aeat.adapters.inbound.borrador._extractors.modelo_100_summary_v2025 module

Modelo 100 (IRPF / Renta) observed-value extractor implementation.

The extractor reads printed casilla/value rows from a Renta artefact using the year-stable NNNN label amount grammar used by the supported 2021-2025 extractor registrations and exercised by the checked 2021-2023 fixture PDFs plus the generated 2025 parser tests. The class name records the original 2025 implementation point, but adapters.inbound.borrador._extractors deliberately maps every supported year to this implementation while the observed row grammar remains stable.

This is a read-only inbound adapter. It does not define Modelo 100 completeness, resolve a RegistrySnapshot, or make filing-grade authority decisions. When callers pass a BorradorExtractionProfile, the extractor filters to that profile and fails hard if observed coverage is insufficient.

Rows are extracted from the concatenated text stream, so ExtractedCasilla records preserve the printed value and confidence but do not capture per-row bounding boxes.

class Modelo100ObservedV2025Extractor[source]

Bases: object

Concrete Modelo 100 observed-value extractor implementation.

Reads the printed text via the backend facade’s extract_pages_text primitive, locates printed casilla rows, and returns a strict InboundBorradorObservation.

Variables:

año – The original implementation year. The extractor registry may map additional years to this class while the observed row grammar stays compatible.

año: ClassVar[int]
extract(pdf_path, artefact_kind, extraction_profile=None)[source]

Parse pdf_path into a InboundBorradorObservation.

Parameters:

  • pdf_path (Path) – Path to the Modelo 100 PDF.

  • artefact_kind (ArtefactKind) – The artefact kind discovered by detect_artefact_kind() (or supplied by the caller as an override).

  • extraction_profile (BorradorExtractionProfile | None) – Optional caller-supplied registry-profile projection that declares target casillas and minimum coverage for this parse.

Return type:

InboundBorradorObservation

Returns:

The strict InboundBorradorObservation with observed casillas extracted.

Raises:

BorradorParseError – When required header fields are missing, when a DECLARACION artefact lacks a CSV stamp, or when a supplied registry-profile projection does not meet its minimum coverage.