Specs

Bases: StrEnum

              flowchart TD
              matchlab.specs.ModelType[ModelType]

              

              click matchlab.specs.ModelType href "" "matchlab.specs.ModelType"
            

Whether a model is a deduper or a linker.

Attributes:

• LINKER –
• DEDUPER –

LINKER = 'linker'

DEDUPER = 'deduper'

Bases: BaseModel

              flowchart TD
              matchlab.specs.SourceSpec[SourceSpec]

              

              click matchlab.specs.SourceSpec href "" "matchlab.specs.SourceSpec"
            

Specification of a source: its name, its key, and the location it reads.

There is no separate list of indexed fields. What the location returns is the single declaration of what a source is. Every column is part of the record, and therefore part of that record's identity. A column you do not want to affect identity is a column that should not come back.

Attributes:

• model_config –
• location_class (str) –
• location_settings (dict[str, Any]) –
• name (str) –
• key_field (str) –

model_config = ConfigDict(frozen=True)

location_class: str = Field(description='The registered name of the Location subclass to read through.')

location_settings: dict[str, Any] = Field(default_factory=dict, description="That location's configuration, dumped to JSON — the query, for a `RelationalDB`. Never its resources, which a document names separately and a fingerprint never sees.")

name: str = Field(description="The source's name within the plan. Part of the spec because it is part of the output: it prefixes every column this source contributes and tags its rows in a resolver's output.")

key_field: str = Field(description=dedent("\n            The name of the key field. This is the source's key for unique\n            entities, such as a primary key in a relational database.\n\n            Keys are always read as strings, whatever the location returns.\n\n            For example, if the source describes companies, it may have used\n            a Companies House number as its key.\n\n            This key is ALWAYS correct. It should be something generated and\n            owned by the source being indexed.\n\n            For example, your organisation's CRM ID is a key field within the CRM.\n\n            A CRM ID entered by hand in another dataset shouldn't be used\n            as a key field.\n        "))

Bases: BaseModel

              flowchart TD
              matchlab.specs.TransformSpec[TransformSpec]

              

              click matchlab.specs.TransformSpec href "" "matchlab.specs.TransformSpec"
            

Specification of a transform, naming the transformer that reshapes the step.

Attributes:

• model_config –
• transformer_class (str) –
• transformer_settings (dict[str, Any]) –

model_config = ConfigDict(frozen=True)

transformer_class: str = Field(description='The registered name of the Transformer subclass.')

transformer_settings: dict[str, Any] = Field(description="That transformer's configuration, dumped to JSON.")

Bases: BaseModel

              flowchart TD
              matchlab.specs.ModelSpec[ModelSpec]

              

              click matchlab.specs.ModelSpec href "" "matchlab.specs.ModelSpec"
            

Specification of a deduper or linker.

Attributes:

• model_config –
• model_type (ModelType) –
• model_class (str) –
• model_settings (dict[str, Any]) –

model_config = ConfigDict(frozen=True)

model_type: ModelType = Field(description='Whether this is a deduper or a linker.')

model_class: str = Field(description='The registered name of the Deduper or Linker subclass.')

model_settings: dict[str, Any] = Field(description="That class's settings, dumped to JSON.")

Bases: BaseModel

              flowchart TD
              matchlab.specs.ResolverSpec[ResolverSpec]

              

              click matchlab.specs.ResolverSpec href "" "matchlab.specs.ResolverSpec"
            

Specification of a resolver over one or more models.

Attributes:

• model_config –
• resolver_class (str) –
• resolver_settings (dict[str, Any]) –

model_config = ConfigDict(frozen=True)

resolver_class: str = Field(description='The registered name of the ResolverMethod subclass.')

resolver_settings: dict[str, Any] = Field(description="That class's settings, dumped to JSON. Settings that point at one of the resolver's inputs key by the input's position, so no name appears here.")