vehicles 0.6.1 → 0.8.0

data/data/plates/_decode/fr-departements.yml

@@ -0,0 +1,191 @@
+# plates/_decode/fr-departements.yml — the French département numbers, and the
+# one thing a parse API must never do with them.
+#
+# THE FINDING THAT SHAPES THIS FILE. France has TWO département dimensions and
+# they behave in opposite ways:
+#
+#   1. FNI (1950/1984-2009) — the département number was the LAST GROUP OF THE
+#      REGISTRATION NUMBER ITSELF, and it was geographic. The arrêté du
+#      5 novembre 1984, art. 2, verbatim: "les certificats d'immatriculation
+#      sont délivrées soit par la préfecture du département du domicile du
+#      propriétaire" — the plate was issued by the préfecture of the owner's
+#      domicile, and a move to another département triggered a re-registration
+#      (art. 10 A/B, mutation). An FNI plate therefore DECODES, and this table
+#      is how.
+#
+#   2. SIV (2009-) — the département number is NOT part of the registration
+#      number at all. It sits in a right-hand blue band called the
+#      "identifiant territorial", it is a PLATE GRAPHIC, and the arrêté du
+#      9 février 2009 fixant les caractéristiques des plaques, art. 9, says in
+#      terms: "Le choix de cet identifiant territorial est libre et peut ne pas
+#      avoir de lien avec le domicile du titulaire du certificat
+#      d'immatriculation." A SIV plate DOES NOT DECODE. Reading region off a
+#      modern French plate is the single most expensive mistake a plate-parse
+#      API can make, and PRD-PLATES §2.4 already names it ("France's SIV
+#      département band is OWNER-CHOSEN, not geographic ... the decode table
+#      must say so or the parse API will lie"). It says so.
+#
+# WHAT THE SIV BAND *DOES* CONSTRAIN, and it is worth carrying: art. 9 requires
+# the band to be "constitué par le logo officiel d'une région ou de la
+# collectivité européenne d'Alsace ET le numéro de l'un des départements de
+# cette collectivité territoriale". So the PAIR (region logo, département
+# number) must be internally consistent even though the pair is freely chosen —
+# which is why every row below carries its région. A band showing the Bretagne
+# logo over "75" is not a legal plate; a Parisian owner showing Bretagne/29 is.
+# The band must also be identical front and rear (art. 9), and it is absent
+# altogether on "véhicule de collection", "véhicule importé en transit" and
+# "véhicule en transit temporaire" plates.
+#
+# SOURCE OF THE TABLE. The rows are the official French administrative
+# découpage as served by the État's own API Géo (geo.api.gouv.fr, DINUM/
+# Etalab), which republishes the INSEE Code officiel géographique. 101 rows:
+# 96 metropolitan départements including Corse-du-Sud (2A) and Haute-Corse (2B),
+# plus the five overseas départements 971-974 and 976. Fetched 2026-08-02.
+#
+# PERIOD DISCIPLINE (PRD-PLATES §2.4 — decode tables rot). This is the list in
+# force TODAY. Historic FNI codes that are NOT in it are NOT invented here:
+#   * 20 = Corse before the 1976 split into 2A/2B;
+#   * 75 covering the whole former Seine before the 1968 Île-de-France
+#     reorganisation that created 91, 92, 93, 94 and 95;
+#   * 78 (Seine-et-Oise) before the same reorganisation;
+#   * 975 Saint-Pierre-et-Miquelon, 977/978 Saint-Barthélemy/Saint-Martin,
+#     986/987/988 (Wallis, Polynésie, Nouvelle-Calédonie) — collectivités, not
+#     départements, and not in this primary;
+#   * the pre-1962 Algerian and Saharan codes.
+# A plate older than the current découpage must be treated as UNRESOLVED rather
+# than mapped through this table. The FNI series in plates/fr.yml carries
+# matching: recall-only for the same reason.
+jurisdiction: fr
+topic: departements
+authority:
+  name: API Géo (découpage administratif) — DINUM/Etalab, sur le Code officiel géographique de l'INSEE
+  url: "https://geo.api.gouv.fr/decoupage-administratif/departements"
+sources:
+  - "https://geo.api.gouv.fr/departements?fields=code,nom,region  # the 101 rows below, code + nom + région, fetched 2026-08-02"
+  - "https://www.insee.fr/fr/information/8740222  # INSEE Code officiel géographique au 1er janvier 2026 — the upstream register API Géo republishes"
+  - "https://www.legifrance.gouv.fr/loda/id/JORFTEXT000020237128  # arrêté du 9 février 2009 fixant les caractéristiques des plaques, art. 9 — the identifiant territorial, its (région, département) pairing rule, and the free-choice sentence"
+  - "https://www.legifrance.gouv.fr/loda/id/LEGITEXT000006075077  # arrêté du 5 novembre 1984, art. 2 and art. 10 — the FNI's préfecture-of-domicile issuance and the mutation-on-move rule that make the FNI suffix decodable"
+period: { start: 2026 }
+period_evidence: instrument-in-force
+period_note: >-
+  2026 is the vintage of the COG edition read here, not the date these codes
+  came into use. Most are unchanged since 1790; the material changes for plate
+  decoding are 1968 (Île-de-France), 1976 (Corse 20 -> 2A/2B) and 2011 (Mayotte
+  976 becomes a département). None of those transitions is sourced in this pass.
+
+# Decoding rules a consumer needs:
+#   * FNI numbers END with the département group, so parse RIGHT to LEFT: take
+#     the trailing 2 or 3 characters, then the letter group, then the digits.
+#   * The group is 2 characters for metropolitan France (including the literals
+#     2A and 2B) and 3 digits for the overseas départements.
+#   * SIV numbers contain no département group at all. If a consumer has the
+#     BAND as a separate observation (from an image, never from the string),
+#     it decodes to a place of the owner's choosing, not of registration.
+applies_to:
+  - "plates/fr.yml#fr-fni-1984      # geographic, decodable"
+  - "plates/fr.yml#fr-siv-standard  # NOT decodable — band only, owner-chosen"
+decodes_registration_number: fni-only
+maximal_munch: false
+
+codes:
+  - { code: "01", name: "Ain",                       region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "02", name: "Aisne",                     region: "Hauts-de-France",         region_code: "32" }
+  - { code: "03", name: "Allier",                    region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "04", name: "Alpes-de-Haute-Provence",   region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "05", name: "Hautes-Alpes",              region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "06", name: "Alpes-Maritimes",           region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "07", name: "Ardèche",                   region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "08", name: "Ardennes",                  region: "Grand Est",               region_code: "44" }
+  - { code: "09", name: "Ariège",                    region: "Occitanie",               region_code: "76" }
+  - { code: "10", name: "Aube",                      region: "Grand Est",               region_code: "44" }
+  - { code: "11", name: "Aude",                      region: "Occitanie",               region_code: "76" }
+  - { code: "12", name: "Aveyron",                   region: "Occitanie",               region_code: "76" }
+  - { code: "13", name: "Bouches-du-Rhône",          region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "14", name: "Calvados",                  region: "Normandie",               region_code: "28" }
+  - { code: "15", name: "Cantal",                    region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "16", name: "Charente",                  region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "17", name: "Charente-Maritime",         region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "18", name: "Cher",                      region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "19", name: "Corrèze",                   region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "21", name: "Côte-d'Or",                 region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "22", name: "Côtes-d'Armor",             region: "Bretagne",                region_code: "53" }
+  - { code: "23", name: "Creuse",                    region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "24", name: "Dordogne",                  region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "25", name: "Doubs",                     region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "26", name: "Drôme",                     region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "27", name: "Eure",                      region: "Normandie",               region_code: "28" }
+  - { code: "28", name: "Eure-et-Loir",              region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "29", name: "Finistère",                 region: "Bretagne",                region_code: "53" }
+  - { code: "2A", name: "Corse-du-Sud",              region: "Corse",                   region_code: "94" }
+  - { code: "2B", name: "Haute-Corse",               region: "Corse",                   region_code: "94" }
+  - { code: "30", name: "Gard",                      region: "Occitanie",               region_code: "76" }
+  - { code: "31", name: "Haute-Garonne",             region: "Occitanie",               region_code: "76" }
+  - { code: "32", name: "Gers",                      region: "Occitanie",               region_code: "76" }
+  - { code: "33", name: "Gironde",                   region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "34", name: "Hérault",                   region: "Occitanie",               region_code: "76" }
+  - { code: "35", name: "Ille-et-Vilaine",           region: "Bretagne",                region_code: "53" }
+  - { code: "36", name: "Indre",                     region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "37", name: "Indre-et-Loire",            region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "38", name: "Isère",                     region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "39", name: "Jura",                      region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "40", name: "Landes",                    region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "41", name: "Loir-et-Cher",              region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "42", name: "Loire",                     region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "43", name: "Haute-Loire",               region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "44", name: "Loire-Atlantique",          region: "Pays de la Loire",        region_code: "52" }
+  - { code: "45", name: "Loiret",                    region: "Centre-Val de Loire",     region_code: "24" }
+  - { code: "46", name: "Lot",                       region: "Occitanie",               region_code: "76" }
+  - { code: "47", name: "Lot-et-Garonne",            region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "48", name: "Lozère",                    region: "Occitanie",               region_code: "76" }
+  - { code: "49", name: "Maine-et-Loire",            region: "Pays de la Loire",        region_code: "52" }
+  - { code: "50", name: "Manche",                    region: "Normandie",               region_code: "28" }
+  - { code: "51", name: "Marne",                     region: "Grand Est",               region_code: "44" }
+  - { code: "52", name: "Haute-Marne",               region: "Grand Est",               region_code: "44" }
+  - { code: "53", name: "Mayenne",                   region: "Pays de la Loire",        region_code: "52" }
+  - { code: "54", name: "Meurthe-et-Moselle",        region: "Grand Est",               region_code: "44" }
+  - { code: "55", name: "Meuse",                     region: "Grand Est",               region_code: "44" }
+  - { code: "56", name: "Morbihan",                  region: "Bretagne",                region_code: "53" }
+  - { code: "57", name: "Moselle",                   region: "Grand Est",               region_code: "44" }
+  - { code: "58", name: "Nièvre",                    region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "59", name: "Nord",                      region: "Hauts-de-France",         region_code: "32" }
+  - { code: "60", name: "Oise",                      region: "Hauts-de-France",         region_code: "32" }
+  - { code: "61", name: "Orne",                      region: "Normandie",               region_code: "28" }
+  - { code: "62", name: "Pas-de-Calais",             region: "Hauts-de-France",         region_code: "32" }
+  - { code: "63", name: "Puy-de-Dôme",               region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "64", name: "Pyrénées-Atlantiques",      region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "65", name: "Hautes-Pyrénées",           region: "Occitanie",               region_code: "76" }
+  - { code: "66", name: "Pyrénées-Orientales",       region: "Occitanie",               region_code: "76" }
+  - { code: "67", name: "Bas-Rhin",                  region: "Grand Est",               region_code: "44" }
+  - { code: "68", name: "Haut-Rhin",                 region: "Grand Est",               region_code: "44" }
+  - { code: "69", name: "Rhône",                     region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "70", name: "Haute-Saône",               region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "71", name: "Saône-et-Loire",            region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "72", name: "Sarthe",                    region: "Pays de la Loire",        region_code: "52" }
+  - { code: "73", name: "Savoie",                    region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "74", name: "Haute-Savoie",              region: "Auvergne-Rhône-Alpes",    region_code: "84" }
+  - { code: "75", name: "Paris",                     region: "Île-de-France",           region_code: "11" }
+  - { code: "76", name: "Seine-Maritime",            region: "Normandie",               region_code: "28" }
+  - { code: "77", name: "Seine-et-Marne",            region: "Île-de-France",           region_code: "11" }
+  - { code: "78", name: "Yvelines",                  region: "Île-de-France",           region_code: "11" }
+  - { code: "79", name: "Deux-Sèvres",               region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "80", name: "Somme",                     region: "Hauts-de-France",         region_code: "32" }
+  - { code: "81", name: "Tarn",                      region: "Occitanie",               region_code: "76" }
+  - { code: "82", name: "Tarn-et-Garonne",           region: "Occitanie",               region_code: "76" }
+  - { code: "83", name: "Var",                       region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "84", name: "Vaucluse",                  region: "Provence-Alpes-Côte d'Azur", region_code: "93" }
+  - { code: "85", name: "Vendée",                    region: "Pays de la Loire",        region_code: "52" }
+  - { code: "86", name: "Vienne",                    region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "87", name: "Haute-Vienne",              region: "Nouvelle-Aquitaine",      region_code: "75" }
+  - { code: "88", name: "Vosges",                    region: "Grand Est",               region_code: "44" }
+  - { code: "89", name: "Yonne",                     region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "90", name: "Territoire de Belfort",     region: "Bourgogne-Franche-Comté", region_code: "27" }
+  - { code: "91", name: "Essonne",                   region: "Île-de-France",           region_code: "11" }
+  - { code: "92", name: "Hauts-de-Seine",            region: "Île-de-France",           region_code: "11" }
+  - { code: "93", name: "Seine-Saint-Denis",         region: "Île-de-France",           region_code: "11" }
+  - { code: "94", name: "Val-de-Marne",              region: "Île-de-France",           region_code: "11" }
+  - { code: "95", name: "Val-d'Oise",                region: "Île-de-France",           region_code: "11" }
+  - { code: "971", name: "Guadeloupe",                region: "Guadeloupe",              region_code: "01" }
+  - { code: "972", name: "Martinique",                region: "Martinique",              region_code: "02" }
+  - { code: "973", name: "Guyane",                    region: "Guyane",                  region_code: "03" }
+  - { code: "974", name: "La Réunion",                region: "La Réunion",              region_code: "04" }
+  - { code: "976", name: "Mayotte",                   region: "Mayotte",                 region_code: "06" }

data/data/plates/_decode/gb-age-identifiers.yml

@@ -0,0 +1,170 @@
+# plates/_decode/gb-age-identifiers.yml — the British age identifier.
+#
+# THE SOURCE FINDING: this is the most precisely decodable date signal in the
+# whole plates dataset, and — like the memory tags — it is NOT law. VERA 1994
+# § 23(4) does not reach the composition of a mark, so the twice-yearly code
+# is an administrative allocation published in DVLA's leaflet INF104 and, for
+# the export series, in HMRC's VAT manual. Both are authority publications and
+# therefore PRIMARY under PRD-PLATES § 4.
+#
+# THE RULE, and it is complete and mechanical:
+#   a vehicle first registered March-August of year YY takes the code YY
+#   a vehicle first registered September of YY to February of YY+1 takes YY+50
+# INF104 § 05 tabulates it from "Sept 2001 - Feb 2002 = 51" to "March 2029 -
+# Aug 2029 = 29 / Sept 2029 - Feb 2030 = 79" and closes: "This pattern will
+# continue until all possible variations have been used." HMRC VATNINMT3500
+# states the SAME rule for the export series and carries its table to
+# "March 2049 - August 2049 = 49" and "Sept 2049 - Feb 2050 = 99".
+#
+# EVERY ROW BELOW CARRIES ITS SOURCE. Rows marked `source: inf104` are printed
+# in INF104's table. Rows marked `source: hmrc-continuation` are printed in
+# HMRC's table but not in INF104's — they are as sourced as the others, from a
+# second government publication, and are marked separately only so a verifier
+# can see which authority each half came from. NOTHING here is extrapolated
+# past 2049/50: the codes after that are undecided ("until all possible
+# variations have been used" is not a formula) and are recorded as a gap.
+#
+# Referenced by: gb-standard-2001, gb-motorcycle-2001, gb-green-zev-2020,
+# gb-export-x (plates/gb.yml) via age_encoding.
+jurisdiction: gb
+topic: age-identifiers
+authority:
+  name: Driver and Vehicle Licensing Agency (DVLA)
+  url: "https://www.gov.uk/government/publications/vehicle-registration-numbers-and-number-plates"
+sources:
+  - "https://assets.publishing.service.gov.uk/media/6694e379fc8e12ac3edafc60/inf104-vehicle-registration-numbers-and-number-plates.pdf  # INF104 (4/22) § 05 'Age identifiers' — every row marked source: inf104, plus the continuation sentence"
+  - "https://www.gov.uk/hmrc-internal-manuals/vat-new-means-transport-northern-ireland/vatninmt3500  # HMRC VATNINMT3500 — the same rule stated for the export series and tabulated to 'Sept 2049 - Feb 2050 = 99'; every row marked source: hmrc-continuation"
+licence: "Crown copyright, Open Government Licence v3.0"
+period: { start: 2001 }
+period_evidence: enabling-instrument
+period_note: >-
+  The scheme began with the current format on 1 September 2001 (INF104 § 03;
+  S.I. 2001/561 reg 1(1)(b)), and 51 is the first identifier issued.
+
+rule:
+  march_to_august: "code = YY (the two-digit year)"
+  september_to_february: "code = YY + 50 (of the year the period STARTS in)"
+  changeover_days: ["1 March", "1 September"]
+  first_code: "51"
+  last_published_code: "99"
+  values_never_issued: ["00", "01", "50"]
+  values_never_issued_note: >-
+    DERIVED, NOT QUOTED. 00 and 50 cannot arise from the rule. 01 would be
+    March-August 2001, which precedes the format's 1 September 2001 start, and
+    INF104's table opens at 51. This is the exclusion `format.regex_strict`
+    encodes in plates/gb.yml; `format.regex` deliberately does not.
+
+direction_of_inference: >-
+  ONE-SIDED, and a parse API must say so. The age identifier gives the period
+  in which the MARK was issued, which is an UPPER bound on the vehicle's age
+  and never a lower one, because a mark may be transferred to an older vehicle
+  and INF104 § 04 forbids only the reverse: "You can't use a registration
+  number to make your vehicle appear younger than it actually is." A 21-plate
+  car was registered no earlier than March 2021; a 21-plate car may lawfully
+  BE older than that only if the mark was assigned to it at first registration
+  in that window, which is the ordinary case — but a personalised mark on the
+  same car could be from any era, so the inference runs only from mark to
+  earliest-possible-registration, never from mark to build year.
+
+codes:
+  - { code: "51", from: "2001-09", to: "2002-02", source: inf104 }
+  - { code: "02", from: "2002-03", to: "2002-08", source: inf104 }
+  - { code: "52", from: "2002-09", to: "2003-02", source: inf104 }
+  - { code: "03", from: "2003-03", to: "2003-08", source: inf104 }
+  - { code: "53", from: "2003-09", to: "2004-02", source: inf104 }
+  - { code: "04", from: "2004-03", to: "2004-08", source: inf104 }
+  - { code: "54", from: "2004-09", to: "2005-02", source: inf104 }
+  - { code: "05", from: "2005-03", to: "2005-08", source: inf104 }
+  - { code: "55", from: "2005-09", to: "2006-02", source: inf104 }
+  - { code: "06", from: "2006-03", to: "2006-08", source: inf104 }
+  - { code: "56", from: "2006-09", to: "2007-02", source: inf104 }
+  - { code: "07", from: "2007-03", to: "2007-08", source: inf104 }
+  - { code: "57", from: "2007-09", to: "2008-02", source: inf104 }
+  - { code: "08", from: "2008-03", to: "2008-08", source: inf104 }
+  - { code: "58", from: "2008-09", to: "2009-02", source: inf104 }
+  - { code: "09", from: "2009-03", to: "2009-08", source: inf104 }
+  - { code: "59", from: "2009-09", to: "2010-02", source: inf104 }
+  - { code: "10", from: "2010-03", to: "2010-08", source: inf104 }
+  - { code: "60", from: "2010-09", to: "2011-02", source: inf104 }
+  - { code: "11", from: "2011-03", to: "2011-08", source: inf104 }
+  - { code: "61", from: "2011-09", to: "2012-02", source: inf104 }
+  - { code: "12", from: "2012-03", to: "2012-08", source: inf104 }
+  - { code: "62", from: "2012-09", to: "2013-02", source: inf104 }
+  - { code: "13", from: "2013-03", to: "2013-08", source: inf104 }
+  - { code: "63", from: "2013-09", to: "2014-02", source: inf104 }
+  - { code: "14", from: "2014-03", to: "2014-08", source: inf104 }
+  - { code: "64", from: "2014-09", to: "2015-02", source: inf104 }
+  - { code: "15", from: "2015-03", to: "2015-08", source: inf104 }
+  - { code: "65", from: "2015-09", to: "2016-02", source: inf104 }
+  - { code: "16", from: "2016-03", to: "2016-08", source: inf104 }
+  - { code: "66", from: "2016-09", to: "2017-02", source: inf104 }
+  - { code: "17", from: "2017-03", to: "2017-08", source: inf104 }
+  - { code: "67", from: "2017-09", to: "2018-02", source: inf104 }
+  - { code: "18", from: "2018-03", to: "2018-08", source: inf104 }
+  - { code: "68", from: "2018-09", to: "2019-02", source: inf104 }
+  - { code: "19", from: "2019-03", to: "2019-08", source: inf104 }
+  - { code: "69", from: "2019-09", to: "2020-02", source: inf104 }
+  - { code: "20", from: "2020-03", to: "2020-08", source: inf104 }
+  - { code: "70", from: "2020-09", to: "2021-02", source: inf104 }
+  - { code: "21", from: "2021-03", to: "2021-08", source: inf104 }
+  - { code: "71", from: "2021-09", to: "2022-02", source: inf104 }
+  - { code: "22", from: "2022-03", to: "2022-08", source: inf104 }
+  - { code: "72", from: "2022-09", to: "2023-02", source: inf104 }
+  - { code: "23", from: "2023-03", to: "2023-08", source: inf104 }
+  - { code: "73", from: "2023-09", to: "2024-02", source: inf104 }
+  - { code: "24", from: "2024-03", to: "2024-08", source: inf104 }
+  - { code: "74", from: "2024-09", to: "2025-02", source: inf104 }
+  - { code: "25", from: "2025-03", to: "2025-08", source: inf104 }
+  - { code: "75", from: "2025-09", to: "2026-02", source: inf104 }
+  - { code: "26", from: "2026-03", to: "2026-08", source: inf104 }
+  - { code: "76", from: "2026-09", to: "2027-02", source: inf104 }
+  - { code: "27", from: "2027-03", to: "2027-08", source: inf104 }
+  - { code: "77", from: "2027-09", to: "2028-02", source: inf104 }
+  - { code: "28", from: "2028-03", to: "2028-08", source: inf104 }
+  - { code: "78", from: "2028-09", to: "2029-02", source: inf104 }
+  - { code: "29", from: "2029-03", to: "2029-08", source: inf104 }
+  - { code: "79", from: "2029-09", to: "2030-02", source: inf104 }
+  - { code: "30", from: "2030-03", to: "2030-08", source: hmrc-continuation }
+  - { code: "80", from: "2030-09", to: "2031-02", source: hmrc-continuation }
+  - { code: "31", from: "2031-03", to: "2031-08", source: hmrc-continuation }
+  - { code: "81", from: "2031-09", to: "2032-02", source: hmrc-continuation }
+  - { code: "32", from: "2032-03", to: "2032-08", source: hmrc-continuation }
+  - { code: "82", from: "2032-09", to: "2033-02", source: hmrc-continuation }
+  - { code: "33", from: "2033-03", to: "2033-08", source: hmrc-continuation }
+  - { code: "83", from: "2033-09", to: "2034-02", source: hmrc-continuation }
+  - { code: "34", from: "2034-03", to: "2034-08", source: hmrc-continuation }
+  - { code: "84", from: "2034-09", to: "2035-02", source: hmrc-continuation }
+  - { code: "35", from: "2035-03", to: "2035-08", source: hmrc-continuation }
+  - { code: "85", from: "2035-09", to: "2036-02", source: hmrc-continuation }
+  - { code: "36", from: "2036-03", to: "2036-08", source: hmrc-continuation }
+  - { code: "86", from: "2036-09", to: "2037-02", source: hmrc-continuation }
+  - { code: "37", from: "2037-03", to: "2037-08", source: hmrc-continuation }
+  - { code: "87", from: "2037-09", to: "2038-02", source: hmrc-continuation }
+  - { code: "38", from: "2038-03", to: "2038-08", source: hmrc-continuation }
+  - { code: "88", from: "2038-09", to: "2039-02", source: hmrc-continuation }
+  - { code: "39", from: "2039-03", to: "2039-08", source: hmrc-continuation }
+  - { code: "89", from: "2039-09", to: "2040-02", source: hmrc-continuation }
+  - { code: "40", from: "2040-03", to: "2040-08", source: hmrc-continuation }
+  - { code: "90", from: "2040-09", to: "2041-02", source: hmrc-continuation }
+  - { code: "41", from: "2041-03", to: "2041-08", source: hmrc-continuation }
+  - { code: "91", from: "2041-09", to: "2042-02", source: hmrc-continuation }
+  - { code: "42", from: "2042-03", to: "2042-08", source: hmrc-continuation }
+  - { code: "92", from: "2042-09", to: "2043-02", source: hmrc-continuation }
+  - { code: "43", from: "2043-03", to: "2043-08", source: hmrc-continuation }
+  - { code: "93", from: "2043-09", to: "2044-02", source: hmrc-continuation }
+  - { code: "44", from: "2044-03", to: "2044-08", source: hmrc-continuation }
+  - { code: "94", from: "2044-09", to: "2045-02", source: hmrc-continuation }
+  - { code: "45", from: "2045-03", to: "2045-08", source: hmrc-continuation }
+  - { code: "95", from: "2045-09", to: "2046-02", source: hmrc-continuation }
+  - { code: "46", from: "2046-03", to: "2046-08", source: hmrc-continuation }
+  - { code: "96", from: "2046-09", to: "2047-02", source: hmrc-continuation }
+  - { code: "47", from: "2047-03", to: "2047-08", source: hmrc-continuation }
+  - { code: "97", from: "2047-09", to: "2048-02", source: hmrc-continuation }
+  - { code: "48", from: "2048-03", to: "2048-08", source: hmrc-continuation }
+  - { code: "98", from: "2048-09", to: "2049-02", source: hmrc-continuation }
+  - { code: "49", from: "2049-03", to: "2049-08", source: hmrc-continuation }
+  - { code: "99", from: "2049-09", to: "2050-02", source: hmrc-continuation }
+
+not_modelled:
+  - "Codes for periods after February 2050. INF104 says the pattern continues 'until all possible variations have been used' and neither authority publishes what happens when 99 is reached; extrapolating past the published tables was declined."
+  - "The Northern Irish and pre-2001 systems encode no date at all — see plates/gb.yml#gb-ni-standard and #gb-prefix-1984 — so this table must never be applied to them."

data/data/plates/_decode/gb-memory-tags.yml

@@ -0,0 +1,202 @@
+# plates/_decode/gb-memory-tags.yml — DVLA memory tags (current British format).
+#
+# THE SOURCE FINDING, and it is the mirror image of the German one: this table
+# is NOT in any statutory instrument and never could be. VERA 1994 § 23(4)
+# empowers regulations over "(a) the size, shape and character of registration
+# marks to be fixed on any vehicle, and (b) the manner in which registration
+# marks are to be displayed" — not over WHICH letters a mark is made of. The
+# composition of a British mark is an administrative allocation by the
+# Secretary of State under § 23(1), and DVLA publishes it in the leaflet
+# INF104. So where plates/_decode/es-provinces.yml is enacted law and the
+# German district list is a Bundesanzeiger publication, this is a leaflet —
+# and a leaflet from the issuing authority is still a PRIMARY (PRD-PLATES § 4:
+# "Government/authority pages ... PRIMARY").
+#
+# Every row below is INF104 (4/22) § 05 "DVLA memory tags and age identifiers",
+# transcribed in the leaflet's own order and its own grouping (postal area
+# letter -> DVLA office -> tag list). WHERE INF104 PRINTS A RANGE it is kept as
+# a range and NOT expanded, because expanding it would require deciding what
+# the range does with I, Q and Z — a decision the leaflet makes elsewhere, in
+# prose, and which is recorded here as `range_rule` rather than baked silently
+# into 25 invented rows.
+#
+# EVERY TAG IS QUOTED, AND IT HAS TO BE. Psych resolves the bare scalar NO to
+# the boolean false (as it does ON, OFF, TRUE, YES), and "NO" is a real
+# Newcastle memory tag. An unquoted tag list therefore silently turns one row
+# into `false` and a decode lookup misses it forever. This is the same class of
+# defect plates/de.yml records for bare YAML dates, found here by round-tripping
+# the file through YAML.safe_load and re-checking every tag against /\A[A-Z]{2}\z/
+# — a check worth keeping in any future decode-table pass.
+#
+# Referenced by: gb-standard-2001, gb-motorcycle-2001, gb-green-zev-2020
+# (plates/gb.yml) via region_encoding.
+jurisdiction: gb
+topic: memory-tags
+authority:
+  name: Driver and Vehicle Licensing Agency (DVLA)
+  url: "https://www.gov.uk/government/publications/vehicle-registration-numbers-and-number-plates"
+sources:
+  - "https://assets.publishing.service.gov.uk/media/6694e379fc8e12ac3edafc60/inf104-vehicle-registration-numbers-and-number-plates.pdf  # INF104 (4/22) § 05 — every row of this file, plus the three charset bullets and the HW / MN+MAN notes"
+  - "https://www.gov.uk/government/publications/vehicle-registration-numbers-and-number-plates  # INF104 landing page, updated 12 April 2022 (stable citation)"
+  - "https://www.legislation.gov.uk/ukpga/1994/22/section/23  # VERA 1994 s 23(1) and (4) — why this table is administrative and not statutory"
+licence: "Crown copyright, Open Government Licence v3.0 (DVLA publication)"
+period: { start: 2001 }
+period_evidence: enabling-instrument
+period_note: >-
+  The memory-tag scheme began with the current format on 1 September 2001
+  (INF104 § 03). Individual OFFICE allocations have changed since — DVLA closed
+  its local offices in 2013 and the office names below are the issuing offices
+  as INF104 (4/22) prints them, which a decode must treat as the historical
+  allocation point, not as a list of open counters. Rows are never deleted:
+  a mark issued under a withdrawn allocation stays on the road, exactly as
+  plates/_decode/de-districts.yml requires for German codes.
+
+# What the tag means, and what it does NOT mean.
+semantics: >-
+  INF104 § 03: the first two characters are a "2 letter memory tag (these refer
+  to the region in the country where a vehicle is first registered)". The FIRST
+  letter is the postal area; the SECOND selects the DVLA office within it.
+caveat: >-
+  A memory tag records WHERE THE VEHICLE WAS FIRST REGISTERED. It is not the
+  keeper's address, not the vehicle's current location, and not stable across
+  a re-registration or a transfer of a personalised mark. Any consumer using
+  it as a present-tense geographic signal is using it wrongly, and a parse API
+  must return it labelled as first-registration provenance.
+
+charset_rules:
+  excluded_from_tags: ["I", "Q", "Z"]
+  z_rule: "Z is used only as a random letter (positions 5-7), never in a memory tag"
+  q_rule: "Q is excluded from memory tags because it remains in service as the prefix of the indeterminate-age Q series (plates/gb.yml#gb-q)"
+  first_letters_used: [A, B, C, D, E, F, G, H, K, L, M, N, O, P, R, S, V, W, Y]
+  first_letters_absent: [I, J, Q, T, U, X, Z]
+  first_letters_absent_note: >-
+    J, T, U and X simply do not appear as postal-area letters in INF104 § 05's
+    table; I, Q and Z are excluded by the leaflet's own bullet. The absence is
+    recorded as a fact about the published table, not as a rule DVLA states.
+  bullets_verbatim:
+    - "We will not use I, Q or Z in local memory tags identifiers"
+    - "We will still issue existing 'Q' marks"
+    - "We will only use Z as a random letter"
+  range_rule: >-
+    Where INF104 prints a range ("BA - BY", "EA - EY", "MA - MY", "OA - OY",
+    "RA - RY", "VA - VY") the second letter runs A to Y inclusive, minus the
+    excluded I and Q (Z is above Y and so is excluded by the range itself).
+    This expansion is DERIVED from the bullets above; it is stated here once
+    rather than written into the rows, so a consumer can see exactly which
+    codes are transcribed and which are computed.
+
+areas:
+  - letter: A
+    name: Anglia
+    offices:
+      - { office: Peterborough, tags: ["AA", "AB", "AC", "AD", "AE", "AF", "AG", "AH", "AJ", "AK", "AL", "AM", "AN"] }
+      - { office: Norwich,      tags: ["AO", "AP", "AR", "AS", "AT", "AU"] }
+      - { office: Ipswich,      tags: ["AV", "AW", "AX", "AY"] }
+  - letter: B
+    name: Birmingham
+    offices:
+      - { office: Birmingham, range: "BA-BY" }
+  - letter: C
+    name: Cymru
+    offices:
+      - { office: Cardiff, tags: ["CA", "CB", "CC", "CD", "CE", "CF", "CG", "CH", "CJ", "CK", "CL", "CM", "CN", "CO"] }
+      - { office: Swansea, tags: ["CP", "CR", "CS", "CT", "CU", "CV"] }
+      - { office: Bangor,  tags: ["CW", "CX", "CY"] }
+  - letter: D
+    name: Deeside to Shrewsbury
+    offices:
+      - { office: Chester,    tags: ["DA", "DB", "DC", "DD", "DE", "DF", "DG", "DH", "DJ", "DK"] }
+      - { office: Shrewsbury, tags: ["DL", "DM", "DN", "DO", "DP", "DR", "DS", "DT", "DU", "DV", "DW", "DX", "DY"] }
+  - letter: E
+    name: Essex
+    offices:
+      - { office: Chelmsford, range: "EA-EY" }
+  - letter: F
+    name: Forest & Fens
+    offices:
+      - { office: Nottingham, tags: ["FA", "FB", "FC", "FD", "FE", "FF", "FG", "FH", "FJ", "FK", "FL", "FM", "FN", "FP"] }
+      - { office: Lincoln,    tags: ["FR", "FS", "FT", "FV", "FW", "FX", "FY"] }
+  - letter: G
+    name: Garden of England
+    offices:
+      - { office: Maidstone, tags: ["GA", "GB", "GC", "GD", "GE", "GF", "GG", "GH", "GJ", "GK", "GL", "GM", "GN", "GO"] }
+      - { office: Brighton,  tags: ["GP", "GR", "GS", "GT", "GU", "GV", "GW", "GX", "GY"] }
+  - letter: H
+    name: Hampshire & Dorset
+    offices:
+      - { office: Bournemouth, tags: ["HA", "HB", "HC", "HD", "HE", "HF", "HG", "HH", "HJ"] }
+      - { office: Portsmouth,  tags: ["HK", "HL", "HM", "HN", "HO", "HP", "HR", "HS", "HT", "HU", "HV", "HW", "HX", "HY"],
+          note: "INF104, verbatim: '(HW will be used exclusively for Isle of Wight residents)' — the only sub-office geographic reservation in the table" }
+  - letter: K
+    name: null
+    name_note: "INF104 gives no regional name for the K area — the 'Letter' and 'Postal area' columns are blank beside Borehamwood and Northampton. Recorded as null rather than invented."
+    offices:
+      - { office: Borehamwood, tags: ["KA", "KB", "KC", "KD", "KE", "KF", "KG", "KH", "KJ", "KK", "KL"] }
+      - { office: Northampton, tags: ["KM", "KN", "KO", "KP", "KR", "KS", "KT", "KU", "KV", "KW", "KX", "KY"] }
+  - letter: L
+    name: London
+    offices:
+      - { office: Wimbledon,   tags: ["LA", "LB", "LC", "LD", "LE", "LF", "LG", "LH", "LJ"] }
+      - { office: Borehamwood, tags: ["LK", "LL", "LM", "LN", "LO", "LP", "LR", "LS", "LT"] }
+      - { office: Sidcup,      tags: ["LU", "LV", "LW", "LX", "LY"] }
+  - letter: M
+    name: Manchester & Merseyside
+    offices:
+      - { office: Manchester, range: "MA-MY",
+          note: "INF104, verbatim: '(MN + MAN Reserved for the Isle of Man)'. The Isle of Man is a separate jurisdiction (ISO 3166-1 IM) with its own registration system; MN is withheld from the British range and MAN is withheld as a three-letter group so that no British mark collides with a Manx one." }
+  - letter: N
+    name: North
+    offices:
+      - { office: Newcastle, tags: ["NA", "NB", "NC", "ND", "NE", "NG", "NH", "NJ", "NK", "NL", "NM", "NN", "NO"],
+          note: "NF is absent from INF104's Newcastle list and is not assigned elsewhere in the table; transcribed as printed, not corrected" }
+      - { office: Stockton,  tags: ["NP", "NR", "NS", "NT", "NU", "NV", "NW", "NX", "NY"] }
+  - letter: O
+    name: Oxford
+    offices:
+      - { office: Oxford, range: "OA-OY" }
+  - letter: P
+    name: Preston
+    offices:
+      - { office: Preston,  tags: ["PA", "PB", "PC", "PD", "PE", "PF", "PG", "PH", "PJ", "PK", "PL", "PM", "PN", "PO", "PP", "PR", "PS", "PT"] }
+      - { office: Carlisle, tags: ["PU", "PV", "PW", "PX", "PY"] }
+  - letter: R
+    name: Reading
+    offices:
+      - { office: Theale, range: "RA-RY", note: "the postal area is Reading; the issuing office INF104 names is Theale" }
+  - letter: S
+    name: Scotland
+    offices:
+      - { office: Glasgow,   tags: ["SA", "SB", "SC", "SD", "SE", "SF", "SG", "SH", "SJ"] }
+      - { office: Edinburgh, tags: ["SK", "SL", "SM", "SN", "SO"] }
+      - { office: Dundee,    tags: ["SP", "SR", "SS", "ST"] }
+      - { office: Aberdeen,  tags: ["SU", "SV", "SW"] }
+      - { office: Inverness, tags: ["SX", "SY"] }
+  - letter: V
+    name: Severn Valley
+    offices:
+      - { office: Worcester, range: "VA-VY" }
+  - letter: W
+    name: West of England
+    offices:
+      - { office: Exeter,  tags: ["WA", "WB", "WC", "WD", "WE", "WF", "WG", "WH", "WJ"] }
+      - { office: Truro,   tags: ["WK", "WL"] }
+      - { office: Bristol, tags: ["WM", "WN", "WO", "WP", "WR", "WS", "WT", "WU", "WV", "WW", "WX", "WY"] }
+  - letter: Y
+    name: Yorkshire
+    offices:
+      - { office: Leeds,     tags: ["YA", "YB", "YC", "YD", "YE", "YF", "YG", "YH", "YJ", "YK", "YL"] }
+      - { office: Sheffield, tags: ["YM", "YN", "YO", "YP", "YR", "YS", "YT", "YV"] }
+      - { office: Beverley,  tags: ["YU", "YW", "YX", "YY"] }
+
+# Ambiguity notes a parse API must carry.
+ambiguity:
+  - "MAXIMAL MUNCH DOES NOT APPLY HERE, unlike es-provinces: a British memory tag is ALWAYS exactly two characters, because the age identifier that follows is always exactly two digits. There is no longest-match problem."
+  - "A tag decodes only in the current (post-1.9.2001) format. The same two letters at the start of a pre-2001 mark are NOT a memory tag: in the suffix system they are the first two of a three-letter area code, and in the prefix system the leading letter is a YEAR code. A parse API must establish the series before consulting this table."
+  - "X is never a memory-tag first letter, which is what makes the export series (plates/gb.yml#gb-export-x, 'XA51 ABC') mechanically separable from an ordinary mark despite sharing its shape."
+  - "Northern Irish marks have no memory tag at all and must not be decoded through this table (see plates/gb.yml#gb-ni-standard)."
+
+# What this table deliberately does NOT contain.
+not_modelled:
+  - "The pre-2001 local-office area codes (the trailing two letters of a suffix or prefix mark). No primary was located; recorded as a gap in the gb dossier."
+  - "The Northern Irish district letters. No primary was located; recorded as a gap."
+  - "Which office issued which tag AT A GIVEN DATE. INF104 publishes one snapshot; DVLA local offices closed in 2013 and the leaflet was reissued 4/22 without a period column, so per-row period discipline is impossible from this source and is not faked."