE, BE (page titles read), GB, NL, FR (reachable). Dropped as unverifiable: Finland (`tietopalvelu.ytj.fi` deep link not documented anywhere I could find), Czechia (the ARES web `?ico=` parameter is undocumented; its REST API is not a page), Germany and Switzerland (no stable per-company address). DK is kept on the strength of the documented `datacvr.virk.dk/enhed/virksomhed/<cvr>` pattern; the site answers 403 to a script, so it was not opened here. ## Library check python-stdnum (LGPL-2.1+, pure Python, maintained) covers FR, CZ, NO, FI and EE, but not Poland's KRS, and ships some two hundred formats for six short sums. I wrote the six as small pure functions with their sources named, and ran them against python-stdnum on 20,000 random numbers each: they agree (the Estonian first-digit rule and the SIRET's SIREN rule were taken from it after the first run disagreed). Czech IČO differs only on numbers starting with 9, which stdnum rejects as a DIČ rule and the IČO algorithm does not. ## Tests `tests/test_register_numbers.py`: valid and invalid numbers per country (including wrong length and the La Poste and Estonian second-pass cases), every link, the BRIS fallback, no link outside it, hostile numbers never reaching an address, the badge markup, `linked()`. Full unit suite green (19460 passed). ## Model use Written with Claude Code, reviewed by the person who opens the PR. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

tiagoagueda added 1 commit 2026-10-19 19:18:44 +00:00

Check and link company register numbers per country, offline ... 2bd655da93

The Company register number scheme only checked the two-letter prefix and
linked nowhere. A typo in a SIREN or an IČO went through, and the number
could not be opened in the register that issued it.

registers.py now holds, without any request, the public check-digit rules
for France (SIREN and SIRET, Luhn), Czechia (IČO), Norway
(organisasjonsnummer), Finland (Y-tunnus), Estonia (registrikood) and the
shape of Poland's KRS, each a small function with its source named, and the
register's own page for France, Norway, Estonia, the United Kingdom, the
Netherlands, Denmark and Belgium. Any other country the EU business
registers search covers links there; the others have no link. Scheme gains
link_by for an address that depends on the value.

The rules were checked against python-stdnum on 20,000 random numbers each
and agree; the library was not adopted as a dependency for six short sums.
No LEI cross-check is added: an LEI carries no country offline, so that
belongs to the online lookup, which stays open in the follow-up.

Refs #935

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

tiagoagueda referenced this pull request 2026-10-19 19:18:55 +00:00

Company register number: look it up in the country's register, through GLEIF, or link to the register #935

tiagoagueda force-pushed 935-register-links from 2bd655da93 to 69acb053ea 2026-10-19 19:26:30 +00:00 Compare

tiagoagueda force-pushed 935-register-links from 69acb053ea to 5a2c922c7e 2026-10-20 06:23:17 +00:00 Compare

tiagoagueda added 1 commit 2026-10-20 15:57:16 +00:00

Merge main into 935-register-links ... ea6231b85e

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

tiagoagueda merged commit 3d806ffdbf into main 2026-10-20 17:02:50 +00:00

tiagoagueda deleted branch 935-register-links 2026-10-20 17:02:50 +00:00

tiagoagueda referenced this pull request from a commit 2026-10-20 17:02:51 +00:00

Check and link company register numbers per country, offline (#935) (#1022)

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

accessibility

Usable by everyone, including people with disabilities: one of the project's stated missions

  

authentication

Signing in, and proving who you are

  

breaking change

Changes behaviour or data in a way operators must act on

  

bug

Something does not do what it says it does

  

documentation

The wiki, the docs, or the words in the interface

  

enhancement

An improvement or a new capability

  

interface

The shape of the pages: what is where, and how it reads

  

internationalisation

Languages, regions, and the formats and symbols that go with them

  

observability

What an operator can see about a running instance: metrics, logs, health

  

performance

Uses fewer resources, or is more reliable, with no loss of features or accessibility

  

security

Keeping the code and the data secure: one of the project's stated missions

  

tier

1

Easy: small, self-contained, no prerequisites

  

tier

2

High-severity security: account access, personal data, the boundaries

  

tier

3

Foundations: prerequisite work that unblocks other issues

  

tier/4

Later: large features and long-horizon work

No labels

accessibility

authentication

breaking change

bug

documentation

enhancement

interface

internationalisation

observability

performance

security

tier

1

tier

2

tier

3

tier/4

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

Postulo/postulo!1022

No description provided.