Override max length for individual props

followthemoney/property.py

@@ -24,6 +24,7 @@ class PropertyDict(TypedDict, total=False):
     hidden: Optional[bool]
     matchable: Optional[bool]
     deprecated: Optional[bool]
+    maxLength: Optional[int]
     # stub: Optional[bool]
     rdf: Optional[str]
     range: Optional[str]
@@ -58,6 +59,7 @@ class Property:
         "type",
         "matchable",
         "deprecated",
+        "max_length",
         "_range",
         "format",
         "range",
@@ -109,6 +111,9 @@ def __init__(self, schema: "Schema", name: str, data: PropertySpec) -> None:
         else:
             self.matchable = self.type.matchable
 
+        #: The maximum length of the property value.
+        self.max_length = int(data.get("maxLength") or self.type.max_length)
+
         #: If the property is of type ``entity``, the set of valid schema to be added
         #: in this property can be constrained. For example, an asset can be owned,
         #: but a person cannot be owned.
@@ -197,6 +202,7 @@ def to_dict(self) -> PropertyToDict:
             "qname": self.qname,
             "label": self.label,
             "type": self.type.name,
+            "maxLength": self.max_length,
         }
         if self.description:
             data["description"] = self.description

followthemoney/schema/Address.yaml

@@ -54,6 +54,7 @@ Address:
     postalCode:
       label: "Postal code"
       description: "Zip code or postcode."
+      maxLength: 16
     region:
       label: "Region"
       description: "Also province or area."

followthemoney/schema/BankAccount.yaml

@@ -27,15 +27,18 @@ BankAccount:
     accountNumber:
       label: Account number
       type: identifier
+      maxLength: 64
     iban:
       label: IBAN
       # type: identifier
       # format: iban
       type: iban
+      maxLength: 64
     bic:
       label: Bank Identifier Code
       type: identifier
       format: bic
+      maxLength: 16
     bank:
       label: Bank
       type: entity

followthemoney/schema/Company.yaml

@@ -31,6 +31,7 @@ Company:
       label: "VOEN"
       description: "Azerbaijan taxpayer ID"
       type: identifier
+      maxLength: 32
     coatoCode:
       label: "COATO / SOATO / OKATO"
       type: identifier
@@ -107,6 +108,8 @@ Company:
       label: PermID
       description: LSEG/Refinitiv code for a company
       type: identifier
+      maxLength: 16
     ricCode:
       label: Reuters Instrument Code
       type: identifier
+      maxLength: 16

followthemoney/schema/CryptoWallet.yaml

@@ -23,6 +23,7 @@ CryptoWallet:
       label: Address
       description: Public key used to identify the wallet
       type: identifier
+      maxLength: 128
     privateKey:
       label: Private key
     creationDate:

followthemoney/schema/Identification.yaml

@@ -35,5 +35,6 @@ Identification:
     number:
       label: "Document number"
       type: identifier
+      maxLength: 64
     authority:
       label: "Authority"

followthemoney/schema/LegalEntity.yaml

@@ -82,6 +82,7 @@ LegalEntity:
       label: "V.A.T. Identifier"
       description: "(EU) VAT number"
       type: identifier
+      maxLength: 32
     jurisdiction:
       label: Jurisdiction
       type: country
@@ -109,30 +110,36 @@ LegalEntity:
       description: "Russian company ID"
       type: identifier
       # format: inn
+      maxLength: 32
     ogrnCode:
       label: "OGRN"
       description: "Major State Registration Number"
       type: identifier
       # format: ogrn
+      maxLength: 32
     leiCode:
       # cf. https://www.gleif.org/en/about-lei/introducing-the-legal-entity-identifier-lei
       label: "LEI"
       description: "Legal Entity Identifier"
       type: identifier
       format: lei
+      maxLength: 32
     dunsCode:
       label: "DUNS"
       description: "Data Universal Numbering System - Dun & Bradstreet identifier"
       type: identifier
+      maxLength: 32
     npiCode:
       label: "NPI"
       description: "National Provider Identifier"
       type: identifier
+      maxLength: 32
     swiftBic:
       label: "SWIFT/BIC"
       description: "Bank identifier code"
       type: identifier
       format: bic
+      maxLength: 16
     parent:
       label: "Parent company"
       description: "If this entity is a subsidiary, another entity (company or organisation) is its parent"

followthemoney/schema/Sanction.yaml

@@ -31,6 +31,7 @@ Sanction:
     unscId:
       label: "UN SC identifier"
       type: identifier
+      maxLength: 16
     program:
       label: "Program"
     provisions:

followthemoney/schema/Security.yaml

@@ -25,6 +25,7 @@ Security:
       description: International Securities Identification Number
       type: identifier
       format: isin
+      maxLength: 16
     registrationNumber:
       label: Registration number
       type: identifier
@@ -35,6 +36,7 @@ Security:
       label: Financial Instrument Global Identifier
       type: identifier
       format: figi
+      maxLength: 16
     issuer:
       label: "Issuer"
       type: entity

followthemoney/schema/Thing.yaml

@@ -58,6 +58,7 @@ Thing:
       label: Wikidata ID
       type: identifier
       format: qid
+      maxLength: 32
     keywords:
       label: Keywords
     topics:

followthemoney/schema/Vessel.yaml

@@ -20,6 +20,7 @@ Vessel:
     imoNumber:
       label: IMO Number
       type: identifier
+      maxLength: 16
     crsNumber:
       label: CRS Number
       type: identifier
@@ -50,3 +51,4 @@ Vessel:
     mmsi:
       label: MMSI
       type: identifier
+      maxLength: 16

followthemoney/types/common.py

@@ -19,7 +19,7 @@ class PropertyTypeToDict(TypedDict, total=False):
     label: str
     plural: str
     description: Optional[str]
-    max_length: int
+    maxLength: int
     group: Optional[str]
     matchable: Optional[bool]
     pivot: Optional[bool]
@@ -198,7 +198,7 @@ def to_dict(self) -> PropertyTypeToDict:
             "label": gettext(self.label),
             "plural": gettext(self.plural),
             "description": gettext(self.docs),
-            "max_length": self.max_length,
+            "maxLength": self.max_length,
         }
         if self.group:
             data["group"] = self.group

followthemoney/types/name.py

@@ -53,7 +53,11 @@ def compare(self, left: str, right: str) -> float:
         right_clean = clean_name_light(right)
         if left_clean is None or right_clean is None:
             return 0.0
-        return levenshtein_similarity(left_clean, right_clean)
+        return levenshtein_similarity(
+            left_clean,
+            right_clean,
+            max_length=self.max_length,
+        )
 
     def node_id(self, value: str) -> Optional[str]:
         slug = slugify(value)