vehicles 0.4.1 → 0.8.0

data/data/plates/_decode/pt-trailer-services.yml

@@ -0,0 +1,96 @@
+# plates/_decode/pt-trailer-services.yml — the Portuguese TRAILER registration
+# prefixes: the issuing regional service, one or two letters.
+#
+# This is the one place geography survives on a Portuguese plate. Car, motorcycle
+# and moped numbers carry none at all — the national car series is purely
+# sequential — but a Portuguese trailer number is, verbatim from the statute,
+# "uma ou duas letras identificadoras do serviço regional que procedeu à
+# matrícula, seguidas de um número de ordem" (Regulamento de Matrícula, art. 4.º
+# n.º 1). The letters are the district office that issued it, not the trailer's
+# home district and not the towing vehicle's.
+#
+# THE TABLE IS A STATUTORY ANNEX, and it was AMENDED ONCE, which is why this
+# file carries two vintages instead of one:
+#
+#   * Decreto-Lei n.º 54/2005, de 3 de março — ANEXO I as originally enacted.
+#     22 rows. Viseu is "VS".
+#   * Decreto-Lei n.º 106/2006, de 8 de junho, art. 1.º n.º 3 — replaces
+#     Anexo I wholesale. Same 22 districts, same 21 codes, and Viseu is "VI".
+#
+# The Viseu change is recorded, not averaged, and not treated as a misprint:
+# both readings were taken from the Diário da República page images (the 2005
+# facsimile at p. 1896 and the 2006 facsimile at p. 4054), so both are what the
+# gazette printed. From the 2006 instrument's entry into force the operative
+# code is VI; VS is on the road on every trailer plated between 2005 and 2006
+# and on every earlier one issued under the same code. A parse API must accept
+# BOTH — which is exactly the "decode tables rot" hazard PRD-PLATES §9 names,
+# caught here inside a single fourteen-month window.
+#
+# WHAT THIS TABLE IS NOT: it is not a district-code table for cars. There has
+# never been one in the modern Portuguese system. And it is not the EXPORT
+# prefix set — export numbers take a single trailing letter from a different,
+# four-member list (L / P / A / M), which is carried inline on pt-export in
+# plates/pt.yml rather than here, because it is a different statutory list in a
+# different article with different members.
+#
+# Referenced by: pt-trailer (plates/pt.yml) via region_encoding.
+jurisdiction: pt
+topic: trailer-services
+authority:
+  name: "Regulamento do Número e Chapa de Matrícula (anexo ao Decreto-Lei n.º 54/2005, de 3 de março), Anexo I — na redação do Decreto-Lei n.º 106/2006, de 8 de junho"
+  url: "https://dre.pt/application/conteudo/346124"
+sources:
+  - "https://files.diariodarepublica.pt/1s/2006/06/111a00/40524060.pdf  # Decreto-Lei n.º 106/2006 (DR I-A n.º 111, 08-06-2006), art. 1.º n.º 3 + ANEXO I, p. 4054: the CURRENT 22-row table, with Viseu = VI. Locator: https://dre.pt/application/conteudo/346124 — accessed 2026-08-02"
+  - "https://files.diariodarepublica.pt/1s/2005/03/044a00/18921897.pdf  # Decreto-Lei n.º 54/2005 (DR I-A n.º 44, 03-03-2005), ANEXO I 'Tabela de dígitos identificadores dos serviços emissores de matrículas de reboques (referente ao n.º 2 do artigo 4.º)', p. 1896: the SUPERSEDED table, with Viseu = VS. Locator: https://dre.pt/application/conteudo/589934 — accessed 2026-08-02"
+period: { start: 2006 }
+period_evidence: instrument-in-force
+period_note: >-
+  2006 is the vintage of the list read here (Decreto-Lei n.º 106/2006, published
+  08-06-2006; no entry-into-force clause of its own, so the general 5-day
+  vacatio of Lei n.º 74/98 applies and the operative date is 13-06-2006 —
+  DERIVED from the default rule, not stated by the instrument, and flagged as
+  such). The CODES themselves are far older than either instrument: the
+  composition "letter(s) + serial" is drawn in the pre-1992 plate model
+  (DL 54/2005 Anexo II, Modelo V, "L-10978"), so trailers were numbered this way
+  for decades before any text reachable in this pass. The instrument that first
+  enacted the list was not located — recorded gap, and the reason this file's
+  period is an instrument vintage rather than a series start.
+maximal_munch: true
+maximal_munch_note: >-
+  Single-letter codes collide with the prefixes of two-letter ones and the
+  digits that follow give no separator help, so a parser MUST take the longest
+  matching code before the serial: A (Ponta Delgada) inside AV (Aveiro) and AN
+  (Angra do Heroísmo); C (Coimbra) inside CB (Castelo Branco); L (Lisboa) inside
+  LE (Leiria); P (Porto) inside PT (Portalegre). E (Évora), H (Horta) and M
+  (Funchal) are unambiguous.
+
+codes:
+  # Mainland districts (distritos), in the order the annex prints them.
+  - { code: AV, name: "Aveiro" }
+  - { code: BE, name: "Beja" }
+  - { code: BR, name: "Braga" }
+  - { code: BN, name: "Bragança" }
+  - { code: CB, name: "Castelo Branco" }
+  - { code: C,  name: "Coimbra" }
+  - { code: E,  name: "Évora" }
+  - { code: FA, name: "Faro" }
+  - { code: GD, name: "Guarda" }
+  - { code: LE, name: "Leiria" }
+  - { code: L,  name: "Lisboa" }
+  - { code: PT, name: "Portalegre" }
+  - { code: P,  name: "Porto" }
+  - { code: SA, name: "Santarém" }
+  - { code: SE, name: "Setúbal" }
+  - { code: VC, name: "Viana do Castelo" }
+  - { code: VR, name: "Vila Real" }
+  - { code: VI, name: "Viseu", superseded_code: VS, superseded_note: "VS in the 2005 Anexo I; VI since the 2006 replacement. Both were printed by the Diário da República; both are on the road." }
+  # Autonomous regions — the annex lists the Azores by ISLAND-GROUP district
+  # (three of them) and Madeira by its single district, under the same heading.
+  - { code: AN, name: "Angra do Heroísmo", kind: autonomous-region, region: "Açores" }
+  - { code: H,  name: "Horta",             kind: autonomous-region, region: "Açores" }
+  - { code: A,  name: "Ponta Delgada",     kind: autonomous-region, region: "Açores" }
+  - { code: M,  name: "Funchal",           kind: autonomous-region, region: "Madeira" }
+
+# Sanity note for the verifier: 22 rows, and the set is complete against
+# Portugal's administrative map — 18 mainland distritos + the 3 Azorean
+# distritos + Funchal. No code is duplicated; no district is missing.

data/data/plates/_decode/ro-counties.yml

@@ -0,0 +1,213 @@
+# plates/_decode/ro-counties.yml — Romania's registration-mark territorial codes.
+#
+# PROVENANCE, STATED FIRST AND WITHOUT DECORATION. Unlike
+# plates/_decode/it-provinces.yml and plates/_decode/at-districts.yml, whose rows
+# are enacted law read off Normattiva and RIS, THE 41 COUNTY ROWS BELOW ARE
+# `secondary-wikipedia`. The statutory annex exists and is named in `authority`;
+# it was NOT reachable in this pass, and every URL tried and every failure mode
+# is recorded in `primary_attempts` below so the next pass starts from a map.
+# Every row carries `evidence_tier: secondary-wikipedia` inline. No row in this
+# file claims `instrument-in-force`, and a reviewer who fingerprints this table
+# against Wikipedia SHOULD find a match — that is the declared provenance, not a
+# defect.
+#
+# THREE STRUCTURAL FACTS THIS FILE EXISTS TO CARRY
+#
+# 1. BUCHAREST IS NOT A COUNTY ROW, AND THAT IS WHY IT IS NOT IN THE COUNTY
+#    TABLE. Romania's territorial code is a ONE-letter code for the capital and a
+#    TWO-letter code for each of the 41 counties, and the two take DIFFERENT
+#    SERIAL FORMATS: a county plate carries two digits between the code and the
+#    three letters, Bucharest carries three. That is a real format fork inside
+#    one jurisdiction, it is modelled as separate SERIES in plates/ro.yml, and
+#    this file keeps `municipality` structurally apart from `counties` so a
+#    decoder cannot silently treat "B" as though it were a 42nd county row. The
+#    ordering and the split below are this dataset's, chosen to make the fork
+#    visible; they are not any source's arrangement.
+#
+# 2. THERE IS NO MAXIMAL-MUNCH PROBLEM, AND THAT IS WORTH SAYING OUT LOUD.
+#    Spain's pre-2000 province codes are one, two or three letters and a decoder
+#    must try longest-first (see plates/_decode/es-provinces.yml). Romania looks
+#    like it should have the same trouble — "B" is Bucharest while "BC", "BH",
+#    "BN", "BR", "BT", "BV" and "BZ" are all counties — but it does not, because
+#    the DIGIT BLOCK ALWAYS FOLLOWS THE CODE. In "B12ABC" the character after B
+#    is a digit, so the code is one letter; in "BC12ABC" it is a letter, so the
+#    code is two. The rule is mechanical and total, and it survives separator
+#    stripping (PRD-PLATES §2.6), which is exactly when a decoder needs it most.
+#
+# 3. THESE ARE NOT THE PRE-1992 CODES. The communist-era system used a different
+#    territorial letter set built on the administrative regions of the day — the
+#    much-repeated example is that the Craiova area ran under "OL" for Oltenia —
+#    and it also placed the letters in the MIDDLE of the mark rather than at the
+#    front. This table decodes the 1992-and-later mark only. Applying it to a
+#    pre-1992 Romanian plate will produce confident nonsense; ro-communist-1966
+#    in plates/ro.yml carries no reference to this file for that reason.
+jurisdiction: ro
+topic: counties
+authority:
+  name: "Regulamentul de aplicare a Ordonanţei de urgenţă a Guvernului nr. 195/2002 privind circulaţia pe drumurile publice, aprobat prin H.G. nr. 1.391 din 4 octombrie 2006 — the instrument whose annex carries the official list of indicative"
+  url: "https://legislatie.just.ro/Public/DetaliiDocument/79058"
+  note: >-
+    Named as the CARRIER of the official list, not as the source of the rows
+    below. Art. 23 of the Regulament composes the mark from "indicativul
+    judeţului sau al municipiului Bucureşti" plus an order number plus three
+    letters, and the indicative themselves are set out in the instrument's annex.
+    That annex was not retrieved — see `primary_attempts`.
+
+evidence:
+  tier: secondary-wikipedia
+  source: "https://en.wikipedia.org/wiki/Vehicle_registration_plates_of_Romania  # the 41 county code-to-name pairs and the Bucharest code, re-expressed in this schema's own shape and ordering. Accessed 2026-08-02"
+  what_is_secondary: "every code, every county name, and the count of 41 + 1"
+  what_is_not_secondary: >-
+    The COMPOSITION RULE — that the mark begins with the indicative of the county
+    or of the municipality of Bucharest — is the Regulament's own words and is
+    cited as such in plates/ro.yml. What is secondary here is the CONTENTS of the
+    list, not the fact that the law keeps one.
+
+primary_attempts:
+  - url: "https://legislatie.just.ro/Public/DetaliiDocument/79058"
+    result: "socket hang up, four separate attempts across the pass"
+  - url: "https://legislatie.just.ro/Public/DetaliiDocumentAfis/79058"
+    result: "socket hang up"
+  - url: "https://www.dsclex.ro/legislatie/2006/octombrie/hg1391_2006.htm"
+    result: "connection reset"
+  - url: "https://r.jina.ai/https://legislatie.just.ro/Public/DetaliiDocument/79058"
+    result: "HTTP 422 (also tried percent-encoded: HTTP 422)"
+  - url: "https://web.archive.org/web/2024/https://legislatie.just.ro/Public/DetaliiDocument/79058"
+    result: "fetch tool refuses web.archive.org"
+  - url: "https://www.legex.ro/Hotararea-1391-2006-76973.aspx"
+    result: "HOST IS REACHABLE — but that document id resolves to Ordinul 2374/2006 (civil aviation), not to HG 1391/2006. legex.ro is a live route for the next pass IF the correct document id is found; the id was not."
+primary_attempts_note: >-
+  Search-engine assistance was unavailable for the second half of this pass: the
+  WebSearch budget was exhausted, DuckDuckGo's HTML endpoint moved to a CAPTCHA,
+  Mojeek returned 403 and SearXNG returned a browser check. The correct legex.ro
+  document id is therefore the single cheapest next step for whoever picks this
+  up.
+
+period: { start: 1992 }
+period_evidence: secondary-wikipedia
+period_note: >-
+  The code set below belongs to the system introduced in 1992. Two period
+  cautions a decoder must carry. (1) LOWER BOUND: these codes do not decode
+  pre-1992 marks at all (header fact 3). (2) STABILITY: Romania's county
+  boundaries have not been re-cut since 1968, so unlike the Italian province
+  sigle (four withdrawn in 2017) and the German district codes (reinstated in
+  waves), this table has no known withdrawn rows and no known period-closed rows.
+  That is an ABSENCE OF EVIDENCE of churn rather than sourced evidence of
+  stability, and it is recorded as the former.
+
+iso_3166_2_correspondence:
+  status: UNVERIFIED
+  claim_not_made: >-
+    The 41 two-letter codes below LOOK like the subdivision codes of ISO
+    3166-2:RO with the "RO-" prefix removed, and if that correspondence held it
+    would give this table a second, non-Wikipedia, structured route. THIS FILE
+    DOES NOT RELY ON IT AND DOES NOT ASSERT IT. The ISO register was not fetched
+    in this pass, so the correspondence is an unchecked hypothesis, and an
+    unchecked hypothesis dressed as corroboration would be worse than no
+    corroboration at all — two sources that were never actually compared are one
+    source with a false witness. Recorded here as the FIRST task for a verifier:
+    fetch the ISO 3166-2:RO register, diff it against these rows, and either
+    upgrade the tier or record where they diverge.
+
+# ---------------------------------------------------------------------------
+# THE CAPITAL — one letter, three digits, its own series in plates/ro.yml.
+# ---------------------------------------------------------------------------
+municipality:
+  code: B
+  name: "București"
+  name_en: Bucharest
+  type: "municipiu (the capital, which is not a județ)"
+  serial_digits: 3
+  serial_digits_history:
+    - { digits: 2, note: "the original 1992 form, B 99 LLL — see ro-bucharest-1992 / ro-bucharest-2007" }
+    - { digits: 3, note: "adopted when the two-digit space was exhausted — see ro-bucharest-2010" }
+  evidence_tier: secondary-wikipedia
+  note: >-
+    THE FORMAT FORK. Every other territorial code in Romania takes two digits;
+    Bucharest takes three. The reason is capacity, not law: art. 23 of the
+    Regulament prescribes "numărul de ordine, format din cifre arabe" and fixes
+    NO number of digits at all, so the capital was able to lengthen its order
+    number without any amendment. That is why plates/ro.yml carries a
+    `regex_statutory` wider than its `regex` for the standard series — the
+    statute is genuinely wider than what is issued, and Bucharest is the proof.
+
+# ---------------------------------------------------------------------------
+# THE 41 COUNTIES — ordered by CODE, this dataset's ordering. Two letters each,
+# two digits in the serial, without exception.
+# ---------------------------------------------------------------------------
+counties_note: "every row: { code, name (Romanian, with diacritics), evidence_tier }. serial_digits is 2 for all 41 and is stated once here rather than repeated per row."
+counties_serial_digits: 2
+counties:
+  - { code: AB, name: "Alba",              evidence_tier: secondary-wikipedia }
+  - { code: AG, name: "Argeș",             evidence_tier: secondary-wikipedia }
+  - { code: AR, name: "Arad",              evidence_tier: secondary-wikipedia }
+  - { code: BC, name: "Bacău",             evidence_tier: secondary-wikipedia }
+  - { code: BH, name: "Bihor",             evidence_tier: secondary-wikipedia }
+  - { code: BN, name: "Bistrița-Năsăud",   evidence_tier: secondary-wikipedia }
+  - { code: BR, name: "Brăila",            evidence_tier: secondary-wikipedia }
+  - { code: BT, name: "Botoșani",          evidence_tier: secondary-wikipedia }
+  - { code: BV, name: "Brașov",            evidence_tier: secondary-wikipedia }
+  - { code: BZ, name: "Buzău",             evidence_tier: secondary-wikipedia }
+  - { code: CJ, name: "Cluj",              evidence_tier: secondary-wikipedia }
+  - { code: CL, name: "Călărași",          evidence_tier: secondary-wikipedia }
+  - { code: CS, name: "Caraș-Severin",     evidence_tier: secondary-wikipedia }
+  - { code: CT, name: "Constanța",         evidence_tier: secondary-wikipedia }
+  - { code: CV, name: "Covasna",           evidence_tier: secondary-wikipedia }
+  - { code: DB, name: "Dâmbovița",         evidence_tier: secondary-wikipedia }
+  - { code: DJ, name: "Dolj",              evidence_tier: secondary-wikipedia }
+  - { code: GJ, name: "Gorj",              evidence_tier: secondary-wikipedia }
+  - { code: GL, name: "Galați",            evidence_tier: secondary-wikipedia }
+  - { code: GR, name: "Giurgiu",           evidence_tier: secondary-wikipedia }
+  - { code: HD, name: "Hunedoara",         evidence_tier: secondary-wikipedia }
+  - { code: HR, name: "Harghita",          evidence_tier: secondary-wikipedia }
+  - { code: IF, name: "Ilfov",             evidence_tier: secondary-wikipedia }
+  - { code: IL, name: "Ialomița",          evidence_tier: secondary-wikipedia }
+  - { code: IS, name: "Iași",              evidence_tier: secondary-wikipedia }
+  - { code: MH, name: "Mehedinți",         evidence_tier: secondary-wikipedia }
+  - { code: MM, name: "Maramureș",         evidence_tier: secondary-wikipedia }
+  - { code: MS, name: "Mureș",             evidence_tier: secondary-wikipedia }
+  - { code: NT, name: "Neamț",             evidence_tier: secondary-wikipedia }
+  - { code: OT, name: "Olt",               evidence_tier: secondary-wikipedia }
+  - { code: PH, name: "Prahova",           evidence_tier: secondary-wikipedia }
+  - { code: SB, name: "Sibiu",             evidence_tier: secondary-wikipedia }
+  - { code: SJ, name: "Sălaj",             evidence_tier: secondary-wikipedia }
+  - { code: SM, name: "Satu Mare",         evidence_tier: secondary-wikipedia }
+  - { code: SV, name: "Suceava",           evidence_tier: secondary-wikipedia }
+  - { code: TL, name: "Tulcea",            evidence_tier: secondary-wikipedia }
+  - { code: TM, name: "Timiș",             evidence_tier: secondary-wikipedia }
+  - { code: TR, name: "Teleorman",         evidence_tier: secondary-wikipedia }
+  - { code: VL, name: "Vâlcea",            evidence_tier: secondary-wikipedia }
+  - { code: VN, name: "Vrancea",           evidence_tier: secondary-wikipedia }
+  - { code: VS, name: "Vaslui",            evidence_tier: secondary-wikipedia }
+
+# ---------------------------------------------------------------------------
+# Decoder cautions.
+# ---------------------------------------------------------------------------
+decoder_cautions:
+  - id: place-of-registration-not-residence
+    text: >-
+      The Romanian indicative denotes where the vehicle is REGISTERED. It is
+      therefore semantically unlike the Italian right-hand sticker (the holder's
+      province of RESIDENCE, and legally not part of the plate at all) and unlike
+      the French SIV département band (owner-CHOSEN, decoding to nothing). A
+      Romanian indicative is real geography. It is still not a claim about where
+      the vehicle IS, only about where it was put on the register.
+  - id: ilfov-is-not-bucharest
+    text: >-
+      IF (Ilfov) is the county that SURROUNDS Bucharest and is a separate
+      indicative with the ordinary two-digit format. A decoder that folds IF into
+      B, or renders IF as "Bucharest area", is wrong on both the code and the
+      format.
+  - id: no-code-is-a-word
+    text: >-
+      Unlike Italy, whose Rome sigla is the literal word "Roma" and forces a
+      special case into every parser, every Romanian code here is exactly one or
+      two letters. There is no four-letter exception and no mixed-case row.
+  - id: three-letter-group-is-not-a-code
+    text: >-
+      The THREE letters at the end of a Romanian mark are a serial group, not a
+      territorial code, even when they spell a plausible abbreviation. Certain
+      combinations are reported to be reserved for institutions (POL, DEP, SNT,
+      SRI, GUV); those are reservations inside the serial space and are recorded
+      in plates/ro.yml, not here, because they decode to an INSTITUTION and not
+      to a place.

data/data/plates/_decode/se-diplomatic.yml

@@ -0,0 +1,220 @@
+# plates/_decode/se-diplomatic.yml — Swedish beskickningsskylt country codes.
+#
+# WHY THIS IS A _decode FILE AND NOT AN INLINE TABLE (PRD-PLATES §2.4): 128
+# rows is past the "small" threshold that put FZV Anlage 2 inline in de.yml,
+# and it is a CLOSED, PRIMARY list — Transportstyrelsen publishes the whole
+# thing, so it qualifies on the other limb of the rule ("a reference … if the
+# full table is primary-sourced and closed").
+#
+# THE SOURCE MOVED, AND THAT MATTERS FOR PROVENANCE. Until 2 June 2025 this
+# table WAS a regulation annex: TSFS 2015:63 Bilaga 1 ("Landskod enligt
+# 12 kap. 3 § 1") and Bilaga 2 ("Kategorikod enligt 12 kap. 3 § 3"). TSFS
+# 2025:33 repealed both annexes and rewrote 12 kap. 3 § so that the codes are
+# no longer enumerated in the föreskrift at all — they now live only on
+# Transportstyrelsen's website. Both are cited below: the website for the
+# CURRENT list, the repealed annex for the fact that this used to be enacted
+# text and for the pre-2025 kategorikod set (which had no "J").
+#
+# DECODE DISCIPLINE: the assigning authority is NOT Transportstyrelsen. 11 kap.
+# 2 § förordningen (2019:383): "Sammansättningen av tecken på skyltar för
+# beskickningsfordon bestäms av Regeringskansliet (Utrikesdepartementet) med
+# hänsyn till det berörda landet och den ställning som den som äger eller
+# använder fordonet har." Missions open and close; a code withdrawn from this
+# list does not become undecodable for plates already issued, so rows are
+# closed by period and never deleted — the same rule de.yml records for German
+# district codes.
+#
+# ALPHABET OBSERVATION (a fact about the table, not a rule): across all 128
+# codes the letters I, O and Q never appear, while V does (EV = IDEA). That is
+# the opposite of the ordinary registration-number alphabet, which excludes
+# I, Q, V, Å, Ä and Ö. No primary states a letter rule for landskoder; the
+# observation is recorded so a parse API does not infer one.
+jurisdiction: se
+topic: diplomatic
+authority:
+  name: Regeringskansliet (Utrikesdepartementet) — codes; Transportstyrelsen — publication
+  url: "https://www.transportstyrelsen.se/sv/vagtrafik/fordon/aga-kopa-eller-salja-fordon/registreringsskyltar/skylt-for-beskickningsfordon-diplomatfordon/"
+sources:
+  - "https://www.transportstyrelsen.se/sv/vagtrafik/fordon/aga-kopa-eller-salja-fordon/registreringsskyltar/skylt-for-beskickningsfordon-diplomatfordon/  # the current 128-row landskod table and the kategorikod table (A-H, 'I eller J'); accessed 2026-08-02"
+  - "https://www.transportstyrelsen.se/TSFS/TSFS%202015_63.pdf  # TSFS 2015:63 as enacted, Bilaga 1 'Landskod enligt 12 kap. 3 § 1' and Bilaga 2 'Kategorikod enligt 12 kap. 3 § 3' (kategorikod A-I, no J) — the version in force 2016-01-01; accessed 2026-08-02"
+  - "https://www.transportstyrelsen.se/TSFS/TSFS%202025_33.pdf  # TSFS 2025:33: 'dels att 3 kap. 17 §, 6 kap. 14 §, 12 kap. 4 § samt bilaga 1 och 2 ska upphöra att gälla' — in force 2 June 2025, which is when the table left the regulation; accessed 2026-08-02"
+  - "https://data.riksdagen.se/dokument/sfs-2019-383.html  # 11 kap. 2 §: the character composition is set by Regeringskansliet (Utrikesdepartementet); accessed 2026-08-02"
+period: { start: 2025 }
+period_evidence: instrument-in-force
+period_note: >-
+  2025 is the date of the SOURCE now read (the Transportstyrelsen web table,
+  which replaced TSFS 2015:63 Bilaga 1 on 2 June 2025), not the date any code
+  was first assigned. The beskickningsskylt regime itself is far older —
+  förordningen (1988:964) om skyltar för beskickningsfordon took effect on
+  1 October 1988 — and the per-mission code assignments are Utrikesdepartementet
+  decisions that are not individually dated in any published source. Treat every
+  row as "in force when read"; do not infer an assignment year from it.
+
+# ---------------------------------------------------------------------------
+# LANDSKOD — two letters, the mission. 128 rows, verbatim from the authority's
+# table, keyed by code (the table is printed mission-first; keying by code is
+# the decode direction a parse API needs).
+# ---------------------------------------------------------------------------
+landskod:
+  AA: "Sydafrikas ambassad"
+  AB: "Albaniens ambassad"
+  AC: "Algeriets ambassad"
+  AD: "Tysklands ambassad"
+  AE: "Armeniens ambassad"
+  AF: "Amerikas förenta staters ambassad"
+  AG: "Angolas ambassad"
+  AH: "Saudiarabiens ambassad"
+  AJ: "Argentinas ambassad"
+  AK: "Australiens ambassad"
+  AL: "Österrikes ambassad"
+  AM: "Bangladesh ambassad"
+  AN: "Belgiens ambassad"
+  AP: "Bolivias ambassad"
+  AR: "Botswanas ambassad"
+  AS: "Brasiliens ambassad"
+  AT: "Bulgariens ambassad"
+  AU: "Canadas ambassad"
+  AW: "Chiles ambassad"
+  AX: "Folkrepubliken Kinas ambassad"
+  AY: "Colombias ambassad"
+  AZ: "Republiken Koreas ambassad"
+  BA: "Demokratiska folkrepubliken Koreasambassad"
+  BB: "Afghanistans ambassad"
+  BC: "Cubas ambassad"
+  BD: "Danmarks ambassad"
+  BE: "Dominikanska republikens ambassad"
+  BF: "Egyptens ambassad"
+  BG: "Ecuadors ambassad"
+  BH: "Spaniens ambassad"
+  BJ: "Etiopiens ambassad"
+  BK: "Finlands ambassad"
+  BL: "Frankrikes ambassad"
+  BM: "Storbritanniens ambassad"
+  BN: "Greklands ambassad"
+  BP: "Guatemalas ambassad"
+  BR: "Kazakstans ambassad"
+  BS: "Ungerns ambassad"
+  BT: "Indiens ambassad"
+  BU: "Indonesiens ambassad"
+  BW: "Iraks ambassad"
+  BX: "Irans ambassad"
+  BY: "Irlands ambassad"
+  BZ: "Islands ambassad"
+  CA: "Israels ambassad"
+  CB: "Italiens ambassad"
+  CC: "Japans ambassad"
+  CD: "Kenyas ambassad"
+  CE: "Laos ambassad"
+  CF: "Libanons ambassad"
+  CG: "Libyens folkkontor"
+  CH: "Malaysias ambassad"
+  CJ: "Marockos ambassad"
+  CK: "Mexikos ambassad"
+  CL: "Moçambiques ambassad"
+  CM: "Nicaraguas ambassad"
+  CN: "Nigerias ambassad"
+  CP: "Norges ambassad"
+  CR: "Qatars ambassad"
+  CS: "Pakistans ambassad"
+  CT: "Panamas ambassad"
+  CU: "Nederländernas ambassad"
+  CW: "Perus ambassad"
+  CX: "Filippinernas ambassad"
+  CY: "Polens ambassad"
+  CZ: "Portugals ambassad"
+  DA: "Rumäniens ambassad"
+  DB: "Senegals ambassad"
+  DD: "Sri Lankas ambassad"
+  DE: "Schweiz ambassad"
+  DF: "Tanzanias ambassad"
+  DG: "Tjeckiens ambassad"
+  DH: "Thailands ambassad"
+  DJ: "Tunisiens ambassad"
+  DK: "Turkiets ambassad"
+  DL: "Rysslands ambassad"
+  DM: "Uruguays ambassad"
+  DN: "Venezuelas ambassad"
+  DP: "Vietnams ambassad"
+  DR: "Serbiens ambassad"
+  DS: "Zambias ambassad"
+  DT: "Zimbabwes ambassad"
+  DU: "FN:s flyktingkommissariat"
+  DW: "Världssjöfartsuniversitet"
+  DX: "Nordiska rådets presidiesekretariat"
+  DY: "Namibias ambassad"
+  DZ: "Sudans ambassad"
+  EA: "Europeiska rymdorganisationen (ESA)"
+  EB: "Europeiska Unionen (EU)"
+  EC: "Burundis ambassad"
+  ED: "Estlands ambassad"
+  EE: "Lettlands ambassad"
+  EF: "Litauens ambassad"
+  EG: "Kroatiens ambassad"
+  EH: "Sloveniens ambassad"
+  EJ: "Slovakiens ambassad"
+  EK: "Bosnien-Hercegovinas ambassad"
+  EL: "Eritreas ambassad"
+  EM: "Ukrainas ambassad"
+  EN: "Cyperns ambassad"
+  EP: "Makedoniens ambassad"
+  ER: "Kuwaits ambassad"
+  ES: "Honduras ambassad"
+  ET: "Rwandas ambassad"
+  EV: "International Institute for Democracy and Electoral Assistance (IDEA)"
+  EX: "Kap Verdes ambassad"
+  EY: "Östersjöstaternas råds sekretariat (CBSS)"
+  FA: "Vitrysslands ambassad"
+  FB: "Demokratiska Republiken Kongos ambassad"
+  FC: "El Salvadors ambassad"
+  FD: "Syriens ambassad"
+  FE: "GIWA (Global International WatersAssessment)"
+  FF: "Apostoliska nuntiaturen"
+  FG: "European Commission/RepresentationIn Sweden"
+  FH: "GWPO (Global Water PartnershipOrganisation)"
+  FJ: "Förenade Arabemiratens ambassad"
+  FK: "Moldaviens ambassad"
+  FL: "Georgiens ambassad"
+  FM: "Azerbajdzjans ambassad"
+  FN: "Nya Zeelands ambassad"
+  FP: "Mongoliets ambassad"
+  FS: "Republiken Kongos ambassad"
+  FT: "Kosovos ambassad"
+  FU: "Paraguays ambassad"
+  FW: "Palestinas ambassad"
+  FX: "Europeiska investeringsbanken"
+  FY: "Sekretariatet för Nordliga dimensionens partnerskap för hälsa och socialt välbefinnande (NDPHS)"
+  FZ: "Uzbekistans ambassad"
+
+# ---------------------------------------------------------------------------
+# KATEGORIKOD — one letter, the holder's standing. Note the I/J pair: the
+# repealed TSFS 2015:63 Bilaga 2 listed only "I  tjänsteman vid internationell
+# organisation"; the current web table reads "I eller J", so J is an addition
+# made after the annex was written. Recorded as two rows with the same meaning
+# rather than as the string "I eller J", so the table is machine-usable.
+# ---------------------------------------------------------------------------
+kategorikod:
+  A: "beskickningschef"
+  B: "ambassad"
+  C: "diplomat"
+  D: "administrativ/teknisk tjänsteman"
+  E: "karriärkonsulat"
+  F: "karriärkonsul"
+  G: "konsulattjänsteman"
+  H: "internationell organisation"
+  I: "tjänsteman vid internationell organisation"
+  J: "tjänsteman vid internationell organisation"
+kategorikod_note: >-
+  TSFS 2015:63 Bilaga 2 as enacted (in force 2016-01-01) ended at I. The
+  current Transportstyrelsen table prints "I eller J" against the single
+  meaning "tjänsteman vid internationell organisation". No instrument dating
+  J's addition was found — recorded gap.
+
+# The middle group is a THREE-DIGIT löpnummer allocated inside each landskod
+# (TSFS 2015:63 12 kap. 3 § 2: "ett löpnummer (tre siffror) inom respektive
+# landskod"). It is a per-mission sequence with no published meaning, so there
+# is nothing to decode below the landskod + kategorikod pair.
+lopnummer:
+  digits: 3
+  scope: "per landskod"
+  decodes_to: null
+  source: "https://www.transportstyrelsen.se/TSFS/TSFS%202015_63k.pdf  # 12 kap. 3 §, consolidated; accessed 2026-08-02"

data/data/plates/_decode/si-regions.yml

@@ -0,0 +1,123 @@
+# plates/_decode/si-regions.yml — Slovenian registracijska območja, and the
+# thing that actually makes a Slovenian plate decodable: THE MUNICIPAL ARMS.
+#
+# TWO LEVELS, and the second one is the interesting one.
+#
+# LEVEL 1 — the two-letter code. Only ELEVEN exist, and both halves of the fact
+# are primary: 29. člen (1) of the Pravilnik o registraciji motornih in
+# priklopnih vozil (Uradni list RS, št. 48/11) names the eleven registration
+# areas and, for each, the upravne enote it covers; 29. člen (2) delegates the
+# LETTERS to Priloga IV ("Oznake registracijskih območij, ki se uporabljajo na
+# registrskih in preskusnih tablicah, so navedene v seznamu oznak
+# registracijskih območij v Prilogi IV tega pravilnika"), which is a one-page
+# table printed with the regulation. Both are reproduced below, in the
+# regulation's own numbering — which is the same in the article and the annex.
+#
+# LEVEL 2 — the coat of arms, which is NOT decoration. 28. člen (1), verbatim:
+# "Za oznako registracijskega območja je grb občine, kjer je sedež upravne
+# enote v tem registracijskem območju." The shield between the code and the
+# serial is the arms of the MUNICIPALITY THAT HOSTS THE SEAT OF THE UPRAVNA
+# ENOTA — not of the registration area. So thirteen visually different plates
+# all read "LJ", and the shield is the only thing that tells Kočevje from
+# Domžale from Ljubljana. A Slovenian plate therefore carries strictly more
+# geography than its letters do, and a text-only parse throws it away.
+# 28. člen (3) adds that the arms must be adapted graphically to the plate.
+#
+# WHAT THE ARMS BIND TO, and when. The 2011 pravilnik states the rule
+# positionally and leaves the binding to the registration rules; its 2004
+# predecessor said it outright, and the sense has not changed —
+# Uradni list RS, št. 29/04, 22. člen (1), verbatim: "se uporabi grb mesta, v
+# katerem je sedež organa, ki vodi vozilo v evidenci in v kateri ima lastnik
+# vozila, OB IZDAJI REGISTRSKIH TABLIC, svoje stalno prebivališče oziroma
+# začasno prebivališče". The arms are fixed AT ISSUE, to the then-owner's
+# residence. Same caution as the Croatian city code: decode to
+# residence-at-issue, never to "where this car is now".
+#
+# NOT A COMPLETE ARMS CATALOGUE. This file names the upravne enote whose seat
+# municipality supplies each shield; it does NOT carry the 58 blazons, their
+# artwork or their licensing. Municipal arms are protected independently of
+# copyright and PRD-PLATES §3 puts them in the placeholder slot
+# (`design.emblem.rendered: placeholder` in plates/si.yml). Recorded gap:
+# a per-upravna-enota arms ledger.
+jurisdiction: si
+topic: regions
+authority:
+  name: "Pravilnik o registraciji motornih in priklopnih vozil, 29. člen in Priloga IV — Ministrstvo za infrastrukturo / upravne enote"
+  url: "https://www.uradni-list.si/glasilo-uradni-list-rs/vsebina/2011-01-2292"
+sources:
+  - "https://www.uradni-list.si/glasilo-uradni-list-rs/vsebina/2011-01-2292  # Uradni list RS, št. 48/2011 z dne 24. 6. 2011 — 29. člen (1) the eleven areas with their upravne enote, (2) the delegation to Priloga IV; 28. člen (1) the grb občine rule; 58. člen 'Ta pravilnik začne veljati 4. julija 2011.' Accessed 2026-08-02"
+  - "https://www.uradni-list.si/files/RS_-2011-048-02292-OB~P004-0000.PDF  # PRILOGA IV, 'REGISTRACIJSKA OBMOČJA — Seznam oznak registracijskih območij', the eleven code letters, read from the gazette PDF. Accessed 2026-08-02"
+  - "https://www.uradni-list.si/glasilo-uradni-list-rs/vsebina/2004-01-1253/pravilnik-o-registrskih-tablicah-motornih-in-priklopnih-vozil  # Uradni list RS, št. 29/2004 z dne 26. 3. 2004, 22. člen (1) — the arms bind to the owner's residence AT PLATE ISSUE. Accessed 2026-08-02"
+period: { start: 2011 }
+period_evidence: enacted-commencement
+period_note: >-
+  2011 is the commencement of the instrument these rows are printed in
+  (58. člen: "Ta pravilnik začne veljati 4. julija 2011."), not the age of the
+  codes. The ELEVEN-AREA system is much older — the 2006 predecessor
+  (Uradni list RS, št. 83/06, 3. člen) lists the same eleven areas with the
+  same upravne enote — and the two-letter codes themselves go back to the first
+  independent-Slovenia instrument (Uradni list RS, št. 5/92, 14. člen, whose
+  own seznam was not fetched in this pass). Per-code start dates are therefore
+  a RECORDED GAP, and 1992-era codes withdrawn since would not appear here.
+maximal_munch: false
+maximal_munch_note: >-
+  Not needed. Slovenian area codes are all exactly two letters (Priloga IV has
+  no one-letter entry), and 30. člen separates the area code from the serial by
+  a shield rather than by ambiguity. The prefix is the first two characters,
+  full stop — unlike Austria, where the split is genuinely ambiguous.
+collisions:
+  - "KR (Kranj) vs KK (Krško) vs KP (Koper) — three K- codes, no overlap, listed only because they are the set most often mistyped for one another."
+  - "MV is NOT an area code. It is the fixed prefix of the starodobno (historic) plate, 40. člen: 'nato sledi oznaka za starodobna vozila »MV«'. A parse API must not look it up here."
+  - "PR is NOT an area code either. It is the red 'preskušnja' field on a test plate, 41. člen (2), and it sits AFTER the area code rather than replacing it."
+  - "CMD / CD / CC / M are activity codes on foreign-mission plates (39. člen (2)) and REPLACE the area code entirely; those plates carry no arms at all (28. člen (2))."
+
+# ---------------------------------------------------------------------------
+# The eleven areas. `upravne_enote` is the list from 29. člen (1) verbatim;
+# each one's SEAT MUNICIPALITY supplies one distinct coat of arms under
+# 28. člen (1), which is why this list is the arms key and not just trivia.
+# ---------------------------------------------------------------------------
+codes:
+  - code: CE
+    region: "Celje"
+    priloga_iv_no: 1
+    upravne_enote: ["Celje", "Laško", "Mozirje", "Slovenske Konjice", "Šentjur pri Celju", "Šmarje pri Jelšah", "Velenje", "Žalec"]
+  - code: KP
+    region: "Koper"
+    priloga_iv_no: 2
+    upravne_enote: ["Izola", "Koper", "Piran", "Sežana", "Ilirska Bistrica"]
+  - code: KR
+    region: "Kranj"
+    priloga_iv_no: 3
+    upravne_enote: ["Jesenice", "Kranj", "Radovljica", "Škofja Loka", "Tržič"]
+  - code: KK
+    region: "Krško"
+    priloga_iv_no: 4
+    upravne_enote: ["Brežice", "Krško", "Sevnica"]
+  - code: LJ
+    region: "Ljubljana"
+    priloga_iv_no: 5
+    upravne_enote: ["Domžale", "Grosuplje", "Hrastnik", "Kamnik", "Kočevje", "Litija", "Ljubljana", "Logatec", "Ribnica", "Trbovlje", "Vrhnika", "Zagorje ob Savi", "Cerknica"]
+  - code: MB
+    region: "Maribor"
+    priloga_iv_no: 6
+    upravne_enote: ["Lenart", "Maribor", "Ormož", "Pesnica", "Ptuj", "Ruše", "Slovenska Bistrica"]
+  - code: MS
+    region: "Murska Sobota"
+    priloga_iv_no: 7
+    upravne_enote: ["Gornja Radgona", "Lendava", "Ljutomer", "Murska Sobota"]
+  - code: GO
+    region: "Nova Gorica"
+    priloga_iv_no: 8
+    upravne_enote: ["Ajdovščina", "Idrija", "Nova Gorica", "Tolmin"]
+  - code: NM
+    region: "Novo mesto"
+    priloga_iv_no: 9
+    upravne_enote: ["Črnomelj", "Metlika", "Novo mesto", "Trebnje"]
+  - code: PO
+    region: "Postojna"
+    priloga_iv_no: 10
+    upravne_enote: ["Postojna"]
+  - code: SG
+    region: "Slovenj Gradec"
+    priloga_iv_no: 11
+    upravne_enote: ["Dravograd", "Radlje ob Dravi", "Ravne na Koroškem", "Slovenj Gradec"]